Document the Audita v2 release

.gitignore

@@ -12,7 +12,6 @@ build/
 
 # ---> Codex
 .codex
-AGENTS.md
 
 # ---> Go
 # If you prefer the allow list template instead of the deny list, see community template:

.woodpecker/release.yml

@@ -3,35 +3,9 @@ when:
 
 steps:
   - name: build-release-assets
-    image: golang:1.25
+    image: golang:1.25.5
     commands:
-      - |
-        set -eu
-
-        version="$CI_COMMIT_TAG"
-        dist="dist"
-        pkg="gitea.maximumdirect.net/eric/audita/cmd/audita"
-
-        rm -rf "$dist"
-        mkdir -p "$dist"
-
-        build_binary() {
-          goos="$1"
-          goarch="$2"
-          suffix="$3"
-          output="$dist/audita-$version-$goos-$goarch$suffix"
-
-          CGO_ENABLED=0 GOOS="$goos" GOARCH="$goarch" \
-            go build -trimpath -ldflags "-s -w -X gitea.maximumdirect.net/eric/audita/internal/buildinfo.Version=$version" \
-            -o "$output" "$pkg"
-        }
-
-        build_binary linux amd64 ""
-        build_binary linux arm64 ""
-        build_binary darwin amd64 ""
-        build_binary darwin arm64 ""
-        build_binary windows amd64 ".exe"
-        build_binary windows arm64 ".exe"
+      - bash scripts/build-release.sh "$CI_COMMIT_TAG" dist
 
   - name: publish-release
     image: woodpeckerci/plugin-release
@@ -41,7 +15,8 @@ steps:
       api_key:
         from_secret: GITEA_RELEASE_TOKEN
       files:
-        - dist/audita-*
+        - dist/audita-*-linux-amd64
+        - dist/audita-*-linux-arm64
       checksum: sha256
       checksum-file: SHA256SUMS
       checksum-flatten: true

AGENTS.md

@@ -0,0 +1,4 @@
+Please review `docs/development.md` for initial orientation in this repository
+and follow its task-specific reading guide.
+
+Review and follow `docs/policy/agents.md` before recommending, planning, or implementing any changes to this repository.

LICENSE

@@ -1,4 +1,4 @@
-Copyright (c) 2026 eric. 
+Copyright (c) 2026 Eric Rakestraw. 
 
 Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
 

README.md

@@ -1,303 +1,58 @@
 # Audita
 
-Audita is a transcript polishing CLI.
+Audita polishes transcript JSON using glossary-aware and LLM-backed correction
+modules. It validates proposed corrections and produces corrected text plus
+reports for review and subprocess integration.
 
-`audita process` validates transcript/glossary input, normalizes and chunks transcript segments, runs the default correction pipeline, and emits corrected transcript output plus machine-readable diagnostics and reports.
+## Quickstart
 
-## What Audita Does
-
-Default module sequence:
-- `glossary`
-- `homophones`
-- `glossary`
-- `spoken_word`
-- `grammar`
-
-Pipeline behavior includes:
-- glossary-backed domain/acoustic corrections
-- conservative homophone and mistranscription corrections
-- conservative spoken-word dysfluency cleanup with semantic guardrails
-- grammar/punctuation/capitalization/formatting cleanup
-- validator-chain enforcement before application
-- run reports and diagnostics artifacts with secret redaction
-
-## Build and Install
-
-Build a local binary:
+From a checkout, build with Go 1.25.5 or newer:
 
 ```sh
 go build -o ./bin/audita ./cmd/audita
 ```
 
-Install into your Go bin directory:
+Processing requires a reachable [compatible LLM endpoint](docs/integrations/openai-compatible-llm.md)
+and access to a model you select. The [minimal config](examples/minimal-config.yml)
+targets an unauthenticated local endpoint; edit its endpoint and
+[local profiles](examples/profiles/) to match your service and models. The
+[service example](examples/production-config.yml) demonstrates a named credential
+reference and a shared backend. Set its referenced environment variable before
+validating it. The [testing overlay](examples/overlays/testing.yml) selects both
+role profiles, but model cost, speed, and quality depend on your provider choices.
+
+Check configuration and profile selection offline:
 
 ```sh
-go install ./cmd/audita
+./bin/audita config validate --config examples/minimal-config.yml
+./bin/audita config print-effective --config examples/minimal-config.yml --view profiles
 ```
 
-CLI help:
+Run the included small input with the minimal config:
 
 ```sh
-audita --help
-audita process --help
-audita config --help
-```
-
-## Test
-
-Run all tests:
-
-```sh
-go test ./...
-```
-
-## Basic Usage
-
-Required inputs:
-- transcript JSON path (positional argument)
-- `--glossary <glossary.yaml>`
-
-Recommended run:
-
-```sh
-audita process transcript.json \
-  --glossary glossary.yaml \
-  --output corrected.json \
-  --report-json report.json
-```
-
-Select an explicit output schema (default is `bare-segments`):
-
-```sh
-audita process transcript.json \
-  --glossary glossary.yaml \
-  --output-schema audita-v1 \
-  --output corrected.json \
-  --report-json report.json
-```
-
-Recommended config-based run:
-
-```sh
-audita process transcript.json \
-  --glossary glossary.yaml \
-  --config audita.yml \
-  --output corrected.json \
-  --report-json report.json
-```
-
-Explicit module override:
-
-```sh
-audita process transcript.json \
-  --glossary glossary.yaml \
-  --modules glossary,homophones,grammar \
-  --output corrected.json \
-  --report-json report.json
-```
-
-Optional transcript background context:
-
-```sh
-audita process transcript.json \
-  --glossary glossary.yaml \
-  --transcript-description "Brief context that may help resolve ambiguous terms." \
+./bin/audita process examples/tiny-transcript.json \
+  --config examples/minimal-config.yml \
+  --glossary examples/tiny-glossary.yaml \
   --output corrected.json
 ```
 
-The transcript description is background context only and does not override transcript content.
-
-Write transcript JSON to stdout (no `--output`):
-
-```sh
-audita process transcript.json --glossary glossary.yaml
-```
-
-Control diagnostics location/retention:
-
-```sh
-audita process transcript.json \
-  --glossary glossary.yaml \
-  --work-dir /tmp/audita \
-  --work-dir-retention auto \
-  --output corrected.json \
-  --report-json report.json
-```
-
-## Stdout/Stderr Contract
-
-- With `--output`, stdout is expected to be empty on success.
-- Without `--output`, stdout contains transcript JSON only on success.
-- `--report-json` writes a file and is never printed to stdout.
-- stderr is human-readable diagnostics/errors.
-
-For subprocess orchestration guidance, see [`docs/subprocess-operations.md`](docs/subprocess-operations.md).
-
-## Configuration
-
-Precedence:
-1. defaults
-2. config file (`--config`, `AUDITA_CONFIG`, or default search paths when present: `/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml`)
-3. environment (`AUDITA_*`)
-4. CLI flags
-
-Config commands:
-
-```sh
-audita config validate --config audita.yml
-audita config print-effective --config audita.yml
-```
-
-For full config-file schema and examples, see [`docs/configuration.md`](docs/configuration.md).
-For output-schema details, see [`docs/output-schemas.md`](docs/output-schemas.md).
-For built-in validator keys and chain definitions, see [`docs/validators.md`](docs/validators.md).
-For embedded prompt assets and prompt metadata behavior, see [`docs/prompts.md`](docs/prompts.md).
-For CLI/process compatibility guarantees, see [`docs/public-contract.md`](docs/public-contract.md).
-
-### Modules
-
-- `AUDITA_MODULES` (CSV)
-- CLI: `--modules`
-
-### Transcript Description
-
-CLI:
-- `--transcript-description`
-
-Behavior:
-- optional background context for proposal and LLM-validator prompts;
-- trimmed and length-limited by CLI validation;
-- does not override transcript content;
-- no `AUDITA_*` environment variable is currently defined for this setting.
-
-### Primary LLM
-
-Environment:
-- `AUDITA_LLM_API_KEY` (or `OPENROUTER_API_KEY` fallback)
-- `AUDITA_MODEL`
-- `AUDITA_BASE_URL`
-- `AUDITA_LLM_TIMEOUT_SECONDS`
-- `AUDITA_MAX_RETRIES`
-
-CLI:
-- `--llm-api-key`
-- `--model`
-- `--base-url`
-- `--llm-timeout-seconds`
-- `--max-retries`
-
-### Validation LLM
-
-Environment:
-- `AUDITA_VALIDATION_LLM_API_KEY`
-- `AUDITA_VALIDATION_MODEL`
-- `AUDITA_VALIDATION_BASE_URL`
-- `AUDITA_VALIDATION_LLM_TIMEOUT_SECONDS`
-- `AUDITA_VALIDATION_MAX_RETRIES`
-- `AUDITA_VALIDATION_LLM_CONCURRENCY`
-- `AUDITA_VALIDATION_MAX_PROMPT_TOKENS`
-
-CLI:
-- `--validation-llm-api-key`
-- `--validation-model`
-- `--validation-base-url`
-- `--validation-llm-timeout-seconds`
-- `--validation-max-retries`
-- `--validation-llm-concurrency`
-- `--validation-max-prompt-tokens`
-
-### LLM Concurrency
-
-Environment:
-- `AUDITA_TOTAL_LLM_CONCURRENCY`
-- `AUDITA_PROPOSAL_LLM_CONCURRENCY`
-- `AUDITA_VALIDATION_LLM_CONCURRENCY`
-- `AUDITA_LLM_CONCURRENCY` (legacy alias for `AUDITA_TOTAL_LLM_CONCURRENCY`)
-
-CLI:
-- `--total-llm-concurrency`
-- `--proposal-llm-concurrency`
-- `--validation-llm-concurrency`
-- `--llm-concurrency` (legacy alias for `--total-llm-concurrency`)
-
-Behavior:
-- all proposal and validation LLM calls are bounded by total LLM concurrency
-- proposal LLM calls are additionally bounded by proposal LLM concurrency
-- when validation concurrency is unset, it inherits total LLM concurrency
-- when explicitly set, proposal and validation concurrency must each be `<= total-llm-concurrency`
-- canonical total settings win when both canonical and legacy alias settings are provided at the same precedence layer
-
-### Confidence Thresholds
-
-Environment:
-- `AUDITA_GLOSSARY_CONFIDENCE_THRESHOLD`
-- `AUDITA_HOMOPHONES_CONFIDENCE_THRESHOLD`
-- `AUDITA_SPOKEN_WORD_CONFIDENCE_THRESHOLD`
-- `AUDITA_GRAMMAR_CONFIDENCE_THRESHOLD`
-
-CLI:
-- `--glossary-confidence-threshold`
-- `--homophones-confidence-threshold`
-- `--spoken-word-confidence-threshold`
-- `--grammar-confidence-threshold`
-
-### Normalization and Chunking
-
-Environment:
-- `AUDITA_NORMALIZE_MAX_SEGMENT_GAP`
-- `AUDITA_NORMALIZE_ELLIPSIS_GAP`
-- `AUDITA_NORMALIZE_MAX_SEGMENT_DURATION`
-- `AUDITA_NORMALIZE_MAX_SEGMENT_TOKENS`
-- `AUDITA_MAX_SECTION_TOKENS`
-- `AUDITA_MIN_SECTION_TOKENS`
-- `AUDITA_TARGET_SECTIONS`
-
-CLI:
-- `--normalize-max-segment-gap`
-- `--normalize-ellipsis-gap`
-- `--normalize-max-segment-duration`
-- `--normalize-max-segment-tokens`
-- `--max-section-tokens`
-- `--min-section-tokens`
-- `--target-sections`
-
-### Work Directory
-
-Environment:
-- `AUDITA_WORK_DIR`
-- `AUDITA_WORK_DIR_RETENTION` (`auto`, `always`, `never`)
-
-CLI:
-- `--work-dir`
-- `--work-dir-retention`
-
-Retention behavior:
-- `always`: keep all run directories
-- `never`: keep successful run directories
-- `auto`: keep failed runs and successful runs with skipped/rejected corrections
-
-## Reports and Diagnostics
-
-Per-run diagnostics include:
-- source transcript artifacts
-- normalized transcript artifact
-- normalization summary
-- chunking summary
-- utilization diagnostics summary
-- correction ledger
-- invocation metadata
-- redacted effective config
-- module/validator prompt-response diagnostics
-- `report.json`
-- `error.log` on failure
-
-Optional external report output:
-- `--report-json <path>`
+This makes LLM requests. Explicit setting flags override the selected file and
+optional overlay; general environment setting overrides are no longer read.
+By default the command writes `errors.yml`, `recoveries.yml`, and `report.md`
+beside `corrected.json`. Use `--reporting=false` to suppress those files, or
+`--debug` to retain detailed run artifacts under the configured debug work
+directory. Model failures affecting some operations can still yield a completed
+transcript with applied corrections; the summary and reports identify them.
+See the [v2 migration guide](docs/migrations/v2.md) for old configurations.
 
 ## Documentation
 
-- Architecture: [`docs/architecture.md`](docs/architecture.md)
-- Diagnostics: [`docs/diagnostics.md`](docs/diagnostics.md)
-- Structured LLM adapter: [`docs/structured-llm.md`](docs/structured-llm.md)
-- Subprocess operations: [`docs/subprocess-operations.md`](docs/subprocess-operations.md)
-- Release checklist: [`docs/release-checklist.md`](docs/release-checklist.md)
+- [CLI reference](docs/cli.md)
+- [Configuration and maintained examples](docs/config.md)
+- [Operations and artifact lifecycle](docs/operations.md)
+- [Transcript and glossary inputs](docs/integrations/transcript-glossary-files.md)
+- [Output files and report formats](docs/integrations/output-files.md)
+- [Subprocess integration](docs/integrations/subprocess.md)
+- [LLM endpoint contract](docs/integrations/openai-compatible-llm.md)
+- [Contributor guide](docs/development.md)

cmd/audita/main.go

@@ -1,11 +1,20 @@
 package main
 
 import (
+	"context"
 	"os"
+	"os/signal"
+	"syscall"
 
 	"gitea.maximumdirect.net/eric/audita/internal/cli"
 )
 
 func main() {
-	os.Exit(cli.Run(os.Args[1:], os.Stdout, os.Stderr))
+	os.Exit(run())
+}
+
+func run() int {
+	ctx, stop := signal.NotifyContext(context.Background(), os.Interrupt, syscall.SIGTERM)
+	defer stop()
+	return cli.Run(ctx, os.Args[1:], os.Stdout, os.Stderr)
 }

docs/adr/README.md

@@ -0,0 +1,7 @@
+# Architectural Decision Records
+
+No ADRs have been recorded here. Use the
+[documentation policy's ADR lifecycle](../policy/documentation.md#architecture-decision-records)
+for significant decisions. Current architecture belongs in the
+[architecture policy](../policy/architecture.md); implementation status belongs in
+[roadmap documents](../roadmap/).

docs/architecture/architecture.md

@@ -1,849 +0,0 @@
-# Audita Architecture
-
-## Scope and intent
-This document describes:
-- the architecture used in production today.
-
-Historical rewrite details live in `docs/rewrite-notes.md`.
-
-## Current implementation status
-Implemented today:
-- Go CLI entrypoint and `audita process` wiring.
-- Config defaults, env loading, CLI override precedence, and validation.
-- Transcript and glossary parsing/validation.
-- Deterministic transcript normalization.
-- Deterministic token estimation and transcript chunking.
-- Per-run diagnostics directory creation plus process-level artifacts.
-- Process report JSON output with diagnostics artifact references.
-- Framework foundation packages for contracts and proposal application.
-- Production runner orchestration package with deterministic sequential module execution.
-- Module-level report structures with applied/skipped change records.
-- Runtime validator models and deterministic validators.
-- Deterministic validator-chain execution in the runner with cardinality enforcement.
-- Module-level validator decision/rejection reporting.
-- Internal structured LLM client contract plus an Audita-owned OpenAI-compatible structured LLM adapter package.
-- Bounded FIFO LLM scheduler infrastructure with context-aware permit handling.
-- Runtime primary/validation LLM effective-config resolution helpers with validation inheritance.
-- Generic JSON prompt/response diagnostics writer primitives with secret redaction.
-- LLM-backed validator models, prompt builders, batching, and runtime execution.
-- Runner wiring for LLM validators via the internal structured LLM abstraction and scheduler hooks.
-- LLM validator diagnostics artifacts and report-level decision metadata paths.
-- Shared LLM proposal-generation helper with structured correction-set parsing.
-- Deterministic proposal-index assignment and enriched proposal mapping for shared generation.
-- Proposal-generation diagnostics artifacts with secret redaction.
-- Production module registry with known-key recognition and explicit unsupported-module errors.
-- Production `grammar` module implementation in `internal/modules/grammar`.
-- Production `glossary` module implementation in `internal/modules/glossary`.
-- Production `homophones` module implementation in `internal/modules/homophones`.
-- Production `spoken_word` module implementation in `internal/modules/spoken_word`.
-- Explicit runtime support for `--modules grammar` through the production runner path.
-- Explicit runtime support for `--modules glossary`, including repeated stages such as `--modules glossary,glossary`.
-- Explicit runtime support for `--modules homophones` through the production runner path.
-- Explicit runtime support for `--modules spoken_word` through the production runner path.
-
-Current reality:
-- all production modules exist and are wired into the default runtime path.
-- a normal `audita process` run without `--modules` now executes the full sequence:
-  - `glossary`
-  - `homophones`
-  - `glossary`
-  - `spoken_word`
-  - `grammar`
-- repeated glossary stages are deterministic and reported distinctly as `glossary_1` and `glossary_2`.
-
-## Actual Go package layout
-
-```text
-cmd/audita/
-  main.go
-
-internal/cli/
-  run.go
-
-internal/core/config/
-  config.go
-  env.go
-  flags.go
-  redaction.go
-  validation.go
-
-internal/core/schema/
-  transcript.go
-  glossary.go
-  errors.go
-
-internal/core/io/
-  files.go
-
-internal/core/normalization/
-  normalize.go
-  tokens.go
-
-internal/core/chunking/
-  sections.go
-  summary.go
-  tokens.go
-
-internal/core/diagnostics/
-  run_dir.go
-
-internal/core/reporting/
-  report.go
-
-internal/framework/contracts/
-  contracts.go
-
-internal/framework/proposals/
-  proposal.go
-  policy.go
-  preview.go
-  apply.go
-
-internal/framework/runner/
-  observability.go
-  runner.go
-
-internal/framework/proposal_generation/
-  generate.go
-
-internal/framework/modules/
-  registry.go
-
-internal/modules/grammar/
-  module.go
-  prompt.go
-
-internal/modules/glossary/
-  module.go
-  prompt.go
-
-internal/modules/homophones/
-  module.go
-  prompt.go
-
-internal/modules/spoken_word/
-  module.go
-  prompt.go
-
-internal/framework/validators/
-  models.go
-  deterministic.go
-  llm_models.go
-  llm_prompt_builders.go
-  llm_batching.go
-  llm_validators.go
-
-internal/validators/
-  metadata/
-    metadata.go
-  registry.go
-  chains.go
-  confidence_threshold/
-    validator.go
-  original_text_presence/
-    validator.go
-  non_empty_corrected_text/
-    validator.go
-  no_effect/
-    validator.go
-  protected_terms/
-    validator.go
-  spoken_form_plausibility/
-    validator.go
-  meaning_reversal_review/
-    validator.go
-  editorial_review/
-    validator.go
-  grammar_review/
-    validator.go
-  spoken_word_review/
-    validator.go
-
-internal/prompts/
-  registry.go
-  render.go
-  assets/
-    shared/
-    modules/
-    validators/
-
-internal/framework/llm/
-  openai_compatible_client.go
-  scheduler.go
-  effective_config.go
-  diagnostics.go
-
-internal/framework/responseschema/
-  registry.go
-  registry_test.go
-
-internal/cli/
-  review_artifacts.go
-  parity_test.go
-  release_fixtures_test.go
-  testdata/
-    parity/
-    release/
-```
-
-## Current CLI behavior
-Primary commands:
-
-```sh
-audita process <transcript.json> --glossary <glossary.yaml> [flags]
-audita config validate --config <config.yml>
-audita config print-effective [--config <config.yml>]
-```
-
-Current runtime flow (`internal/cli/run.go`):
-1. Build runtime config from:
-   - defaults;
-   - file config source (`--config`, `AUDITA_CONFIG`, or default search paths when present: `/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml`);
-   - environment overrides;
-   - CLI overrides.
-2. Parse flags and apply CLI overrides.
-3. Validate transcript positional argument and required `--glossary`.
-4. Create per-run diagnostics directory.
-5. Read transcript and glossary files.
-6. Parse/validate transcript and glossary.
-7. Write source transcript artifacts.
-8. Normalize transcript.
-9. Write normalized transcript and normalization summary artifacts.
-10. Chunk normalized transcript and compute chunk summaries.
-11. Write chunking summary artifact.
-12. Execute runner modules sequentially:
-    - default run path uses configured default sequence (`glossary,homophones,glossary,spoken_word,grammar`);
-    - explicit `--modules` overrides the default sequence;
-    - test/injected module factory path remains available for deterministic runtime tests.
-    - each module recomputes chunks from the current working transcript, runs chunk proposal work concurrently, aggregates deterministically, validates, and applies approved proposals once.
-13. Output working transcript to `--output` file or stdout.
-14. Build process report metadata.
-15. Optionally write `--report-json`; always write run-dir `report.json`.
-16. Apply work-dir retention.
-
-Config command behavior (`internal/cli/run.go`):
-- `audita config validate --config <path>`:
-  - loads and validates a versioned YAML config file;
-  - does not require transcript or glossary inputs.
-- `audita config print-effective [--config <path>]`:
-  - builds effective config from defaults + file config + env overrides;
-  - prints redacted JSON to stdout;
-  - does not require transcript or glossary inputs.
-
-Parity fixture status:
-- representative Python-parity fixture coverage exists under `internal/cli/testdata/parity`;
-- parity tests use fake structured LLM responses for deterministic behavior, including default full-pipeline shape assertions;
-- parity comparisons intentionally ignore nondeterministic metadata (timestamps, run IDs, temp paths, token usage) and remain strict for deterministic contract fields (transcript content, module order/instance naming, applied/skipped/rejected counts, and status).
-- intentional Python-vs-Go differences and open parity gaps are documented in `docs/python-parity.md`.
-
-Important behavior details:
-- Glossary is validated and is used for explicit glossary/grammar/homophones/spoken_word module correction paths.
-- Default production CLI behavior now executes the full production module sequence unless `--modules` override is supplied.
-- Explicit `--modules grammar`, `--modules glossary`, `--modules homophones`, and `--modules spoken_word` continue to run production module paths with LLM-backed proposal generation and validator-chain execution.
-- Default runs (without explicit module selection) perform LLM calls through production module and validator paths.
-- Success path is generally quiet on stderr.
-- Source IDs are preserved into a canonical transcript before normalization; normalization then reassigns output IDs sequentially from `1`.
-
-## Implemented data contracts
-
-### Transcript input
-Accepted top-level forms:
-- bare JSON array of segments
-- object with `segments` array
-
-Source segment contract:
-- `id` optional integer
-- `speaker` non-empty string
-- `start` finite non-negative number
-- `end` finite non-negative number with `end >= start`
-- `text` non-empty string
-- `categories` optional array of non-empty strings
-
-Additional checks:
-- duplicate explicit source IDs are rejected.
-
-### Transcript output
-Transcript output is selected through an output schema registry (`internal/core/outputschema`).
-
-Supported output schemas:
-- `bare-segments` (default):
-  - top-level JSON array of normalized segments;
-  - each segment includes `id`, `speaker`, `start`, `end`, `text`, optional `categories`.
-- `audita-v1`:
-  - top-level object with:
-    - `schema: "audita-v1"`
-    - `version: "v1"`
-    - `segments: [...]` (same normalized segment payload).
-
-Current status:
-- `seriatim-intermediate` is not implemented yet; selecting it fails clearly as an unsupported output schema.
-
-Selection behavior:
-- CLI: `--output-schema <name>`
-- file config: `output.schema: <name>`
-- precedence remains runtime-wide defaults -> file config -> env -> CLI.
-
-Both stdout transcript output and `--output` file output use the same selected output encoder.
-
-### Glossary input
-YAML with `glossary` entries. Required fields per entry:
-- `name`, `category`, `summary`
-
-Optional:
-- `aliases`, `plural`
-
-## Implemented config/env/flag behavior
-Precedence for `audita process`:
-1. defaults (`config.Default()`)
-2. config file (if resolved from `--config`, `AUDITA_CONFIG`, or default path)
-3. environment overrides
-4. CLI flags (`ApplyCLIOverrides`)
-
-File-config source behavior:
-- explicit `--config <path>`:
-  - required to exist, otherwise process fails clearly.
-- `AUDITA_CONFIG` (when `--config` is not provided):
-  - required to exist, otherwise process fails clearly.
-- default paths `/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml` (when neither explicit source is provided):
-  - first existing path in that order is used;
-  - both missing is silently ignored.
-
-Versioned file-config behavior (`internal/core/config/file_config.go`):
-- supported version: `version: 1`;
-- missing version fails;
-- unsupported version fails;
-- strict unknown-field rejection is enabled.
-
-`api_key_env` behavior:
-- file config can declare API key environment variable names for proposal/validation LLM settings;
-- runtime resolves those names from the process environment during config application;
-- no direct API-key value field is supported in file config.
-
-Redaction behavior:
-- effective config artifacts and `audita config print-effective` both use the same redaction path (`Config.Redacted()`), so API keys are not emitted in plaintext.
-
-Implemented config surfaces include:
-- module list
-- primary and validation LLM settings
-- total/proposal/validation LLM concurrency controls
-- transcript description context (`--transcript-description`)
-- section token controls and target sections
-- confidence thresholds
-- normalization controls
-- work-dir and retention mode
-
-Current caveat:
-- LLM/module-related settings are active for default and explicit module-run paths.
-- compatibility environment variables and lower-level CLI tuning flags remain available while the preferred config-driven surface is adopted.
-
-Transcript description behavior:
-- `--transcript-description` is a process-flag input for optional user-supplied background context.
-- runtime config stores this value in `Config.TranscriptDescription` after CLI trimming and length validation.
-- default value is empty; empty values produce no prompt context section.
-- this value is intentionally non-secret and appears in effective config and invocation metadata artifacts.
-
-## Implemented transcript description prompt context
-Transcript description context is wired through production prompt paths:
-- proposal prompts for `glossary`, `homophones`, `spoken_word`, and `grammar`;
-- LLM-backed validator prompts for spoken-form plausibility, meaning reversal, editorial review, grammar review, and spoken-word review.
-
-Prompt guardrail semantics are consistent across modules and validators:
-- transcript description is labeled as "background context only";
-- it may help interpret ambiguous terms;
-- it must not override transcript content;
-- the model must not invent corrections, facts, names, events, motivations, or speaker intent from this description.
-
-Generated transcript descriptions remain deferred and are not implemented in the current runtime.
-
-## Implemented embedded prompt assets
-Prompt assets are now built-in embedded Markdown files under `internal/prompts/assets`:
-- `assets/modules/*` for production module proposal prompts;
-- `assets/validators/*` for LLM-backed validator prompts;
-- `assets/shared/prompt_hardening.md` for shared prompt-injection hardening text.
-
-Prompt source behavior:
-- built-in embedded prompts are the only supported source in current runtime;
-- filesystem prompt overrides and prompt-source selection flags are not implemented.
-
-`internal/prompts` registry responsibilities:
-- register stable prompt IDs;
-- register prompt version and source metadata;
-- load embedded assets;
-- compute deterministic SHA-256 prompt source hashes;
-- render system/user prompts with `text/template` using missing-key errors.
-
-Prompt metadata fields:
-- `prompt_id`
-- `prompt_version`
-- `prompt_source` (`builtin`)
-- `embedded_path`
-- `sha256`
-
-Prompt rendering flow:
-- module proposal builders construct typed template data (section JSON, glossary JSON, transcript-description block) and render via `internal/prompts`;
-- validator prompt builders construct typed template data (validation payload JSON, transcript-description block) and render via `internal/prompts`.
-
-Shared prompt hardening:
-- the same centralized hardening fragment is included in every proposal and LLM-validator prompt;
-- hardening text enforces untrusted transcript handling, no instruction-following from transcript content, and no invented facts/corrections.
-
-Prompt metadata diagnostics flow:
-- proposal-generation diagnostics request metadata includes prompt metadata;
-- LLM-validator diagnostics request metadata includes prompt metadata;
-- detailed prompt metadata is diagnostics-scoped today and is not yet expanded into broad report-level prompt registries.
-
-## Implemented structured LLM infrastructure
-`internal/framework/contracts` now defines a typed structured-completion contract:
-- `StructuredLLMClient.CompleteStructured(ctx, req, out)`
-- caller-owned typed decode target via `out` pointer.
-- caller-selected structured response schema metadata via `StructuredCompletionRequest.ResponseSchema`.
-
-`internal/framework/llm` provides `OpenAICompatibleClient`, a direct `net/http` adapter over OpenAI-compatible chat completions:
-- configurable `base_url`, model, optional API key, retries, HTTP client, and request timeout;
-- OpenAI-compatible endpoint behavior (for example OpenAI/OpenRouter/local-compatible base URLs);
-- request message translation from `contracts.LLMMessage` to chat-completions messages;
-- strict `response_format.type = json_schema` with registered structured response schemas (`strict: true`, schema name, and schema body);
-- response metadata mapping (provider/model/token usage) into Audita-owned response types;
-- API-key redaction in adapter-returned errors;
-- context cancellation and timeout propagation through request contexts and HTTP client timeouts;
-- bounded retry behavior for transient request failures and malformed retryable structured responses.
-
-Structured response schemas are owned by Audita in `internal/framework/responseschema` and currently include:
-- key `correction_set`:
-  - id `audita.correction_set`
-  - version `v1`
-  - name `audita_correction_set_v1`
-  - sha256 `05f8ff3fa04f68115c0cb1859d2656f51aa5c0bae8ff2470b2d4f6f531953195`
-- key `validator_decision_set`:
-  - id `audita.validator_decision_set`
-  - version `v1`
-  - name `audita_validator_decision_set_v1`
-  - sha256 `b73f4790b98fbb955f0aec5496dd8ce9a8fe14aa2f35c700b4b4e5634f106fd5`
-
-Provider-level structured output is treated as a guardrail, not a trust boundary:
-- the adapter decodes assistant message content into caller-owned structs;
-- proposal-generation and validator layers continue local validation (shape, cardinality, confidence bounds, and proposal-index semantics) before changes can be applied.
-
-Current runtime boundary:
-- the default CLI runtime path (without explicit module selection) instantiates the full production module sequence.
-- LLM calls are exercised in production in both default full-pipeline runs and explicit `--modules` runs, and in tests when fake/injected clients are used.
-- normal `go test ./...` does not require real LLM credentials or Python dependencies.
-
-`internal/framework/llm` also provides:
-- a bounded FIFO `Scheduler` for controlled concurrent LLM calls with reliable permit release on success, error, and cancellation;
-- primary/validation effective-config resolution helpers, including validation inheritance fallback to total LLM concurrency settings;
-- generic interaction diagnostics primitives that write machine-readable JSON artifacts for request metadata, request payload, response payload, and optional error payload with secret redaction.
-
-Structured LLM diagnostics behavior:
-- proposal-generation and validator diagnostics include structured response schema metadata (`id`, `version`, `name`, `sha256`) when schema-driven calls are made;
-- API keys and bearer tokens are redacted from request/response/error diagnostics artifacts and surfaced errors.
-
-Dependency posture:
-- the runtime no longer depends on `instructor-go`;
-- structured LLM behavior is implemented through Audita-owned code paths behind `StructuredLLMClient`.
-
-LLM concurrency runtime behavior:
-- `total` concurrency bounds all proposal and validation LLM calls.
-- `proposal` concurrency adds a proposal-only sub-cap, composed with total.
-- `validation` concurrency adds a validation-only sub-cap, composed with total.
-- legacy `llm-concurrency` inputs remain compatibility aliases for total concurrency.
-- modules execute serially, chunk proposals run concurrently within each module, and approved proposals are applied once per module in deterministic order.
-
-## Implemented normalization behavior
-Normalization (`internal/core/normalization`) currently:
-- sorts by segment start time;
-- merges adjacent same-speaker segments when constraints pass;
-- uses gap-based joiners:
-  - gap `< ellipsis_gap` -> single space join
-  - gap `>= ellipsis_gap` -> `... ` join
-- enforces merged duration and token-limit constraints;
-- reassigns output IDs sequentially from `1`;
-- returns `NormalizationSummary` with merge and skip counters.
-
-Note: merged categories are concatenated (not deduplicated).
-
-## Implemented chunking behavior
-Chunking (`internal/core/chunking`) currently provides:
-- deterministic heuristic token estimation;
-- contiguous sectioning with section metadata;
-- max/min section token validation;
-- optional `target_sections` override for section-count planning;
-- summary and detailed summary generation.
-
-Current behavior details:
-- if a single segment exceeds max tokens, it is emitted as its own section (not hard-failed);
-- default section count is planned from `ceil(total_tokens / max_section_tokens)`;
-- section sizing targets `ceil(total_tokens / section_count)` with a deterministic forward pass;
-- sections remain contiguous and ordered, and segments are never split.
-
-## Implemented proposal/replacement infrastructure
-`internal/framework/proposals` provides deterministic proposal composition logic:
-- `CorrectionProposal` and `EnrichedCorrectionProposal` models;
-- replacement policies: `require_unique`, `replace_all`;
-- safe preview (`PreviewProposalForSegment`) with stable skip reasons;
-- deterministic apply (`ApplyProposals`) in ascending `proposal_index` order;
-- applied/skipped change records suitable for reporting.
-
-`internal/framework/contracts` provides interfaces and run-spec metadata scaffolding, including deterministic repeated module instance naming (`ResolveModuleRunSpecs`).
-
-These primitives are wired into the production runner and report model. The grammar, glossary, homophones, and spoken_word modules are implemented.
-
-## Implemented validator runtime infrastructure
-`internal/framework/validators` provides deterministic validator infrastructure:
-- runtime validation request/result models;
-- stable validator reason codes;
-- cardinality enforcement for validator decisions:
-  - missing proposal indexes fail
-  - duplicate proposal indexes fail
-  - unknown proposal indexes fail
-- deterministic validators:
-  - confidence threshold by module key/config threshold
-  - original-text presence against current working transcript
-  - non-empty corrected text
-  - identical/no-effect rejection
-  - conservative protected glossary-term guard for non-glossary modules
-
-`internal/framework/runner` executes module pipelines with deterministic boundaries:
-- modules still execute serially over the working transcript;
-- section proposal work is launched promptly and can run concurrently;
-- section-level validator-chain work starts as section proposals become available (deterministic validators before LLM-backed validators);
-- proposal-generation and LLM-validator calls can overlap under composed scheduler limits;
-- approved proposals are still applied once per module after section work settles.
-
-Validator rejections are reported distinctly from proposal-application skips.
-
-Validator composition is now explicit and registry-backed through `internal/validators`:
-- built-in validator registry with stable keys and lookup/build failure for unknown keys;
-- built-in chain definitions per production module key;
-- production modules resolve validator chains from those built-in definitions.
-
-Package ownership boundary:
-- `internal/validators/<validator_key>` owns built-in validator construction and stable key identity.
-- `internal/framework/validators` remains shared runtime machinery:
-  - request/result models;
-  - decision cardinality enforcement;
-  - protected-vocabulary helpers;
-  - generic LLM-backed validator runtime, batching, and diagnostics glue.
-
-Validator execution classification metadata:
-- `internal/validators/metadata` defines execution class markers:
-  - `deterministic`
-  - `llm_backed`
-- runner ordering uses this metadata interface rather than concrete framework validator type assertions.
-- validators without classification metadata default to deterministic ordering.
-
-`protected_terms` construction ownership:
-- `internal/validators/protected_terms.New()` builds the general (non-glossary-stage) variant.
-- `internal/validators/protected_terms.NewGlossaryStage()` builds the glossary-stage variant used by glossary chains.
-- both variants preserve the stable key `protected_terms`.
-
-Stable built-in validator keys:
-- deterministic:
-  - `confidence_threshold`
-  - `original_text_presence`
-  - `non_empty_corrected_text`
-  - `no_effect`
-  - `protected_terms`
-- LLM-backed:
-  - `spoken_form_plausibility`
-  - `meaning_reversal_review`
-  - `editorial_review`
-  - `grammar_review`
-  - `spoken_word_review`
-
-Built-in module chains:
-- `glossary`:
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `spoken_form_plausibility`
-  - `meaning_reversal_review`
-- `homophones`:
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `spoken_form_plausibility`
-  - `meaning_reversal_review`
-- `spoken_word`:
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `spoken_word_review`
-  - `meaning_reversal_review`
-- `grammar`:
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `grammar_review`
-  - `meaning_reversal_review`
-
-1.0 boundary:
-- validator chains are built-in and not user-configurable from config/CLI.
-- existing threshold and batching knobs remain configurable.
-
-## Implemented LLM-backed validator infrastructure
-`internal/framework/validators` now includes LLM-backed validator support:
-- typed request/response models for structured LLM validation;
-- prompt builders for:
-  - spoken-form plausibility
-  - meaning reversal detection
-  - editorial review
-  - grammar review
-  - spoken-word review
-- deterministic batching by `validation_max_prompt_tokens`;
-- strict cardinality validation of structured LLM decisions (missing/duplicate/unknown indexes fail);
-- safe failure behavior for malformed/invalid structured responses.
-
-`internal/framework/runner` wires LLM validators into existing validator chains using:
-- the internal structured LLM client abstraction (`contracts.StructuredLLMClient`);
-- bounded scheduler hooks for validator call execution;
-- diagnostics writer hooks for machine-readable prompt/response artifacts with secret redaction.
-
-## Implemented shared proposal-generation infrastructure
-`internal/framework/proposal_generation` provides a reusable, prompt-agnostic helper for future real modules:
-- structured request model including module key/instance, replacement policy, working transcript context, optional section metadata, glossary, config, and diagnostics context;
-- structured correction-set response model (`corrections`) mapped into existing `proposals.CorrectionProposal` and `proposals.EnrichedCorrectionProposal` models;
-- deterministic proposal-index assignment through a caller-provided `start_index`;
-- structured LLM calls through `contracts.StructuredLLMClient` only (no direct provider calls);
-- optional bounded execution through scheduler hooks (`contracts.LLMScheduler`);
-- prompt/response diagnostics artifact writing via the generic `internal/framework/llm` diagnostics primitives with redaction of API keys/secrets.
-
-This helper only produces candidate proposals; validator-chain execution and proposal application remain runner responsibilities.
-
-## Implemented production module-registry scaffolding
-`internal/framework/modules` now provides a production registry scaffold:
-- recognizes intended module keys:
-  - `glossary`
-  - `homophones`
-  - `spoken_word`
-  - `grammar`
-- supports explicit constructor registration with dependency injection for:
-  - run spec
-  - config
-  - glossary
-  - proposal/validation structured LLM clients
-  - proposal/validation schedulers
-  - diagnostics directory context
-- returns explicit errors for unknown keys (`unsupported_module`).
-
-The `grammar`, `glossary`, `homophones`, and `spoken_word` module keys are now registered and constructible.
-
-## Implemented grammar production module
-`internal/modules/grammar` now provides the first production module:
-- prompt builder faithfully constrained to punctuation/capitalization/spacing/article cleanup;
-- explicit guardrails against meaning-changing rewrites, style rewrites, summarization, and invention;
-- proposal generation through `internal/framework/proposal_generation` and `contracts.StructuredLLMClient`;
-- scheduler-aware proposal calls through existing `contracts.LLMScheduler` hooks;
-- replacement policy `require_unique` (current runtime policy);
-- validator chain integration using existing deterministic + LLM-backed validators;
-- grammar confidence threshold enforcement through existing validator/config infrastructure;
-- module-level reporting and diagnostics capture through existing runner/reporting paths.
-
-## Implemented glossary production module
-`internal/modules/glossary` now provides the second production module:
-- prompt builder aligned to Python glossary-module intent, constrained to glossary-backed domain/acoustic corrections;
-- prompt context includes glossary names, aliases, categories, summaries, and plural forms where available;
-- guardrails against broad style rewriting and against replacing unrelated terms simply because they appear in glossary entries;
-- proposal generation through `internal/framework/proposal_generation` and `contracts.StructuredLLMClient`;
-- scheduler-aware proposal calls through existing `contracts.LLMScheduler` hooks;
-- replacement policy `replace_all` (matching Python glossary behavior);
-- validator chain integration using existing deterministic + LLM-backed validators;
-- glossary confidence threshold enforcement through existing validator/config infrastructure;
-- module-level reporting and diagnostics capture through existing runner/reporting paths;
-- explicit support for repeated glossary stages with deterministic instance names (`glossary_1`, `glossary_2`, ...), where later stages see prior-stage working transcript changes.
-
-## Implemented protected-term behavior
-`internal/framework/validators/protected_terms.go` provides deterministic glossary-derived protected vocabulary:
-- extracts protected terms from glossary names and aliases;
-- includes explicit plural fields and synthetic plural forms where safe;
-- deduplicates and returns stable ordering for repeatable behavior/tests.
-
-This vocabulary is used by deterministic validators for both glossary-stage and non-glossary-stage protection checks, keeping protected-term guardrails active across modules.
-
-## Implemented homophones production module
-`internal/modules/homophones` now provides the third production module:
-- prompt builder aligned to Python homophones-module intent, constrained to conservative homophone/near-homophone/mistranscription corrections;
-- prompt context includes protected glossary names/aliases/plurals to avoid damaging known terms;
-- explicit guardrails against punctuation cleanup, grammar cleanup, style rewriting, summarization, and content invention;
-- proposal generation through `internal/framework/proposal_generation` and `contracts.StructuredLLMClient`;
-- scheduler-aware proposal calls through existing `contracts.LLMScheduler` hooks;
-- replacement policy `require_unique` (matching Python homophones behavior);
-- validator chain integration using existing deterministic + LLM-backed validators;
-- homophones confidence threshold enforcement through existing validator/config infrastructure;
-- protected-term guardrails for non-glossary modules remain active and are exercised through the homophones path;
-- module-level reporting and diagnostics capture through existing runner/reporting paths.
-
-## Implemented spoken_word production module
-`internal/modules/spoken_word` now provides the fourth production module:
-- prompt builder aligned to Python spoken_word-module intent, constrained to conservative dysfluency cleanup;
-- strong prompt guardrails preserving meaning/intent/voice/named entities/domain terms and substantive content;
-- explicit guardrails against summarization, style rewriting, grammar-only cleanup, punctuation-only cleanup, invention, and meaning-changing rewrites;
-- proposal generation through `internal/framework/proposal_generation` and `contracts.StructuredLLMClient`;
-- scheduler-aware proposal calls through existing `contracts.LLMScheduler` hooks;
-- replacement policy `require_unique` (matching Python spoken_word behavior);
-- validator chain integration using existing deterministic + LLM-backed validators, including strong semantic guardrails (`spoken_word_review`, `meaning_reversal_review`);
-- spoken_word confidence threshold enforcement through existing validator/config infrastructure;
-- protected-term guardrails for non-glossary modules remain active and are exercised through the spoken_word path;
-- module-level reporting and diagnostics capture through existing runner/reporting paths.
-
-## Reports and diagnostics (implemented)
-Current per-run artifacts include:
-- `source-transcript.json`
-- `source-transcript-parsed.json`
-- `normalized-transcript.json`
-- `normalization-summary.json`
-- `chunking-summary.json`
-- `utilization-diagnostics.json`
-- `correction-ledger.json`
-- `invocation.json`
-- `effective-config.json` (redacted credentials)
-- `report.json`
-- `error.log` on failure
-
-`--report-json` writes a separate report file when requested.
-
-Current process reports include diagnostics metadata references for:
-- diagnostics directory path;
-- source transcript artifact path;
-- parsed source transcript artifact path;
-- normalized transcript artifact path;
-- normalization summary artifact path;
-- chunking summary artifact path;
-- utilization diagnostics artifact path;
-- correction ledger artifact path;
-- invocation metadata artifact path;
-- redacted effective-config artifact path;
-- error-log artifact path on failure.
-
-Current process reports also include:
-- module-level results (when runner modules execute), including applied/skipped proposal changes;
-- run-level module summary totals and failed module instance metadata.
-- module-level validator decisions and validator rejections.
-- optional decision-level diagnostic artifact paths for validator LLM interactions when available.
-- stable validator keys in `validator_name` fields for validator decisions/rejections.
-- explicit report metadata:
-  - report schema name;
-  - report schema version;
-  - selected output schema;
-  - config file version when config file input is used.
-- review/observability artifacts:
-  - run-level and module-level utilization/timing summaries;
-  - flattened correction ledger entries for applied/rejected/skipped/failed correction dispositions.
-
-Utilization diagnostics collection:
-- collection is performed in the runner path via lightweight instrumentation around LLM scheduler and structured-client execution (`internal/framework/runner`);
-- instrumentation is observational only and does not change scheduler acquisition/release semantics or module execution order;
-- serialized artifact: `utilization-diagnostics.json`.
-
-Utilization diagnostics high-level shape:
-- `effective_concurrency`:
-  - `total_llm`, `proposal_llm`, `validation_llm`;
-- `run_timing`:
-  - `run_wall_time_ms`;
-  - `scheduler_queue_wait_ms`;
-  - `llm_execution_time_ms`;
-  - `deterministic_validation_time_ms`;
-  - `max_in_flight_llm_calls`;
-  - `average_in_flight_llm_calls`;
-- `llm_calls`:
-  - `total_proposal_calls`;
-  - `total_validation_calls`;
-- `modules`:
-  - per-module key/instance timing summaries including module wall time and per-module call counts;
-- `validators`:
-  - per-validator summaries keyed by stable validator key with elapsed time and LLM-backed marker.
-
-Correction ledger construction:
-- ledger entries are built from runner module results in the CLI report/diagnostics path (`internal/cli/review_artifacts.go`);
-- serialized artifact: `correction-ledger.json`;
-- one flattened record per applied/validator-rejected/application-skipped outcome where data is available, plus module-failed records for failed module instances.
-
-Correction ledger high-level shape:
-- run/module/proposal identity:
-  - `run_id`, `module_key`, `module_instance`, `proposal_index`, `segment_id`;
-- correction payload:
-  - `original_text`, `proposed_corrected_text`, `applied_corrected_text` (when applied), `replacement_policy`;
-- disposition:
-  - `disposition` in `{applied,rejected,skipped,failed}`;
-  - `disposition_reason_code`, `disposition_message`;
-- validator decision snapshots:
-  - `deterministic_validator_decisions[]`;
-  - `llm_validator_decisions[]`;
-  - each decision uses stable validator keys and reason codes.
-
-Identity and metadata boundaries:
-- stable module keys/instance names and stable validator keys are included directly in ledger records;
-- prompt metadata and structured response schema metadata remain in LLM interaction diagnostics payloads and are not duplicated into every ledger row;
-- reports reference artifact paths for utilization and ledger files through diagnostics metadata.
-
-Redaction and retention:
-- secret redaction guarantees continue to apply to diagnostics/report artifacts;
-- utilization and ledger artifacts are emitted within the existing run-directory retention model (`auto|always|never`) and are retained/removed with the run directory.
-
-Current report schema metadata values:
-- `report_metadata.report_schema_name = "audita-process-report"`
-- `report_metadata.report_schema_version = "v1"`
-
-Retention modes implemented in `ApplyRetention`:
-- `always`: keep all run directories.
-- `never`: keep successful run directories.
-- `auto`: keep failed runs and successful runs with skipped corrections.
-- failed runs are always retained.
-
-Current runtime note:
-- default non-explicit runs usually have no module-level skipped corrections, so `auto` commonly removes clean successful run directories.
-- explicit grammar/glossary/homophones/spoken_word runs can produce validator rejections and application skips, which are reflected in reports and retention input.
-
-## Current tests and quality posture
-Implemented tests currently cover:
-- CLI argument handling and behavior (`internal/cli/run_test.go`)
-- subprocess stdout/stderr and exit-code behavior (`cmd/audita/main_integration_test.go`)
-- config/env/override validation (`internal/core/config/*_test.go`)
-- transcript and glossary schema validation (`internal/core/schema/*_test.go`)
-- deterministic normalization (`internal/core/normalization/*_test.go`)
-- deterministic chunking and summaries (`internal/core/chunking/*_test.go`)
-- proposal preview/apply semantics (`internal/framework/proposals/*_test.go`)
-- contracts/foundation composition tests (`internal/framework/contracts/*_test.go`)
-- runner sequencing and failure behavior with deterministic fake modules (`internal/framework/runner/*_test.go`)
-- CLI runner integration through injected fake module factories (`internal/cli/run_test.go`)
-- validator models, cardinality enforcement, and deterministic validators (`internal/framework/validators/*_test.go`)
-- LLM-backed validator batching, prompt builders, structured-response safety, scheduler hooks, and diagnostics redaction (`internal/framework/validators/*_test.go`, `internal/framework/runner/*_test.go`)
-- shared proposal-generation request/response parsing, deterministic indexing, scheduler hooks, and diagnostics redaction (`internal/framework/proposal_generation/*_test.go`, `internal/framework/runner/*_test.go`)
-- production module-registry known-key recognition and unsupported/internal-registry error behavior (`internal/framework/modules/*_test.go`, `internal/cli/run_test.go`)
-- production grammar module prompt constraints, proposal mapping, validator-chain behavior, confidence-threshold enforcement, diagnostics redaction, and explicit CLI/runtime integration (`internal/modules/grammar/*_test.go`, `internal/cli/run_test.go`, `internal/framework/runner/*_test.go`)
-- production glossary module prompt constraints, proposal mapping, validator-chain behavior, confidence-threshold enforcement, diagnostics redaction, repeated-stage behavior, and explicit CLI/runtime integration (`internal/modules/glossary/*_test.go`, `internal/cli/run_test.go`, `internal/framework/runner/*_test.go`)
-- production homophones module prompt constraints, proposal mapping, validator-chain behavior, confidence-threshold enforcement, diagnostics redaction, protected-term behavior, and explicit CLI/runtime integration (`internal/modules/homophones/*_test.go`, `internal/cli/run_test.go`, `internal/framework/runner/*_test.go`)
-- production spoken_word module prompt constraints, proposal mapping, validator-chain behavior, semantic guardrail behavior, confidence-threshold enforcement, diagnostics redaction, protected-term behavior, and explicit CLI/runtime integration (`internal/modules/spoken_word/*_test.go`, `internal/cli/run_test.go`, `internal/framework/runner/*_test.go`)
-- glossary-derived protected-term extraction and stable behavior (`internal/framework/validators/protected_terms_test.go`)
-- default full-pipeline runtime shape and ordering (`internal/cli/run_test.go`, `cmd/audita/main_integration_test.go`, `internal/cli/parity_test.go`)
-- subprocess operational hardening behavior including large-input, failure-mode, timeout/cancellation, backend-failure, and partial-progress paths (`cmd/audita/main_integration_test.go`)
-- report/diagnostics redaction and artifact-shape behavior across success and failure paths (`internal/cli/run_test.go`, `cmd/audita/main_integration_test.go`)
-- curated release-fixture and idempotence-oriented readiness checks using fake structured LLM responses (`internal/cli/release_fixtures_test.go`, `internal/cli/testdata/release`)
-
-## Operational hardening status
-The runtime now includes hardened subprocess behavior for parent-process callers:
-- deterministic success/failure exit codes;
-- strict stdout/stderr separation suitable for machine orchestration;
-- failure stderr summaries that include diagnostics location when available;
-- retained failure diagnostics (`report.json`, `error.log`, and artifacts written before failure);
-- deterministic timeout/cancellation behavior in tests;
-- redaction coverage for API keys/secrets across reports, diagnostics artifacts, and surfaced errors.
-- stable output routing behavior:
-  - with `--output`, stdout remains empty on success;
-  - without `--output`, stdout contains only transcript JSON in the selected output schema;
-  - `--report-json` writes report data to file only (never stdout).
-
-Operational caller guidance is documented in [`docs/subprocess-operations.md`](docs/subprocess-operations.md).
-
-## Final status
-- Audita's default full module-sequence runtime is implemented and tested.
-- Parity fixtures and operational hardening coverage are in place.
-- Historical migration context is documented in [`docs/migration-from-python.md`](docs/migration-from-python.md).

docs/architecture/diagnostics.md

@@ -1,98 +0,0 @@
-# Audita Diagnostics
-
-This document describes the run-directory diagnostics artifacts produced by `audita process`.
-
-## Purpose
-
-Diagnostics provide machine-readable run context and execution artifacts for:
-- failure debugging;
-- validator/correction review;
-- post-run performance analysis.
-
-Diagnostics are written under the configured work directory (`--work-dir`) when run-directory initialization succeeds.
-
-## Core artifacts
-
-Typical artifacts in each run directory:
-- `source-transcript.json`
-- `source-transcript-parsed.json`
-- `normalized-transcript.json`
-- `normalization-summary.json`
-- `chunking-summary.json`
-- `invocation.json`
-- `effective-config.json` (redacted)
-- module/validator LLM interaction artifacts
-- `report.json`
-- `error.log` on failure
-
-## Utilization diagnostics artifact
-
-Artifact:
-- `utilization-diagnostics.json`
-
-High-level fields:
-- `effective_concurrency`:
-  - total/proposal/validation LLM concurrency limits in effect.
-- `run_timing`:
-  - run wall time;
-  - scheduler queue wait time;
-  - LLM execution time;
-  - deterministic validator time;
-  - max/average in-flight LLM calls.
-- `llm_calls`:
-  - total proposal and validation LLM call counts.
-- `modules`:
-  - module-level timing summaries.
-- `validators`:
-  - per-validator timing summaries keyed by stable validator key.
-
-## Correction ledger artifact
-
-Artifact:
-- `correction-ledger.json`
-
-Ledger records are flattened review entries derived from module results and include:
-- module/proposal identity (`module_key`, `module_instance`, `proposal_index`, `segment_id`);
-- correction text fields and replacement policy when available;
-- disposition:
-  - `applied`
-  - `rejected`
-  - `skipped`
-  - `failed`
-- stable reason codes/messages;
-- deterministic and LLM validator decision snapshots using stable validator keys.
-
-Validator rejection and proposal-application skip are distinct dispositions.
-
-## Report references
-
-`report.json` and optional `--report-json` output include diagnostics metadata paths for:
-- utilization diagnostics artifact;
-- correction ledger artifact;
-- existing transcript/normalization/chunking/invocation/effective-config artifacts.
-
-## Retention behavior
-
-Run-directory retention follows configured policy:
-- `always`: keep all run directories;
-- `never`: keep successful run directories;
-- `auto`: keep failed runs and successful runs with skipped/rejected corrections.
-
-## Redaction guarantees
-
-API keys and other configured secrets are redacted from:
-- `effective-config.json`;
-- LLM interaction diagnostics artifacts;
-- reports and surfaced errors.
-
-## Debugging guide
-
-When debugging:
-- slow runs:
-  - inspect `utilization-diagnostics.json` (`run_timing`, `modules`, `validators`, in-flight metrics).
-- validator rejections:
-  - inspect `correction-ledger.json` rejected entries and matching validator decisions;
-  - inspect validator response diagnostics payloads.
-- application skips:
-  - inspect `correction-ledger.json` skipped entries and skip reason codes;
-  - compare with validator decisions to distinguish validation rejection vs apply-time skip.

docs/architecture/output-schemas.md

@@ -1,88 +0,0 @@
-# Audita Output Schemas
-
-This document describes the built-in transcript output schema registry used by `audita process`.
-
-## Supported schema names
-
-### `bare-segments`
-
-Status:
-- implemented
-- default output schema
-
-Shape:
-- top-level JSON array of transcript segments
-
-Segment fields:
-- `id`
-- `speaker`
-- `start`
-- `end`
-- `text`
-- optional `categories`
-
-Compatibility:
-- this preserves the long-standing output shape used by existing consumers.
-
-### `audita-v1`
-
-Status:
-- implemented
-
-Shape:
-- top-level JSON object:
-  - `schema`: `"audita-v1"`
-  - `version`: `"v1"`
-  - `segments`: transcript segment array
-
-Segment fields inside `segments` match `bare-segments` segment fields.
-
-Compatibility:
-- this is the Audita-native object format with explicit schema/version metadata.
-
-### `seriatim-intermediate`
-
-Status:
-- deferred / not implemented
-
-Current behavior:
-- selecting `seriatim-intermediate` fails clearly as an unsupported output schema.
-
-Reason:
-- a concrete, repository-backed contract for this schema has not been finalized yet.
-
-## Selection
-
-Choose output schema with CLI:
-
-```sh
-audita process <transcript.json> --glossary <glossary.yaml> --output-schema audita-v1
-```
-
-Or in file config:
-
-```yaml
-version: 1
-output:
-  schema: audita-v1
-```
-
-Precedence remains:
-1. defaults
-2. file config
-3. environment overrides
-4. CLI overrides
-
-`--output-schema` overrides `output.schema` when both are supplied.
-
-## Output routing behavior
-
-- With `--output`, transcript JSON is written to file using the selected schema and stdout stays empty on success.
-- Without `--output`, stdout contains transcript JSON only, using the selected schema.
-- `--report-json` writes report JSON to file and does not write report payloads to stdout.
-
-## Backward-compatibility expectations
-
-- default schema stays `bare-segments` for compatibility unless explicitly changed in a future breaking release;
-- supported schema names are treated as stable public contract values;
-- unsupported schema names fail before output write.

docs/architecture/prompts.md

@@ -1,118 +0,0 @@
-# Audita Prompts
-
-This document describes Audita's built-in embedded prompt assets and prompt registry behavior.
-
-## Why embedded prompt assets
-
-Audita embeds production prompt text into the binary so runtime behavior is:
-- deterministic;
-- auditable;
-- dependency-light;
-- not dependent on external prompt files at execution time.
-
-Prompt text is authored as Markdown assets and rendered by Go code using typed template data.
-
-## Built-in prompt registry
-
-The prompt registry lives in `internal/prompts` and is responsible for:
-- loading embedded prompt assets;
-- registering stable prompt IDs and versions;
-- recording prompt source metadata;
-- computing deterministic SHA-256 source hashes;
-- rendering system/user prompts with strict missing-key failures.
-
-Current prompt source behavior:
-- built-in embedded prompts only (`prompt_source = builtin`).
-- filesystem prompt overrides are not supported.
-
-## Built-in prompt IDs
-
-Module proposal prompts:
-- `modules.glossary.proposal`
-- `modules.homophones.proposal`
-- `modules.spoken_word.proposal`
-- `modules.grammar.proposal`
-
-LLM-backed validator prompts:
-- `validators.spoken_form_plausibility`
-- `validators.meaning_reversal_review`
-- `validators.editorial_review`
-- `validators.grammar_review`
-- `validators.spoken_word_review`
-
-## Prompt version semantics
-
-Current built-in prompt version value is `v1`.
-
-Version is a stable metadata identifier for diagnostics and debugging. It is not a dynamic prompt-selection mechanism.
-
-## Prompt hash semantics
-
-Each registered prompt includes a deterministic SHA-256 hash of embedded source text.
-
-Hash purpose:
-- identify exact prompt source used in a run;
-- support diagnostics reproducibility and change auditing.
-
-Current hash scope:
-- source prompt text (system + user assets for a registered prompt), not a runtime secret-bearing payload.
-
-## Template rendering behavior
-
-Prompt rendering uses Go `text/template` with typed template data from module/validator builders.
-
-Missing-key behavior:
-- rendering uses missing-key errors;
-- missing/renamed template fields fail quickly instead of silently producing incomplete prompts.
-
-Go code still owns:
-- structured request/response models;
-- response schema selection;
-- transcript/glossary/payload formatting;
-- module and validator selection;
-- diagnostics wiring.
-
-## Shared prompt hardening policy
-
-A shared hardening fragment is embedded once and included in every module proposal prompt and every LLM-validator prompt.
-
-Hardening policy includes:
-- transcript text is untrusted data;
-- glossary entries and transcript descriptions are reference data, not instructions;
-- instructions found inside transcript text must not be obeyed;
-- model must perform only the requested correction/validation task;
-- no invention of facts, names, events, motivations, speaker intent, or corrections;
-- transcript remains the source of truth.
-
-## Transcript description behavior
-
-Transcript description remains background-only prompt context:
-- it may help interpret ambiguous terms;
-- it is explicitly non-authoritative and must not override transcript content;
-- empty descriptions do not render awkward blank context sections.
-
-Generated transcript descriptions are not implemented in this workstream.
-
-## Diagnostics and report metadata boundaries
-
-Current metadata flow:
-- proposal-generation diagnostics request metadata includes prompt metadata;
-- LLM-validator diagnostics request metadata includes prompt metadata.
-
-Prompt metadata fields used in diagnostics:
-- `prompt_id`
-- `prompt_version`
-- `prompt_source`
-- `embedded_path`
-- `sha256`
-
-Current boundary:
-- detailed prompt metadata is diagnostics-first;
-- broad report-level prompt registries/ledgers are deferred.
-
-## 1.0 boundary
-
-Not implemented for 1.0 in this workstream:
-- filesystem prompt overrides;
-- user-configurable prompt selection;
-- external prompt directories.

docs/architecture/public-contract.md

@@ -1,165 +0,0 @@
-# Audita Public Contract
-
-This document defines stability expectations for Audita's external process and data interfaces.
-
-## Scope
-
-This contract covers:
-- CLI invocation and behavior
-- versioned config file behavior
-- transcript/glossary input forms
-- transcript output schema selection
-- process report schema metadata
-- stable validator key identifiers in report/diagnostics records
-- prompt metadata identifiers in diagnostics
-- diagnostics directory behavior
-- utilization diagnostics and correction-ledger artifact presence/pathing in diagnostics metadata
-- stdout/stderr and exit-code behavior
-- secret redaction guarantees
-- compatibility and deprecation policy
-
-## CLI stability expectations
-
-Stable commands:
-- `audita process`
-- `audita config validate`
-- `audita config print-effective`
-
-For `audita process`, stable high-value flags include:
-- `--config`
-- `--glossary`
-- `--output`
-- `--report-json`
-- `--modules`
-- `--output-schema`
-
-Compatibility flags and lower-level tuning flags remain available; they may be narrowed over time with explicit compatibility notes.
-
-## Config file stability expectations
-
-Supported file format:
-- YAML
-- strict unknown-field rejection
-- explicit `version`
-
-Supported version:
-- `version: 1`
-
-Precedence for `audita process`:
-1. built-in defaults
-2. file config
-3. environment overrides
-4. CLI overrides
-
-Config source behavior:
-- `--config <path>`: missing path is a clear failure
-- `AUDITA_CONFIG`: missing path is a clear failure
-- defaults `/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml`: both missing is non-fatal
-
-## Supported transcript input forms
-
-Audita accepts transcript JSON as either:
-- a top-level array of segments
-- an object with a `segments` array
-
-Segments must satisfy the schema and validation rules enforced by `internal/core/schema`.
-
-## Supported glossary input form
-
-Audita accepts glossary YAML with a top-level `glossary` entry list and validates required fields per entry.
-
-## Supported output schema names
-
-Built-in output schema registry supports:
-- `bare-segments` (default)
-- `audita-v1`
-
-`seriatim-intermediate` is planned but not implemented.
-
-Unknown output schema names fail clearly.
-
-## Report schema/versioning expectations
-
-Process report payloads include `report_metadata` with:
-- `report_schema_name`
-- `report_schema_version`
-- `output_schema`
-- `config_version` when file config is used
-
-Current values:
-- `report_schema_name`: `audita-process-report`
-- `report_schema_version`: `v1`
-
-`--report-json` output and diagnostics run-dir `report.json` use the same report schema metadata.
-
-Validator decision/rejection records in reports use stable validator keys in `validator_name`.
-Report diagnostics metadata includes artifact-path fields for utilization diagnostics and correction ledger when diagnostics initialization succeeds.
-
-## Diagnostics directory behavior
-
-When diagnostics directory creation succeeds, Audita writes run artifacts including:
-- invocation metadata
-- redacted effective config
-- transcript/normalization/chunking artifacts
-- utilization diagnostics (`utilization-diagnostics.json`)
-- correction ledger (`correction-ledger.json`)
-- report and failure error log (when applicable)
-- module/LLM diagnostics artifacts as available
-
-Retention behavior is controlled by configured retention mode; failed runs are retained.
-
-Diagnostics metadata for LLM interactions may include semi-public prompt identifiers:
-- `prompt_id`
-- `prompt_version`
-- `prompt_source`
-- `embedded_path`
-- `sha256`
-
-These are diagnostic identifiers, not user-facing prompt override controls.
-
-## Stdout/stderr behavior
-
-Success behavior:
-- with `--output`, stdout is empty
-- without `--output`, stdout contains only transcript JSON in selected output schema
-- report JSON is not written to stdout
-
-Failure behavior:
-- stderr contains human-readable error summary
-- nonzero exit
-- diagnostics path is printed when available
-
-## Exit-code behavior
-
-- `0`: success
-- nonzero: failure
-
-Treat any nonzero exit as a failed invocation.
-
-## Secret redaction guarantees
-
-Audita redacts API keys and authorization secrets from:
-- effective config outputs (`audita config print-effective`, diagnostics effective-config artifact)
-- report artifacts
-- LLM diagnostics artifacts
-- surfaced request/response error messages
-
-Config files should reference secrets via environment variable names (`api_key_env`) rather than embedding secret values.
-
-## Compatibility and deprecation policy
-
-- Existing stable schema names, report metadata keys, and top-level command behavior are treated as public contract.
-- Compatibility inputs (legacy flags/env aliases) may remain during transition windows.
-- Any planned removal or behavior change should include clear compatibility notes and migration guidance.
-
-## Breaking changes after 1.0
-
-After 1.0, breaking changes include, for example:
-- changing default success/failure exit-code semantics
-- changing stdout/stderr routing semantics
-- silently changing default output schema shape
-- removing supported output schema names without compatibility strategy
-- changing report schema fields or meanings incompatibly
-- changing config version semantics incompatibly without version bump
-
-Additive fields, additive diagnostics, and new optional schema names are generally non-breaking when existing behavior remains intact.

docs/architecture/structured-llm.md

@@ -1,90 +0,0 @@
-# Structured LLM Architecture
-
-## Purpose
-
-This document describes Audita's structured LLM runtime boundary and adapter behavior.
-
-## Why Audita owns the adapter
-
-Audita owns a small structured LLM adapter so that core runtime behavior is controlled inside the repository:
-- request construction and schema handling are explicit and testable;
-- retries, timeouts, cancellation, and error redaction are consistent across modules and validators;
-- provider SDK types are not exposed outside the adapter boundary;
-- dependency weight and transitive provider-specific behavior are reduced.
-
-At runtime, the rest of Audita depends only on the internal contract:
-- `StructuredLLMClient`
-- `CompleteStructured(ctx, req, out)`
-
-## OpenAI-compatible request shape
-
-At a conceptual level, Audita sends chat completion requests with:
-- `model`
-- `messages` (role/content pairs)
-- `response_format`:
-  - `type = "json_schema"`
-  - `json_schema.name` (stable schema name)
-  - `json_schema.strict = true`
-  - `json_schema.schema` (registered JSON Schema payload)
-
-The adapter uses OpenAI-compatible `POST {base_url}/chat/completions` over `net/http`.
-
-## Structured response schema registry
-
-Structured response schemas are registered in `internal/framework/responseschema` with stable metadata:
-- schema key
-- schema ID
-- schema version
-- schema name (OpenAI-compatible `response_format` name)
-- raw JSON Schema payload
-- SHA-256 hash
-
-Current schemas:
-- `correction_set`:
-  - id `audita.correction_set`
-  - version `v1`
-  - name `audita_correction_set_v1`
-- `validator_decision_set`:
-  - id `audita.validator_decision_set`
-  - version `v1`
-  - name `audita_validator_decision_set_v1`
-
-## Provider compatibility assumptions
-
-Audita assumes an OpenAI-compatible chat-completions endpoint that:
-- accepts message arrays with model selection;
-- accepts `response_format.type = json_schema`;
-- returns a completion with assistant message content and optional usage metadata.
-
-Provider-specific differences are expected in strictness and error payload shapes, so the adapter treats provider output as untrusted until locally decoded.
-
-## Local decode and validation remain mandatory
-
-Provider-level structured output is a transport guardrail, not final validation.
-
-After receiving a response, Audita still:
-- decodes assistant content into typed request-specific structs;
-- validates proposal and validator payload invariants locally;
-- enforces deterministic validator/cardinality rules before any transcript application.
-
-This protects runtime correctness even when provider responses are malformed, partial, or semantically inconsistent.
-
-## Diagnostics and redaction
-
-When structured schemas are used, diagnostics metadata records:
-- schema ID
-- schema version
-- schema name
-- schema hash
-
-Diagnostics and surfaced errors preserve secret redaction:
-- API keys and bearer tokens are redacted from request/response/error artifacts;
-- redaction is applied before diagnostic files are written.
-
-## Runtime behavior guarantees
-
-The structured LLM path preserves existing runtime guarantees:
-- bounded LLM call execution through schedulers;
-- context-aware cancellation and timeout propagation;
-- retry behavior for transient failures and retryable malformed structured responses;
-- deterministic module/chunk/proposal/validator behavior outside provider nondeterminism.

docs/architecture/validators.md

@@ -1,148 +0,0 @@
-# Audita Validators
-
-This document describes Audita's built-in validator registry and module validator chains.
-
-For LLM-backed validator prompt asset details, see [`docs/prompts.md`](prompts.md).
-
-## Package ownership
-
-Built-in validator construction is package-owned under `internal/validators/<validator_key>`:
-- `internal/validators/confidence_threshold`
-- `internal/validators/original_text_presence`
-- `internal/validators/non_empty_corrected_text`
-- `internal/validators/no_effect`
-- `internal/validators/protected_terms`
-- `internal/validators/spoken_form_plausibility`
-- `internal/validators/meaning_reversal_review`
-- `internal/validators/editorial_review`
-
-Registry and chain wiring stay in:
-- `internal/validators/registry.go`
-- `internal/validators/chains.go`
-
-Shared validator runtime mechanics stay in `internal/framework/validators`:
-- request/result/decision models
-- decision cardinality helpers
-- protected vocabulary helpers
-- shared LLM validator runtime, batching, and diagnostics helpers
-
-Execution classification metadata is defined in `internal/validators/metadata`:
-- `deterministic`
-- `llm_backed`
-
-Runner ordering uses this metadata so deterministic validators run before LLM-backed validators without concrete framework type assertions.
-
-## Scope
-
-Validator chains are built-in runtime behavior.
-
-Current 1.0 boundary:
-- built-in validator keys and built-in module chains are stable runtime identifiers;
-- thresholds and batching knobs remain configurable where already supported;
-- arbitrary user-defined validator chains are deferred.
-
-## Built-in validator keys
-
-### Deterministic validators
-
-- `confidence_threshold`
-  - checks proposal confidence against module-specific configured threshold.
-- `original_text_presence`
-  - ensures target segment exists and `original_text` exists in current working segment text.
-- `non_empty_corrected_text`
-  - rejects blank/whitespace-only `corrected_text`.
-- `no_effect`
-  - rejects proposals where `original_text == corrected_text`.
-- `protected_terms`
-  - protects glossary-derived terms from unsafe mutations in non-glossary modules.
-  - glossary stages use glossary-specific protection logic but still report this same stable key.
-
-### LLM-backed validators
-
-- `spoken_form_plausibility`
-  - checks whether proposed spoken-form change remains plausible in transcript context.
-- `meaning_reversal_review`
-  - checks for likely meaning reversal or semantic contradiction.
-- `editorial_review`
-  - performs conservative editorial safety review.
-
-## Built-in module chains
-
-Current built-in chains resolved from `internal/validators/chains.go`:
-
-- `glossary`
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `spoken_form_plausibility`
-  - `meaning_reversal_review`
-
-- `homophones`
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `spoken_form_plausibility`
-  - `meaning_reversal_review`
-
-- `spoken_word`
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `editorial_review`
-  - `meaning_reversal_review`
-
-- `grammar`
-  - `no_effect`
-  - `original_text_presence`
-  - `confidence_threshold`
-  - `protected_terms`
-  - `non_empty_corrected_text`
-  - `editorial_review`
-  - `meaning_reversal_review`
-
-## Protected terms construction
-
-`protected_terms` has explicit constructors:
-- general constructor used by non-glossary modules through the built-in registry
-- glossary-stage constructor used by glossary chain resolution
-
-Both variants preserve existing behavior and report the stable key `protected_terms`.
-
-## Execution semantics
-
-- modules execute serially;
-- section proposal work can run concurrently within a module;
-- deterministic validators run before LLM-backed validators;
-- malformed/missing/duplicate/unknown LLM validator decisions fail safely;
-- approved proposals are applied once per module after section work settles.
-
-## Validator rejections vs proposal-application skips
-
-- validator rejection:
-  - proposal is denied by validator-chain review and appears in validator rejection reporting with validator key and reason code.
-- proposal-application skip:
-  - proposal passed validators but could not be applied under replacement-policy semantics (for example no matching span at apply time).
-
-These are separate outcomes and are reported separately.
-
-## Reporting and diagnostics identity
-
-- report validator decision/rejection entries use stable validator keys in `validator_name`.
-- validator LLM diagnostics include validator identity in interaction metadata and structured response schema metadata.
-- correction ledger entries include deterministic and LLM validator decision snapshots keyed by the same stable validator keys, and keep validator rejection distinct from application-level skip.
-
-Prompt assets are unchanged by the validator package-ownership refactor and remain built-in under `internal/prompts`.
-
-## Configurable knobs that remain supported
-
-- per-module confidence thresholds (`thresholds.*` / equivalent env+CLI overrides)
-- validation batching limits (`validation_max_prompt_tokens` / equivalent env+CLI overrides)
-- validation LLM model/base URL/timeout/retries/concurrency settings
-
-These tune validator behavior without exposing arbitrary user-defined chains.

docs/cli.md

@@ -0,0 +1,95 @@
+# Audita CLI Reference
+
+The shared version-2 configuration parser and offline profile inspection serve
+`process`, `config validate`, and `config print-effective`.
+The [configuration reference](config.md) owns fields, defaults, and validation;
+the [migration guide](migrations/v2.md) maps retired inputs.
+
+## Commands and help
+
+```sh
+audita process <transcript.json> --glossary <glossary.yml> [flags]
+audita config validate --config <config.yml> [flags]
+audita config print-effective [flags]
+```
+
+Root invocation, `--help`, `-h`, and command help succeed without loading a config
+file or resolving secrets. Unknown commands and flags return exit code `2`.
+`process` accepts exactly one transcript path, either before the flags or after
+them. With flags first, put the transcript last. Both config commands reject
+positional arguments and process-only flags.
+
+All three commands accept `--config <path>` and `--overlay <name>`. `config validate`
+requires an explicit nonempty `--config`; the other commands select the explicit
+file, then `/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml`.
+Missing system candidates are skipped; inaccessible selected files are errors.
+No file means a configuration error. `AUDITA_CONFIG` is ignored. The selected
+overlay is loaded beside the main file, not used to locate profile files.
+
+## Setting overrides
+
+Only flags actually supplied override file or overlay settings. `--modules ''`
+clears the module pipeline, and an explicit numeric zero reaches configuration
+validation. The shared setting flags are:
+
+| Setting | Flags |
+| --- | --- |
+| Profile selection | `--proposal-profile`, `--validation-profile` |
+| Pipeline and output | `--modules`, `--output-schema` |
+| Validation budget | `--validation-max-prompt-tokens` |
+| Chunking | `--target-sections`, `--max-section-tokens`, `--min-section-tokens` |
+| Normalization (numeric seconds) | `--normalize-max-segment-gap`, `--normalize-ellipsis-gap`, `--normalize-max-segment-duration`, `--normalize-max-segment-tokens` |
+| Confidence thresholds | `--glossary-confidence-threshold`, `--homophones-confidence-threshold`, `--spoken-word-confidence-threshold`, `--grammar-confidence-threshold` |
+| Context | `--transcript-description` |
+| Reporting | `--reporting[=true\|false]`, `--report-dir DIR` |
+| Debug | `--debug[=true\|false]`, `--work-dir DIR` |
+
+`process` alone accepts `--glossary` (required) and `--output`.
+An omitted or empty output value selects stdout. Nonempty output filenames are
+used literally, including whitespace.
+Reporting defaults to enabled. `--reporting=false` disables all three ordinary
+reports; an explicit false overrides a true file or overlay setting. With file
+output, reports go beside the transcript and `--report-dir` has no effect. With
+stdout output, `--report-dir` selects the report directory, defaulting to the
+invocation working directory. `--debug` separately enables retained detailed
+artifacts under `--work-dir`. Configuration inspection does not create or probe
+these output directories.
+Per-role model, base URL, key, timeout, retry, and concurrency flags have been
+removed. Models and generation controls belong to PromptKit profiles, backend
+credentials and capacity to configuration, and retry/deadline settings to the
+shared LLM section. Legacy configuration environment overrides are ignored.
+
+## Streams and exit status
+
+`config validate` loads settings and both selected PromptKit profiles offline,
+then prints `config is valid` on success. `config print-effective` accepts
+`--view settings|profiles|explain` and defaults to `settings`. Settings writes
+ConfigKit's redacted effective-settings YAML. Profiles writes an Audita-owned
+YAML summary of selected ID, backend, model, sampling controls, per-attempt
+timeout, service tier, and reasoning effort for each role. It omits credentials,
+endpoints, and arbitrary extra provider parameters. Explain writes a YAML
+sequence of field paths, source origins, and recorded derivations without
+resolved setting values. These views share the same settings and profile
+validation path as `process`. Unknown views are usage errors. The effective
+settings YAML is for inspection and is not guaranteed reloadable.
+
+For `process`, `--output` writes the corrected transcript to a file and leaves
+stdout empty on success. Without it, stdout carries transcript JSON only on
+success. Enabled reporting creates `errors.yml`, `recoveries.yml`, and
+`report.md` in the selected directory. Successful stderr begins with
+`Completed: N corrections applied; E work units with errors; R recoveries.`
+Report and debug locations follow only when created. Model-operation failures
+can coexist with a completed run and applied corrections; read the YAML reports
+for details. Debug artifacts are created only with `--debug`. See the
+[report contract](integrations/output-files.md).
+
+Exit codes are `0` for success, `1` for runtime or output failure, and `2` for
+usage or configuration input errors. Invalid settings and profiles return `2`;
+broken embedded assets, caller cancellation, fatal processing, and required
+output failures return `1`. A late report-write failure can follow a published
+transcript; exit status and report status remain authoritative.
+Configuration commands inspect profiles without contacting a provider.
+
+See the [input files](integrations/transcript-glossary-files.md),
+[subprocess contract](integrations/subprocess.md), and
+[artifact lifecycle](operations.md) for their respective details.

docs/config.md

@@ -0,0 +1,88 @@
+# Audita Configuration
+
+This reference describes the version-2 configuration loader in `internal/core/config`.
+Use the maintained [local](../examples/minimal-config.yml) and
+[service](../examples/production-config.yml) examples with their PromptKit
+[profiles](../examples/profiles/) and optional [testing overlay](../examples/overlays/testing.yml).
+Choose real models for your endpoint; a testing profile makes no cost or quality
+guarantee. See the [migration guide](migrations/v2.md) for retired inputs.
+
+## Loading and precedence
+
+The application selects one main YAML file: an explicit `--config`, then
+`/usr/local/etc/audita/config.yml` if present, then `/etc/audita/config.yml`
+if present. Missing system candidates are skipped; an explicit missing or empty
+path, inaccessible selected file, or absence of any selected file is an error.
+`config validate` requires an explicit file. ConfigKit loads built-in defaults,
+the selected file, an optional named overlay from `overlays/<name>.yml` beside the
+main file, and only explicitly supplied setting overrides, in that order. Imports
+and secret-directory bootstrap are disabled. Unknown fields and missing, null,
+or non-integer `version: 2` are errors. Version-1 files require migration; no
+compatibility loader or implicit environment configuration remains.
+
+Validation runs after overlays, explicit overrides, and derived settings have
+resolved. An override can therefore repair an invalid lower-precedence setting
+when that setting can be decoded. ConfigKit retains field origins and derivations
+for effective-configuration inspection. Its effective YAML redacts resolved
+credentials.
+
+Profile files are resolved against the selected main file's directory, or may
+use an absolute `profiles_dir`. Both selected profiles must exist and select a
+registered backend. Their effective endpoints must match that backend's endpoint
+after surrounding whitespace and trailing slashes are removed. Profiles cannot
+supply a competing `api_key_env` lookup. The backend credential supplies the
+runtime key when required; inspection makes no provider call. The
+[CLI inspection views](cli.md#streams-and-exit-status) show selected controls and
+field provenance separately from ConfigKit's redacted settings YAML.
+
+## Settings
+
+| Field | Default | Rule |
+| --- | --- | --- |
+| `profiles_dir` | `profiles` | Nonempty; profile files are resolved relative to the selected main file. |
+| `backends.<id>.endpoint` | None | Required nonempty endpoint for each backend. At least one backend is required. |
+| `backends.<id>.concurrency` | `1` | Positive finite capacity; shared by both roles using that backend. |
+| `backends.<id>.queue_capacity` | `1024` | Nonnegative; zero disables waiting admission. |
+| `backends.<id>.credential` | Unset | Optional environment-variable secret reference; omission or null means unauthenticated use. |
+| `llm.proposal_profile` | None | Required nonempty profile ID. |
+| `llm.validation_profile` | Final proposal profile | Omission inherits the final proposal selection; explicit empty or null fails. |
+| `llm.transient_retries` | `2` | Nonnegative additional attempts per generation. |
+| `llm.repair_attempts` | `2` | Additional structural repair generations, from zero through three. |
+| `llm.operation_timeout` | `15m` | Positive Go duration string per logical call. |
+| `pipeline.modules` | `[glossary, homophones, glossary, spoken_word, grammar]` | Ordered module keys; repeats are meaningful and an explicit empty list runs no correction modules. |
+| `output.schema` | `bare-segments` | `bare-segments` or `audita-v1`. See [output formats](integrations/output-files.md). |
+| `validation.max_prompt_tokens` | `2048` | Positive validation batching budget. |
+| `chunking.target_sections` | Unset | Optional positive target; omission or null selects automatic chunking. |
+| `chunking.max_section_tokens` | `8192` | Positive maximum estimate. |
+| `chunking.min_section_tokens` | `2048` | Positive and no greater than maximum. |
+| `normalization.max_segment_gap` | `4.0` | Nonnegative numeric seconds. |
+| `normalization.ellipsis_gap` | `3.5` | Nonnegative numeric seconds. |
+| `normalization.max_segment_duration` | `60.0` | Positive numeric seconds. |
+| `normalization.max_segment_tokens` | `2048` | Positive token estimate. |
+| `thresholds.glossary`, `.homophones`, `.spoken_word`, `.grammar` | `0.8` each | Confidence thresholds in `[0, 1]`. |
+| `context.description` | Empty | Trimmed prompt context, at most 500 bytes. |
+| `reporting.enabled` | `true` | Write the ordinary three-file report bundle when processing. |
+| `reporting.stdout_dir` | `.` | Report directory for stdout transcript output; ignored for file output or disabled reporting. |
+| `diagnostics.debug` | `false` | Persist detailed debug artifacts when processing. |
+| `diagnostics.work_dir` | `/tmp/audita` | Parent of the retained debug run directory; required only when debug is enabled. |
+
+Module keys select correction modules; validator chains remain built into them.
+The loader preserves module order and repetitions. Normalization time settings
+accept numeric seconds only. The operation timeout uses a duration string.
+`reporting.stdout_dir` must be nonblank only for reporting-enabled stdout runs;
+`diagnostics.work_dir` must be nonblank only with debug enabled. These path
+requirements are checked when choosing process destinations, not during
+configuration inspection.
+
+## Credentials
+
+A backend credential is a ConfigKit secret reference containing an environment
+variable name, such as `credential: AUDITA_SERVICE_KEY`. Every reference that
+survives overlay and override resolution must name a set, nonempty variable,
+including credentials on backends unused by the selected roles. A replacement
+reference resolves only the replacement; null removes the reference. General
+`AUDITA_*` and `OPENROUTER_API_KEY` variables have no implicit setting effect.
+Secret values are available to the runtime through the resolved reference, while
+ConfigKit's effective output and provenance contain no resolved secret values.
+
+Current invocation syntax belongs in [CLI](cli.md).

docs/configuration.md

@@ -1,190 +0,0 @@
-# Audita Configuration
-
-This document describes Audita's versioned YAML config support and related commands.
-
-## Purpose
-
-Audita's config file provides a stable place for pipeline defaults and runtime tuning that would otherwise require many environment variables or CLI flags.
-
-Use config files for baseline settings, then use environment variables and CLI flags for deployment and per-run overrides.
-
-## Supported version
-
-Current supported config version:
-
-- `version: 1`
-
-Rules:
-
-- missing `version` fails validation;
-- unknown versions fail validation;
-- unknown fields fail validation (strict decoding).
-
-## Config path resolution
-
-For `audita process`, config path resolution is:
-
-1. `--config <path>` if provided
-2. `AUDITA_CONFIG` if set and `--config` is not provided
-3. default `/usr/local/etc/audita/config.yml` if present
-4. fallback default `/etc/audita/config.yml` if present
-
-Missing-file behavior:
-
-- missing `--config` path: hard failure;
-- missing `AUDITA_CONFIG` path: hard failure;
-- missing both default-path files: non-fatal, run continues.
-
-## Precedence model
-
-Effective config precedence is:
-
-1. built-in defaults
-2. file config
-3. environment overrides
-4. CLI overrides
-
-## Supported YAML fields
-
-```yaml
-version: 1
-
-pipeline:
-  modules: [glossary, homophones, glossary, spoken_word, grammar]
-
-output:
-  schema: bare-segments
-
-llm:
-  proposal:
-    base_url: https://openrouter.ai/api/v1
-    model: openrouter/google/gemma-4-31b-it
-    api_key_env: AUDITA_LLM_API_KEY
-    timeout: 120s
-    max_retries: 3
-
-  validation:
-    base_url: https://openrouter.ai/api/v1
-    model: openrouter/google/gemma-4-31b-it
-    api_key_env: AUDITA_VALIDATION_LLM_API_KEY
-    timeout: 120s
-    max_retries: 3
-
-concurrency:
-  total_llm: 2
-  proposal_llm: 2
-  validation_llm: 1
-
-chunking:
-  target_sections: 8
-  max_section_tokens: 8192
-  min_section_tokens: 2048
-
-normalization:
-  max_segment_gap: 4s
-  ellipsis_gap: 3.5s
-  max_segment_duration: 60s
-  max_segment_tokens: 2048
-
-thresholds:
-  glossary: 0.8
-  homophones: 0.8
-  spoken_word: 0.8
-  grammar: 0.8
-
-context:
-  description: "optional transcript background context"
-
-diagnostics:
-  work_dir: /tmp/audita
-  retention: auto
-```
-
-`context.description` provides background-only transcript context for prompts.
-If both config and CLI provide a description, `--transcript-description` takes precedence.
-
-`output.schema` supports the built-in output schema registry values:
-- `bare-segments` (default)
-- `audita-v1`
-
-Unknown schema names fail clearly before transcript output is written.
-
-Duration-like fields accept either:
-
-- numeric seconds (for example `120`, `3.5`), or
-- duration strings (for example `120s`, `2m`).
-
-For LLM timeouts, duration strings must resolve to whole seconds.
-
-## Secret handling
-
-Use `api_key_env` for secrets:
-
-- `llm.proposal.api_key_env`
-- `llm.validation.api_key_env`
-
-These fields must contain environment variable names, not secret values.
-
-At runtime, Audita resolves those names from the process environment.
-
-Redaction behavior:
-
-- run diagnostics `effective-config.json` is redacted;
-- `audita config print-effective` output is redacted;
-- API keys are never emitted in plaintext by those outputs.
-
-## Config commands
-
-Validate a config file:
-
-```sh
-audita config validate --config ./audita.yml
-```
-
-Print redacted effective config:
-
-```sh
-audita config print-effective --config ./audita.yml
-```
-
-`print-effective` loads defaults, then file config, then environment overrides.
-
-## Example: local OpenAI-compatible endpoint
-
-```yaml
-version: 1
-
-llm:
-  proposal:
-    base_url: http://localhost:8000/v1
-    model: local/proposal-model
-    api_key_env: AUDITA_LLM_API_KEY
-    timeout: 90s
-    max_retries: 2
-
-  validation:
-    base_url: http://localhost:8000/v1
-    model: local/validation-model
-    api_key_env: AUDITA_VALIDATION_LLM_API_KEY
-    timeout: 90s
-    max_retries: 2
-
-pipeline:
-  modules: [glossary, homophones, glossary, spoken_word, grammar]
-
-diagnostics:
-  work_dir: /tmp/audita
-  retention: auto
-```
-
-## Compatibility notes
-
-Existing environment variables and lower-level CLI flags remain available for compatibility.
-
-Current guidance:
-
-- prefer file config for baseline behavior;
-- keep environment variables for secrets/deployment-specific overrides;
-- use CLI flags for per-run overrides.
-- validator chains are built-in and are not user-configurable in config.
-- prompt source selection and filesystem prompt overrides are not config options.

docs/development.md

@@ -0,0 +1,49 @@
+# Developing Audita
+
+Audita is a Go CLI for transcript polishing. Start with the
+[README](../README.md) for product orientation and the
+[internal overview](internal/overview.md) for the implemented component map.
+
+## Before Changing the Repository
+
+Coding agents must read and follow the [agent policy](policy/agents.md) before
+recommending, planning, or implementing changes. Read the
+[architecture policy](policy/architecture.md) for boundaries and invariants,
+then use the task guide below rather than reading every reference indiscriminately.
+
+Confirm the requested scope and behavior contract, make focused changes in the
+owning component, and update the relevant tests and canonical documentation
+when behavior changes. The [testing policy](policy/testing.md) governs test
+sufficiency; the [documentation policy](policy/documentation.md) governs content
+ownership. Commit discipline and agent Git restrictions are in the
+[agent policy](policy/agents.md#git-commit-policy).
+
+## Baseline Validation
+
+Use Go 1.25.5 or newer. From the repository root:
+
+```sh
+go build -o ./bin/audita ./cmd/audita
+go test ./...
+```
+
+Run targeted package tests for touched behavior and the complete suite for
+substantial changes. Documentation-only work normally needs affected-link,
+example, and whitespace checks; use `git diff --check`. Detailed component
+validation belongs with the change recipes below.
+
+## Task-Specific Reading Guide
+
+| Task | Read |
+| --- | --- |
+| Locate components or change orchestration | [Internal overview](internal/overview.md), then [pipeline](internal/pipeline.md) |
+| Change configuration or CLI behavior | [Configuration contract](config.md), [CLI contract](cli.md), and [configuration/CLI change recipes](internal/cli-config.md) |
+| Change modules or validators | [Modules](internal/modules.md) or [validators](internal/validators.md); follow their prompt, registry, and test references |
+| Change LLM transport or concurrency | [LLM integration contract](integrations/openai-compatible-llm.md), [LLM runtime](internal/llm-runtime.md), and [pipeline](internal/pipeline.md) |
+| Change prompts | [Prompt registry](internal/prompts.md) and the consuming [module](internal/modules.md) or [validator](internal/validators.md) |
+| Change input/output formats | [Input file contracts](integrations/transcript-glossary-files.md), [output-file contract](integrations/output-files.md), [encoder internals](internal/output-schemas.md), and [configuration](config.md) |
+| Change reports or diagnostics | [Operations](operations.md), [output-file contract](integrations/output-files.md), [subprocess integration](integrations/subprocess.md), and [diagnostics/reporting internals](internal/diagnostics-reporting.md) |
+| Change tests | [Testing policy](policy/testing.md) and the affected component's internal documentation |
+| Change docs or examples | [Documentation policy](policy/documentation.md), the canonical topic owner, and [example validation](internal/cli-config.md#validating-examples) |
+| Prepare a release | [Release policy](policy/release.md) and [release procedure](release.md) |
+| Record a significant architectural decision | [ADR lifecycle](policy/documentation.md#architecture-decision-records) and [ADR index](adr/README.md) |

docs/integration/subprocess-operations.md

@@ -1,96 +0,0 @@
-# Audita Subprocess Operations
-
-This document describes how parent processes should invoke `audita process` safely in production orchestration.
-
-## Recommended command form
-
-Use explicit file outputs for orchestrated runs:
-
-```sh
-audita process <transcript.json> \
-  --transcript-description "Brief context that may help resolve ambiguous terms." \
-  --glossary <glossary.yaml> \
-  --output <output-transcript.json> \
-  --report-json <report.json>
-```
-
-Additional flags that may be situationally appropriate:
-- `--config <path>` to select an explicit versioned config file.
-- `--output-schema <bare-segments|audita-v1>` to select transcript output shape.
-- `--work-dir <dir>` to control diagnostics location.
-- `--work-dir-retention <always|auto|never>` to control retained run directories.
-- `--total-llm-concurrency`, `--proposal-llm-concurrency`, and `--validation-llm-concurrency` when orchestration needs to set explicit LLM throughput controls.
-- `--modules ...` only when intentionally overriding the default sequence.
-
-For config-driven orchestration, validate config files in CI/preflight:
-
-```sh
-audita config validate --config <path>
-```
-
-## Stdout behavior
-
-- With `--output`: stdout is expected to be empty on success.
-- Without `--output`: stdout contains transcript JSON only on success.
-- Report JSON is never written to stdout.
-
-## Stderr behavior
-
-- Success path should be quiet or minimal human-readable logs.
-- Failure path writes concise human-readable errors.
-- When a diagnostics run directory exists, failure stderr includes its path.
-- Prompt/response diagnostic payloads are not streamed to stderr.
-
-## Output file behavior
-
-- `--output` writes transcript JSON in the selected output schema to the provided path.
-- Output write failures return nonzero and surface actionable errors.
-- The command does not silently ignore output write errors.
-
-## Report JSON behavior
-
-- `--report-json` writes a machine-readable process report to the requested path.
-- Run-directory `report.json` is written independently under diagnostics.
-- Best-effort failure reports are emitted when possible without masking the primary failure.
-- Report write failures return nonzero with clear stderr messaging.
-- Report diagnostics metadata references run-directory artifacts including utilization diagnostics and correction ledger paths when available.
-
-## Diagnostics directory behavior
-
-- Each run creates (when possible) a per-run diagnostics directory.
-- Typical artifacts include transcript, normalization, chunking, invocation, effective config, LLM diagnostics, `utilization-diagnostics.json`, `correction-ledger.json`, `report.json`, and `error.log` on failure.
-- Failed runs retain diagnostics.
-- Under `auto` retention, successful runs with skipped/rejected corrections are retained; clean successful runs may be removed.
-
-## Exit codes
-
-- `0`: success.
-- Nonzero: failure (input/schema/config/module/LLM/runtime/output/report/diagnostics errors).
-
-Treat any nonzero as a failed subprocess invocation.
-
-## Timeout and cancellation
-
-- Runtime operations propagate context cancellation and request timeouts through LLM/scheduler paths.
-- On cancellation or timeout, the process exits nonzero and should not hang.
-- If diagnostics were initialized before failure, failure artifacts remain available for debugging.
-
-## Secret redaction expectations
-
-API keys and configured secret values are redacted from:
-- reports (`--report-json` and run-dir `report.json`);
-- diagnostics artifacts (including effective config and LLM interaction artifacts);
-- surfaced adapter/runtime errors;
-- test fixtures and regression outputs.
-
-Parent-process logs should still avoid printing raw environment variables.
-
-## Parent-process pipe guidance
-
-To avoid deadlocks in orchestrators:
-- always read both stdout and stderr concurrently when invoking as a subprocess;
-- prefer file outputs (`--output`, `--report-json`) for machine workflows;
-- treat stderr as human-readable diagnostics, not structured data;
-- parse structured results from output/report files.
-
-For Go callers, prefer `exec.CommandContext` with explicit timeout/cancellation and buffered/streamed readers for both pipes.

docs/integrations/openai-compatible-llm.md

@@ -0,0 +1,34 @@
+# OpenAI-Compatible LLM Integration
+
+Audita uses the pinned [PromptKit](https://gitea.maximumdirect.net/eric/promptkit)
+built-in OpenAI-compatible transport. PromptKit owns chat-completions wire framing,
+provider envelope parsing, transient retries, per-generation timeouts, and JSON
+Schema repair. A backend's configured base endpoint must support the selected
+profile's model and structured JSON-schema responses; local configuration
+inspection does not contact the provider.
+
+Audita selects a role profile, supplies its registered backend credential, and
+passes the embedded response schema as a complete per-call output contract. It
+prepares each call once, checks that the prepared target still matches the
+inspected backend and endpoint, then executes that same prepared request. A
+profile cannot introduce a separate credential environment lookup. An absent
+backend credential permits an unauthenticated local service.
+
+`llm.transient_retries` counts additional attempts for eligible transient
+failures; explicit zero makes one attempt. `llm.repair_attempts` counts
+additional generations for schema-invalid model output. The operation timeout
+covers preparation, capacity waiting, retries, generation, and repair. Caller
+cancellation or an earlier deadline also applies. See [configuration](../config.md#settings)
+for defaults and [PromptKit's consumer contract](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.11.0/docs/consumers/pkg-promptkit.md)
+for the pinned protocol and retry details.
+
+A model operation that exhausts structural repair, a live-parent local operation
+timeout, capacity rejection, transport failure, malformed provider response,
+or provider HTTP failure is recoverable for that operation. Proposal generation
+then contributes no candidates from that response; a failed validation batch
+marks its candidates unavailable while other batches and later modules can
+continue. Caller cancellation, broken assets or configuration, and unknown
+internal failures remain fatal and cancel peers. Audita decodes into domain
+types only after PromptKit validates the JSON. Ordinary reports use safe reason
+codes and counts rather than provider messages. Full request/response content
+is written only with `--debug`; see [operations](../operations.md#diagnostics-directory).

docs/integrations/output-files.md

@@ -0,0 +1,112 @@
+# Output Files and Reports
+
+## Corrected Transcript
+
+The corrected transcript uses the selected `output.schema`: `bare-segments`
+or `audita-v1`. Its schema and text rules are unchanged by resilience
+reporting. See [configuration](../config.md#settings) for selection and
+[operations](../operations.md#required-output-files) for placement and writes.
+
+With reporting enabled, Audita produces three ordinary files:
+
+- `errors.yml`: affected work units and run-level failures;
+- `recoveries.yml`: successful logical model operations with repair, service
+  retry, or response filtering;
+- `report.md`: a concise overview and complete per-chunk matrices.
+
+The YAML documents have Audita-owned schemas `audita-errors` and
+`audita-recoveries`, each with integer `version: 1`. Both contain, in order,
+`schema`, `version`, `run_id`, `status`, `started_at`, `completed_at`, `summary`,
+and an always-present `errors` or `recoveries` sequence. Status is
+`in_progress`, `completed`, or `failed`. Times are UTC RFC3339 strings;
+`completed_at` is null until finalization. A recoverable model error can appear
+with `status: completed`.
+
+## Summary Counts
+
+Both documents carry the same summary, in this order:
+
+| Count | Meaning |
+| --- | --- |
+| `work_units_total` | Configured module instances × initial chunks. |
+| `work_units_started`, `work_units_completed`, `work_units_interrupted`, `work_units_not_started` | Explicit processing states. |
+| `work_units_with_errors` | Distinct units with recoverable operation failures. |
+| `run_errors` | Fatal processing or output failures outside that unit count. |
+| `corrections_applied` | Approved proposals applied, not replacement occurrences. |
+| `llm_operations` | Structured-client invocations, including pre-dispatch failures. |
+| `proposal_responses` | Structurally usable proposal responses. |
+| `recoveries` | Successful operations with one or more recovery reasons. |
+| `proposal_responses_with_filtering` | Proposal responses that filtered at least one defective proposal. |
+
+Unknown counts before processing are `null`; completed empty counts are `0`.
+Retries and repair generations are details of one operation, not additional
+operations. Provider token usage, execution timing, and attempt-level evidence
+remain in opt-in debug files.
+
+## Error and Recovery Records
+
+An `errors` record with `scope: work_unit` identifies zero-based `stage_index`
+and `chunk_index`, module key/instance, initial segment range, applied-correction
+count, and a `failures` sequence. Each failure identifies the operation
+(`proposal`, `validation`, or `validation_input`), optional validator and
+zero-based batch index, safe reason/message, sorted proposal indices when
+known, and execution counts when available. Multiple failed batches in one
+work unit form one error record. `scope: run` instead records phase, safe reason
+and message, optional origin, and output path where relevant. It can describe
+a fatal error or secondary output-write failure without inventing a failed
+model work unit. Raw provider errors and transcript text are absent.
+For `validation_decisions_invalid`, proposal indices identify only candidates
+whose required decisions were unavailable. Other candidates in the same batch
+may still be approved or rejected normally.
+
+Each recovery identifies its work unit and successful operation, with sorted
+`reasons`, a safe message, available execution counts, and sorted
+`filtered_proposals` details. One recovery can contain `output_repaired`,
+`service_recovered`, and `proposals_filtered` together. Filtering covers
+defective response proposals; low confidence, protected-term rejection,
+ordinary editorial rejection, and an apply-time conflict caused by an earlier
+accepted correction are not filtering recoveries. Failed operations are errors,
+not recoveries. The initial chunk identity remains stable even if earlier
+modules change its text.
+`validation_decisions_filtered` records unexpected validator decision indices
+ignored from a successful validation batch. It may accompany repair or service
+recovery reasons, but it does not add `filtered_proposals` or increment
+`proposal_responses_with_filtering`. A batch with unavailable requested decisions
+is an error even when unexpected indices were also ignored.
+
+## Markdown Overview
+
+`report.md` includes run ID, status, corrections, work coverage, error and
+recovery totals, input/output paths, and full work-unit error and recovery
+matrices. Rows show every initial chunk with 1-based chunk number and initial
+segment range; columns follow configured module-instance order, including
+repeats. Completed error cells are `0` or `1`; recovery cells are operation
+counts. `—` means not started and `n*` means interrupted with n facts recorded.
+An empty stage or chunk set is identified as having no work units. Short reason breakdowns count
+records containing each reason; a recovery can have several reasons, so those
+counts need not sum to the recovery total. The report refers to both YAML
+files. It does not include transcript excerpts or correction lists.
+
+## Debug-Only Files
+
+With `--debug`, `correction-ledger.json` contains proposed candidates and
+their application, rejection, or interruption details. `llm-execution.json`
+holds logical calls and PromptKit execution events; `utilization-diagnostics.json`
+groups backend capacity and timing observations while preserving known and
+complete usage semantics. Interaction request/response artifacts, selected
+profiles, configuration provenance, and input snapshots are also debug-only.
+These files are diagnostic evidence, not an alternate ordinary report schema.
+Each candidate-backed ledger entry includes `proposed_corrected_text`, the
+literal replacement span, even when it is `""` for a deletion. Applied entries
+also include `applied_corrected_text`, likewise `""` for an applied deletion.
+Rejected, unavailable, skipped, and interrupted entries omit the applied field; a module
+failure without a candidate omits both. These fields are replacement spans,
+not full resulting-segment text. Deterministic decision records now identify
+the blank-result guard as `non_empty_resulting_segment` (formerly
+`non_empty_corrected_text`); its `empty_resulting_segment` reason is unchanged.
+Validator decision records carry `disposition`: `approved`, `rejected`, or
+`unavailable`. A missing or duplicate required decision is `unavailable`, not
+an editorial rejection. The candidate ledger uses the same distinction in its
+top-level disposition. Interaction response artifacts retain supplied model
+decisions and the reconciliation result, including unavailable and ignored
+indices; unexpected indices never create candidate ledger entries.

docs/integrations/subprocess.md

@@ -0,0 +1,48 @@
+# Subprocess Integration
+
+This guide explains how a parent process can invoke Audita and consume its results.
+The [CLI reference](../cli.md) owns command syntax, streams, and exit semantics;
+[configuration](../config.md) owns discovery and precedence.
+
+## Invocation
+
+Use an argument array with the transcript path and required glossary argument,
+following the [process examples](../cli.md#commands-and-help). For machine workflows,
+select an explicit transcript file destination and a configuration path. Reports
+then go beside the transcript. For stdout output, select `--report-dir` or use
+the invocation working directory. Debug is independent and off by default.
+Supplying a configuration path also
+makes file selection predictable; only explicitly referenced credential variables
+are read from the environment.
+
+Resolve the executable deliberately and set the child's working directory so
+relative input/output paths and reported artifact paths have a known base.
+Read stdout and stderr concurrently to avoid pipe blocking, and impose timeout
+or cancellation in the parent process when needed.
+
+## Result Handling
+
+Interpret the exit result according to the
+[CLI output and exit contract](../cli.md#streams-and-exit-status).
+Consume transcript/report data through the
+[output-file contract](output-files.md), and treat stderr as human-readable detail.
+The three reports may be absent after early configuration or destination
+failures. After initialization, failed reports can describe invalid input or
+partial processing. Exit `0` can include recoverable model errors and applied
+corrections; inspect `errors.yml` and `recoveries.yml` for those outcomes. A
+transcript may have been written before a later required report failure, so
+the file alone does not prove completion.
+
+If processing fails, preserve available evidence and follow
+[operations recovery](../operations.md#failure-inspection-and-recovery).
+Ordinary reports are independent of optional retained debug evidence. Parent
+cancellation reaches the child through its context; Audita joins workers and
+attempts failed-status report finalization before exiting. Uncatchable
+termination can leave `in_progress` reports. Neither the existence of an output
+file nor a debug directory alone establishes successful completion.
+
+## Related Contracts
+
+- [Transcript and glossary inputs](transcript-glossary-files.md)
+- [LLM endpoint compatibility](openai-compatible-llm.md)
+- [Configuration inspection commands](../cli.md#commands-and-help)

docs/integrations/transcript-glossary-files.md

@@ -0,0 +1,50 @@
+# Transcript and Glossary Input Files
+
+This document owns accepted input formats. Use the [CLI reference](../cli.md) for
+invocation and error exit semantics, [operations](../operations.md) for artifact
+handling, and [output files](output-files.md) for corrected-transcript serialization.
+
+## Transcript JSON Contract
+
+Accepted top-level shapes are an array of segments or an object containing a
+`segments` array. The array must contain at least one segment. Unknown object
+fields are ignored by the current parser; an input envelope does not require a
+schema/version marker.
+
+| Segment field | Type and validation |
+| --- | --- |
+| `id` | Optional integer (or null). Supplied IDs must be unique; missing IDs are accepted. |
+| `speaker` | Non-empty string |
+| `start`, `end` | Numeric seconds, finite and non-negative; end must be at least start |
+| `text` | Non-empty string |
+| `categories` | Optional array of non-empty strings |
+
+Supply explicit start and end times for meaningful timing. The current decoder
+uses zero for missing or null timestamps rather than enforcing their presence.
+String validation checks empty values, not whitespace trimming. Normalization can
+merge segments and reassign IDs, so source IDs need not match output IDs.
+
+The [tiny transcript](../../examples/tiny-transcript.json) is a complete copyable
+input. Invalid JSON, unsupported top-level shapes, empty transcripts, invalid field
+types, duplicate supplied IDs, and the validation failures above are rejected.
+
+## Glossary YAML Contract
+
+The top-level object contains a `glossary` list with at least one entry.
+
+| Entry field | Type and validation |
+| --- | --- |
+| `name` | Non-empty string |
+| `category` | Non-empty string |
+| `summary` | Non-empty string |
+| `aliases` | Optional list of non-empty strings |
+| `plural` | Optional string |
+
+The parser checks empty strings without trimming them. Unknown YAML fields are
+ignored by the glossary parser, unlike the stricter
+[configuration-file schema](../config.md#loading-and-precedence).
+
+The [tiny glossary](../../examples/tiny-glossary.yaml) is a complete copyable input.
+Malformed YAML, an empty/missing glossary, invalid shapes, missing required text,
+and empty aliases are rejected. These formats describe input acceptance, not
+additional vocabulary restrictions.

docs/internal/cli-config.md

@@ -0,0 +1,66 @@
+# Configuration and CLI Change Recipes
+
+This document owns contributor editing and validation steps for configuration,
+CLI wiring, and maintained examples. Definitions of behavior belong to the
+[configuration](../config.md) and [CLI](../cli.md) contracts.
+
+## Changing Configuration
+
+1. Update fields, defaults, loading, and validation in `internal/core/config`.
+2. Update file, selected overlay, credential reference, and CLI mapping as
+   applicable. Preserve the precedence defined in the
+   [configuration contract](../config.md#loading-and-precedence), including the
+   distinct validation and effective-config commands.
+3. Update redaction when secrets are involved.
+4. Review configuration tests and related CLI tests for precedence, validation,
+   and source-specific behavior. Add or adjust tests for the changed contract
+   according to the [testing policy](../policy/testing.md).
+5. Update the configuration reference and affected [examples](../../examples/).
+
+## Changing CLI Behavior
+
+1. Implement parsing and wiring in `internal/cli`.
+2. Preserve [stream and exit contracts](../cli.md) unless a change is intentional
+   and documented. Review [subprocess integration](../integrations/subprocess.md)
+   when changing output, error, or report handling.
+3. Review `internal/cli` tests and `cmd/audita` subprocess tests; update coverage
+   for the changed behavior.
+4. Update the CLI reference and affected integration guidance.
+
+## Validating Examples
+
+Keep complete copyable files in [examples](../../examples/) and link to them from
+their canonical contract documentation. Keep them valid and secret-free.
+From the repository root, validate the maintained configuration examples:
+
+```sh
+go run ./cmd/audita config validate --config examples/minimal-config.yml
+AUDITA_SERVICE_KEY=synthetic-example go run ./cmd/audita config validate --config examples/production-config.yml
+AUDITA_SERVICE_KEY=synthetic-example go run ./cmd/audita config validate --config examples/production-config.yml --overlay testing
+go run ./cmd/audita config print-effective --config examples/minimal-config.yml --view profiles
+```
+
+Only referenced credentials read environment values; use a controlled synthetic
+value when validating the service example. Check affected CLI examples against the
+parser and use schema tests for maintained transcript/glossary inputs. Example
+validation should not invoke a live LLM merely to check documentation.
+
+## Runtime Change Checks
+
+Select checks according to the [testing policy](../policy/testing.md):
+
+- Configuration or CLI changes: inspect `internal/core/config` tests,
+  `internal/cli` tests, and `cmd/audita` integration tests.
+- Process/output changes: verify stdout, stderr, exit, output-file, and report
+  behavior against the [CLI contract](../cli.md) and
+  [subprocess guidance](../integrations/subprocess.md), including failures.
+- Report or diagnostics changes: review [report assembly tests and redaction
+  checks](diagnostics-reporting.md#key-tests), metadata fields, and artifact paths.
+- Schema or serialization changes: review [output-schema tests](output-schemas.md#key-tests)
+  and `internal/core/schema` tests, including supported formats and rejection of
+  unsupported schemas.
+- Changes affecting assembled output: inspect controlled HTTP subprocess workflows
+  and retained report/artifact assertions.
+
+Use the [contributor guide](../development.md#baseline-validation) for baseline
+commands. These recipes supplement, rather than duplicate, public contracts.

docs/internal/diagnostics-reporting.md

@@ -0,0 +1,63 @@
+# Diagnostics and Reporting Internals
+
+The [output-file contract](../integrations/output-files.md) owns public
+schemas and the [operations guide](../operations.md) owns physical placement.
+
+`internal/framework/outcomes` contains processing-owned operation references,
+counts, failure reasons, filtering details, and work states. It imports no
+reporting, CLI, or provider package. The runner records those facts per initial
+section and module instance. `internal/framework/processreport` maps them once
+into `internal/core/reporting.Snapshot`, including nullable counts, grouped
+work-unit errors, separate run errors, and one recovery per successful logical
+operation. Reports order stages, sections, validators, and batches by semantic
+identity, never concurrent completion time. The core reporting package owns
+Audita's versioned YAML DTOs, fixed Markdown rendering, and brief CLI
+completion formatter. YAML redaction walks serialized string nodes so numeric
+and boolean values retain their types. Ordinary reports omit raw causes,
+transcript text, complete decisions, and token/timing payloads.
+Validation operations with unavailable decisions report only affected requested
+indices, even when other candidates in the batch continue. Ignored extra
+validator entries yield `validation_decisions_filtered` only for a successful
+operation; they do not count as filtered correction proposals. Failed operations
+retain their execution counts and never produce a recovery record.
+
+`internal/core/diagnostics.OutputSession` owns active destination checks,
+transcript preflight, report initialization, regular-file replacement, and direct
+stream publication. Transcript input is protected from early report writes but
+may be the final transcript destination. The LLM runtime records successful reads
+through PromptKit's profile filesystem and supplies those source paths to CLI;
+no second filename-based discovery rule is used.
+CLI composes it after configuration/profile inspection and before reading
+inputs or dispatching model work. It does not create ordinary report parents.
+On a final write error it attempts other enabled reports and one failed-status
+pass; a published transcript is never erased by a later report failure.
+
+Debug is independent. Only `diagnostics.debug` creates a unique retained child
+under `diagnostics.work_dir`. CLI passes empty diagnostics paths and nil writers
+when debug is off. With debug on, source/normalized snapshots, invocation,
+effective settings, origins, profiles, initial chunk summary, correction ledger,
+indexed execution/utilization, and model interactions are written. Runner
+candidate capture includes proposals later rejected or interrupted. Ledger
+mapping uses optional string pointers so an empty proposed replacement is
+serialized for every candidate and an empty applied replacement appears only
+when that candidate applied; see the [public artifact contract](../integrations/output-files.md#debug-only-files).
+Debug write failures reach the CLI through an explicit concurrency-safe diagnostic
+callback and produce redacted stderr messages without aborting processing.
+Completed validator decisions returned with a fatal batch error remain available
+to the debug ledger; they never authorize applying an incompletely validated edit.
+Runner decision records carry the validator's explicit approved, rejected, or
+unavailable disposition into the ledger. Unavailable candidates have a distinct
+top-level ledger disposition, while fatal interruptions remain interrupted when
+no rejection or unavailable decision was returned. Successful validation
+interaction payloads contain the full decoded response, final content, and
+reconciliation outcomes; ignored indices do not become ledger candidates.
+Every debug writer uses all configured backend secrets for redaction, including
+unselected backend credentials and source snapshots. Debug can still contain
+sensitive transcript and prompt content.
+
+## Key Tests
+
+Useful tests: `internal/core/reporting/*_test.go`,
+`internal/core/diagnostics/*_test.go`,
+`internal/framework/processreport/*_test.go`, and
+`cmd/audita/main_integration_test.go`.

docs/internal/llm-runtime.md

@@ -0,0 +1,39 @@
+# Audita LLM Runtime
+
+## Client Boundary
+
+`llm.NewRuntime` assembles one PromptKit engine for both roles. It loads embedded
+prompt and schema assets, registers application backends, and inspects the
+selected proposal and validation profiles offline. A role-bound `llm.RoleClient`
+implements `structuredoutput.Client`; proposal and validator contracts alias that
+interface. Both roles share engine-local backend capacity.
+
+For each logical call, the adapter starts an operation context and a diagnostic
+record before preparation. It supplies the reserved hardening text, the selected
+profile and resolved backend credential, and the complete embedded schema/repair
+contract. `PrepareExecution` produces the exact messages and hashes retained in
+an Audita snapshot. The adapter checks the prepared backend and endpoint before
+`RunPrepared`. It records available final content, usage, model, and profile
+metadata and decodes only a validated result. Call hooks expose logical
+start/finish records and PromptKit execution events for debug utilization.
+
+The [integration contract](../integrations/openai-compatible-llm.md) describes
+wire ownership and failure behavior. `structuredoutput.Failure` has a safe
+message, fixed reason code, typed cause, and recoverable/fatal/canceled class.
+Counts distinguish provider attempts, transient retries, and structural
+repairs, including on errors. Secrets from all configured backends, including unselected
+ones, are available to the diagnostics redaction boundary.
+
+## Orchestration
+
+CLI processing supplies two role clients from the same engine to the runner.
+PromptKit's per-backend admission applies to both roles and every generation,
+retry, and repair. The runner still processes module instances in order,
+collects concurrent section work deterministically, and cancels remaining
+section work only on fatal failures. `llm.DiagnosticsWriter` remains the protected
+request/response artifact writer. `llm.ObservationCollector` captures metadata
+events under a short mutex and freezes them after runner work joins. Its logical
+records include preparation failures; summary usage comes from terminal events
+once per call and retains knownness and completeness for each field. Runner wall
+and validator timings remain separate from provider attempts. Ordinary reports
+derive compact runner outcomes; full observations are debug-only.

docs/internal/modules.md

@@ -0,0 +1,91 @@
+# Audita Modules
+
+## Scope
+
+This document covers internal module contracts and packages. User-selectable keys
+and the default sequence belong to [configuration](../config.md#settings).
+
+## Module Contract
+
+Modules implement `contracts.TranscriptModule`:
+- `Key()`
+- `ReplacementPolicy()`
+- `Validators()`
+- `Propose(ctx, req)`
+
+Runner resolves configured module specs to module instances through `internal/framework/modules`.
+
+## Built-In Modules
+
+Current module packages:
+- `internal/modules/glossary`
+- `internal/modules/homophones`
+- `internal/modules/spoken_word`
+- `internal/modules/grammar`
+
+Current replacement policies:
+- `glossary`: `replace_all`
+- `homophones`: `require_unique`
+- `spoken_word`: `require_unique`
+- `grammar`: `require_unique`
+
+## Proposal Generation Ownership
+
+Shared proposal-generation plumbing is centralized in:
+- `internal/framework/proposal_generation`
+
+Module packages own:
+- prompt selection (`internal/prompts` prompt IDs).
+
+`proposal_generation.BuildProposalVariables` encodes the glossary, selected
+transcript section, and trimmed description once for all four modules. PromptKit
+renders their embedded definitions from those variables and validates the
+correction-set schema before Audita maps proposals. Domain validators still
+check confidence, applicability, protected glossary terms, and replacement
+policy. Model-facing segment IDs are one-based positions within the proposal
+request. Generation maps returned references back to the current chunk's actual
+segment IDs, preserving response order and multiple proposals per segment.
+Unknown references become invalid-target candidates for the existing domain
+checks; they are never interpreted as actual transcript IDs. Debug request
+metadata records `segment_ids` in local-reference order (reference 1 is the
+first entry). Shared context encoding lives in `internal/framework/promptcontext`.
+An empty `corrected_text` deletes the identified span when the resulting segment
+remains nonblank. Whitespace-only original spans can express spacing corrections;
+matching stays literal under each module's replacement policy. No automatic
+whitespace cleanup follows application.
+
+## Validator Chain Ownership
+
+Built-in chains are resolved in `internal/validators` per module key.
+Module packages call the built-in chain resolver at construction.
+
+## Failure and Outcome Behavior
+
+Module construction failures are fatal. Recoverable proposal model failures
+record an unsuccessful operation and yield no candidates for that section;
+other sections and module instances continue. Fatal adapter or application
+failures cancel peers and retain partial observations. Consumer-visible
+outcomes are defined by the [report contract](../integrations/output-files.md).
+
+## Changing Modules
+
+1. Add or update the narrowly scoped package under `internal/modules/<module_key>`.
+2. Keep module prompt selection in the module and its definition in
+   `internal/prompts`; see [prompt assets](prompts.md).
+3. Wire module registration and catalog resolution through framework module
+   registration and the core module catalog. Update selectable keys in the
+   [configuration contract](../config.md) when they change.
+4. Verify replacement policy and validator-chain selection against the
+   [architecture requirements](../policy/architecture.md#correction-and-concurrency-invariants).
+5. Review module, prompt-variable, and proposal-generation tests. Inspect parity
+   and release fixtures when public output changes, choosing sufficient coverage
+   under the [testing policy](../policy/testing.md).
+6. Update this document for implementation changes and the appropriate public
+   contract owner for externally visible changes.
+
+## Key Tests
+
+- `internal/modules/*/module_test.go`
+- `internal/framework/modules/registry_test.go`
+- `internal/framework/proposal_generation/*_test.go`
+- `internal/cli/run_test.go` (pipeline/report integration)

docs/internal/output-schemas.md

@@ -0,0 +1,34 @@
+# Output Encoding Internals
+
+## Scope and Ownership
+
+`internal/core/outputschema` owns encoder registration and resolution. External
+JSON envelopes and segment fields are defined by the
+[output-file contract](../integrations/output-files.md#corrected-transcript);
+selectable values and defaults belong to
+[configuration](../config.md#settings).
+
+## Resolution and Encoding
+
+Configuration validation checks registry support. CLI orchestration resolves the
+selected encoder, passes the final transcript to it, and routes the encoded bytes
+to the destination defined by the [CLI contract](../cli.md#streams-and-exit-status).
+An unsupported key fails validation or resolution rather than selecting an
+implicit fallback.
+
+The registry delegates bare segment serialization to the transcript schema
+package and uses an envelope encoder for the versioned form. Report construction
+receives the selected key separately; see [report assembly](diagnostics-reporting.md).
+
+## Changing an Encoder
+
+Update registration, configuration validation, and serialization together. Inspect
+round-trip and shape tests, then update the external output-file contract if the
+serialized form changes. Keep internal helper changes separate from decisions
+about externally visible compatibility.
+
+## Key Tests
+
+- `internal/core/outputschema/registry_test.go`
+- `internal/core/config/config_test.go`
+- `internal/cli/run_test.go`

docs/internal/overview.md

@@ -0,0 +1,97 @@
+# Audita Internal Overview
+
+## Scope
+
+This document owns the implemented component map. For normative boundaries, see
+[architecture policy](../policy/architecture.md); for task-specific reading and
+baseline commands, see the [contributor guide](../development.md).
+
+## Package Map
+
+Executable and command orchestration:
+- `cmd/audita`: executable entry point.
+- `internal/cli`: argument parsing, config commands, streams, filesystem wiring, and process orchestration.
+
+Core deterministic components:
+- `internal/core/config`
+- `internal/core/schema`
+- `internal/core/normalization`
+- `internal/core/chunking`
+- `internal/core/outputschema`
+- `internal/core/diagnostics`
+- `internal/core/reporting`
+- `internal/core/modulecatalog`
+
+Framework orchestration and contracts:
+- `internal/framework/contracts`
+- `internal/framework/modules`
+- `internal/framework/proposal_generation`
+- `internal/framework/proposals`
+- `internal/framework/runner`
+- `internal/framework/validators`
+- `internal/framework/llm`
+- `internal/framework/responseschema`
+- `internal/framework/structuredoutput`
+- `internal/framework/processreport`
+- `internal/framework/promptcontext`
+- `internal/framework/stagename`
+
+Domain implementations:
+- `internal/modules/*`
+- `internal/validators/*`
+- `internal/prompts`
+
+Supporting areas:
+- `docs/`: canonical documentation.
+- `examples/`: maintained copyable inputs and configuration.
+
+## Current Scope
+
+Runtime state is held in memory; optional debug artifacts are retained only
+when requested. Audita does
+not currently implement a plugin framework, generic workflow engine,
+resume/checkpoint/manifest system, remote storage, or multi-process service mode.
+Its role-bound PromptKit adapter implements the structured LLM client contract.
+These are current implementation boundaries, not bans on future work.
+
+## Main Execution Path (`audita process`)
+
+The CLI inspects configuration and profiles, preflights and initializes active
+outputs, optionally creates debug evidence, invokes the runner, and finalizes
+transcript and reports.
+The [pipeline](pipeline.md) owns detailed state transitions; the
+[operations guide](../operations.md) owns the operator workflow and physical
+artifact lifecycle.
+
+## Component Responsibilities
+
+Core components implement configuration defaults/loading/validation/redaction,
+transcript and glossary schemas, normalization and chunking, output encoding,
+diagnostics artifact helpers, and public report structures.
+
+Framework components implement module and validator contracts, proposal
+generation and application, prompt context, runner orchestration, PromptKit
+execution and observation mapping, response schemas, and process-report/correction-ledger assembly.
+Domain packages provide correction modules, built-in validator chains and
+execution metadata, and embedded prompt assets.
+
+These describe the implementation; dependency and ownership requirements belong
+to the [architecture policy](../policy/architecture.md#ownership-and-dependency-direction).
+
+## Change Recipes
+
+- [Configuration, CLI, and example changes](cli-config.md)
+- [Module changes](modules.md#changing-modules)
+- [Validator changes](validators.md#changing-validators)
+- Runtime orchestration changes: use [pipeline](pipeline.md), [LLM runtime](llm-runtime.md),
+  or [diagnostics/reporting](diagnostics-reporting.md) and their test references.
+
+## Related Internal Docs
+
+- [`docs/internal/pipeline.md`](pipeline.md)
+- [`docs/internal/modules.md`](modules.md)
+- [`docs/internal/validators.md`](validators.md)
+- [`docs/internal/llm-runtime.md`](llm-runtime.md)
+- [`docs/internal/diagnostics-reporting.md`](diagnostics-reporting.md)
+- [`docs/internal/prompts.md`](prompts.md)
+- [`docs/internal/output-schemas.md`](output-schemas.md)

docs/internal/pipeline.md

@@ -0,0 +1,45 @@
+# Audita Internal Pipeline
+
+The [input contract](../integrations/transcript-glossary-files.md) owns input
+formats; [configuration](../config.md) owns settings; the
+[output contract](../integrations/output-files.md) owns serialized results.
+
+1. CLI filesystem adapters read transcript and glossary files. Schema parsing
+   produces source objects; transcript segments enter canonical normalization.
+2. Normalization sorts and merges segments using configured time and token
+   limits, reassigning IDs where needed.
+3. The runner partitions the normalized transcript exactly once before its
+   module loop. Initial section identity, segment membership, and token
+   estimates stay fixed while text evolves.
+4. Module instances execute in configured order. Sections may propose
+   concurrently, but proposal indices follow section order and response order.
+5. Deterministic validators precede LLM-backed validators. Every submitted
+   candidate receives one decision per validator on a normal return. Fatal returns
+   may preserve a checked subset of completed decisions as diagnostic evidence.
+   Recoverable failed batches mark their candidates unavailable; successful
+   survivors continue through all required validators.
+6. Only fully approved proposals reach application. Corrections apply in
+   proposal-index order within their owning section. Existing replacement
+   policies and apply-time safety checks resolve stale conflicts.
+7. CLI encodes the final transcript and publishes active files through the
+   output session. `framework/processreport` builds one snapshot for YAML,
+   Markdown, and CLI totals. Debug evidence is captured separately when enabled.
+
+`runner.Run` returns the final working transcript, initial section inventory,
+explicit `not_started`/`interrupted`/`completed` work units, compact model
+operation outcomes, and per-module results. Candidate payloads are retained
+in the returned output only when debug capture is requested. A work unit can
+complete with model errors and applied corrections. Recoverable model failures
+never trigger shared cancellation. Fatal setup, internal, application, or
+parent-cancellation errors cancel peers, join workers, and preserve earlier
+committed module text for partial reports; no completed transcript is published.
+
+Required transcript/report write failures are fatal. Optional debug writes
+report safely to stderr without changing decisions or exit status. See
+[operations](../operations.md) for file placement and failure handling, and
+[architecture policy](../policy/architecture.md#correction-and-concurrency-invariants)
+for invariants.
+
+Key tests: `internal/framework/runner/*_test.go`,
+`internal/framework/validators/*_test.go`, and
+`cmd/audita/main_integration_test.go`.

docs/internal/prompts.md

@@ -0,0 +1,59 @@
+# Audita Prompt Registry
+
+`internal/prompts` embeds seven PromptKit YAML definitions under
+`assets/prompts`. The IDs and selected version remain:
+
+- `modules.glossary.proposal` (`v1`)
+- `modules.homophones.proposal` (`v1`)
+- `modules.spoken_word.proposal` (`v1`)
+- `modules.grammar.proposal` (`v1`)
+- `validators.spoken_form_plausibility` (`v1`)
+- `validators.meaning_reversal_review` (`v1`)
+- `validators.editorial_review` (`v1`)
+
+PromptKit reads the embedded definitions and renders their system and user
+messages from named string variables. The role-bound adapter supplies
+`HardeningPolicy` alongside `ProposalContract` and `ValidationContract` from
+embedded fragments under `assets/shared`, after caller variables so callers cannot replace those
+instructions. Common literal-span and identity guidance lives in the proposal
+fragment; shared candidate-reference, context, and confidence guidance lives in
+the validation fragment. Module-specific policies and examples remain in YAML.
+Proposal prompts receive the section JSON, glossary JSON where used, and transcript description block.
+Validator prompts receive the correction payload and description block.
+The prompt assets retain the prior task instructions and output-schema
+association. The role-bound runtime supplies the complete JSON Schema output contract
+and repair count; the definition does not own an operator-facing repair default.
+The correction-set schema requires a string replacement and accepts an empty
+string. Proposal guidance uses it for eligible span deletion and permits literal
+spacing edits; validator guidance evaluates the full before/after segment under
+each validator's existing remit. A blank resulting segment remains a domain
+rejection, not a structural schema error.
+
+`internal/framework/responseschema` remains the only embedded source for
+`correction_set.v1.json` and `validator_decision_set.v1.json`. It exposes that
+filesystem to PromptKit and retains Audita's schema-key lookup. PromptKit's
+prepared execution supplies definition and rendered hashes; Audita does not
+recompute or promise equality with hashes from the removed template renderer.
+The small Audita registry records only prompt ID, version, source, and embedded
+path. Prepared runtime metadata belongs in diagnostics rather than the registry.
+
+The structured-call contract carries prompt ID/version, string variables,
+response-schema key, and logical-call attribution. Module and validator callers
+pass these values through the role-bound adapter. See the
+[LLM runtime](llm-runtime.md) for runtime assembly and
+[configuration](../config.md) for operator-selected profiles.
+
+Proposal instructions describe literal matching, each module's replacement policy,
+source-based proposal identity, and confidence in whether an edit is warranted.
+Proposal `id` and validator `correction_index` values are short, one-based
+references scoped to each request, not internal transcript/proposal identities.
+Mapping belongs to [proposal generation](modules.md) and
+[validator reconciliation](validators.md). Validator instructions distinguish
+confidence in the verdict from the verdict itself. Compact output examples
+illustrate the supplied schema without requiring fixed JSON formatting or key
+order. The grammar article exception and glossary mishearing exception apply
+consistently across their instructions.
+
+Only the three active validator definitions are retained. The unused internal
+`grammar_review` and `spoken_word_review` definitions and type selections were
+removed; built-in chains and operator configuration are unchanged.

docs/internal/validators.md

@@ -0,0 +1,156 @@
+# Audita Validators
+
+## Scope
+
+This document describes internal validator composition, execution order, and
+decision handling. These registry keys identify implementation components; they
+are not independently selectable through [configuration](../config.md#settings).
+
+## Ownership
+
+Built-in validator keys and chains:
+- `internal/validators`
+
+Shared validator runtime mechanics:
+- `internal/framework/validators`
+
+Execution-class metadata:
+- `internal/validators/metadata`
+
+## Built-In Validator Keys
+
+Deterministic:
+- `proposal_shape`
+- `confidence_threshold`
+- `original_text_presence`
+- `non_empty_resulting_segment` — rejects edits that would leave the segment
+  blank, regardless of replacement-span length.
+- `no_effect`
+- `protected_terms`
+
+LLM-backed:
+- `spoken_form_plausibility`
+- `meaning_reversal_review`
+- `editorial_review`
+
+## Built-In Chains
+
+Module chains are defined in `internal/validators/chains.go`.
+Glossary, homophones, spoken_word, and grammar each resolve a fixed ordered chain.
+
+## Runtime Execution
+
+For each module section:
+1. run deterministic validators;
+2. run LLM-backed validators;
+3. record decisions and compact operation outcomes;
+4. carry only approved proposals forward.
+
+Decision cardinality is enforced inside Audita: each candidate proposal must
+receive exactly one outcome per validator it reaches. The runner still rejects
+duplicate, missing, or unknown identities in an internal validator result.
+
+## LLM Validator Batching
+
+LLM validators:
+- build canonical validation request payloads;
+- batch using `validation.max_prompt_tokens` and the existing domain token estimator;
+- select an embedded prompt ID and pass payload/description variables to the
+  role-bound structured client with the validator-decision schema.
+
+Oversized single inputs are rejected with a consolidated
+`validation_input_too_large` operation, without a provider call. Recoverable
+model failures make only their batch candidates unavailable while other
+batches continue. PromptKit validates each complete response against the
+validator-decision JSON Schema and performs configured structural repair before
+Audita receives decoded decisions. Each actual batch presents candidates
+numbered from 1, independent of their internal proposal indices. Audita then
+reconciles those local references: one usable entry keeps its
+approval or editorial rejection; a missing or duplicate requested entry makes
+only that candidate unavailable; unexpected indices are ignored. An empty list
+makes all requested candidates unavailable. Decisions and unavailable indices
+map back to internal proposal identities; results are ordered by request identity
+and satisfy strict internal cardinality before the runner advances
+approved candidates. An unavailable candidate never advances as approved.
+Invalid confidence in a decoded test double also makes only its candidate
+unavailable; production schema validation still rejects out-of-range confidence
+before reconciliation. A partially unavailable batch records one failed
+operation with only those requested indices, while a successful batch with
+ignored extras records a decision-filtering recovery. Fatal errors stop the
+chain without returning unchecked candidates as approved. A local result builder
+retains already-collected decisions and operations on early returns; full
+cardinality checks and final sorting remain explicit on completed paths.
+Reconciliation makes no semantic retry or extra model call; guided recovery remains
+[deferred](../roadmap/future.md#domain-aware-response-recovery).
+
+Debug request metadata records `proposal_indices` in local-reference order
+(reference 1 is the first entry). Raw decoded responses and ignored indices use
+local references; reconciled decisions and unavailable indices use internal
+proposal indices. Batch token estimates use the same locally numbered payload
+as prompt encoding.
+
+## Decision and Rejection Reporting
+
+Runner collects decisions, rejected candidates, and operation outcomes. Ledger
+assembly groups decisions by deterministic versus LLM-backed execution metadata.
+The serialized names and fields belong to the
+[report and ledger contract](../integrations/output-files.md); implementation mapping
+is covered by [report assembly](diagnostics-reporting.md).
+
+## Changing Validators
+
+1. Implement validator behavior under `internal/validators/<validator_key>`;
+   update registry and chain wiring in `internal/validators`.
+2. Put shared runtime mechanics in `internal/framework/validators` only when
+   needed. Keep deterministic versus LLM-backed execution metadata explicit.
+3. Preserve exposed identities and decision semantics under the
+   [architecture policy](../policy/architecture.md#correction-and-concurrency-invariants).
+4. Review validator, registry, chain, batching, malformed-output, diagnostics,
+   and decision-cardinality tests; update meaningful coverage under the
+   [testing policy](../policy/testing.md).
+5. Update implementation guidance here and any affected external contract in its
+   [canonical owner](../policy/documentation.md#canonical-ownership).
+
+## Key Tests
+
+- `internal/validators/*_test.go`
+- `internal/framework/validators/*_test.go`
+- `internal/framework/processreport/correction_ledger_test.go`
+- `internal/cli/run_test.go`
+
+## Supplemental Validation Context
+
+Request construction prepares one segment-position lookup and protected vocabulary
+for all candidates. Glossary selection remains candidate-specific, including the
+proposed replacement text.
+
+Each candidate includes the immediately preceding and following source segments
+available within its fixed chunk, labeled `preceding` or `following` with speaker,
+text, and categories, plus glossary entries whose names, aliases, or
+recognized plurals appear in its original/resulting text or those neighbors.
+Glossary matching reuses the protected vocabulary; entries retain glossary order.
+Model-facing candidate and neighboring context omit internal segment IDs.
+This supplies evidence such as a troupe definition for a proposed `Jesters`
+correction or an adjacent speaker turn explaining deliberate repetition. A correction
+already clear from its own segment does not require such evidence. No audio,
+out-of-chunk context, or complete transcript is implied.
+
+Context is serialized inside the validation item and included in the existing
+payload token estimate. If one enriched item exceeds the configured limit, optional
+neighbors are removed from the end first, then glossary entries from the end,
+until it fits or only required candidate data remains. Required before/after text
+is never truncated. Existing oversized-candidate handling still applies if that
+required data does not fit. Context is supplemental, not a promise that every
+relevant definition or neighboring utterance always fits. This uses the existing
+payload estimator, not a new guarantee about the provider's full context window.
+
+## Protected-Name Repetition
+
+Spoken-word cleanup may reduce repeated protected-name occurrences when every
+original protected identity still appears in the resulting segment and surviving
+spellings satisfy existing capitalization checks. Full-segment preview supports
+both a replacement of the repeated phrase and deletion of a redundant span. This
+exception only relaxes the deterministic count guard; editorial and meaning-reversal
+validation still decide whether the repetition was accidental and the edit preserves
+meaning. Intentional emphasis and substantive references remain protected by those
+reviews. Other modules retain their existing protection rules.

docs/migrations/v2.md

@@ -0,0 +1,135 @@
+# Migrating to Audita v2
+
+Audita v2 adopts ConfigKit for configuration and PromptKit for LLM execution.
+Version-1 configuration files, environment overlays, and retired flags are not
+accepted. Copy a maintained [local](../../examples/minimal-config.yml) or
+[service](../../examples/production-config.yml) example, then select real models
+in the accompanying `examples/profiles/` files. A testing overlay is shown at
+`examples/overlays/testing.yml`; its model choice carries no cost, speed, or
+quality guarantee.
+
+## Configuration and invocation
+
+Use `version: 2` in a YAML main file. `--config` selects that file; otherwise
+`process` and `config print-effective` look at
+`/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml`.
+`config validate` requires `--config`. `AUDITA_CONFIG` no longer selects a file.
+Settings resolve from defaults, main file, optional `--overlay <name>` in
+`overlays/<name>.yml`, then explicitly supplied setting flags. General `AUDITA_*`
+environment overlays no longer apply. Unknown fields and version-1 files fail;
+there is no automatic converter.
+
+Keep transcript as the positional path and `--glossary` and `--output` as
+invocation paths. The former `--report-json` output has been replaced by the
+default `errors.yml`, `recoveries.yml`, and `report.md` bundle; use
+`--reporting=false` to disable it or `--report-dir` for stdout output. Keep
+`pipeline.modules` / `--modules` and
+`output.schema` / `--output-schema`. An explicit `pipeline.modules: []` or
+`--modules ''` runs no correction modules; omission uses the default sequence.
+Ordered repetitions still matter.
+
+Move `llm.proposal.base_url` and `llm.validation.base_url` to a shared
+`backends.<id>.endpoint` when both roles use one service. Move old role models to
+PromptKit profile files and select them with `llm.proposal_profile` and
+`llm.validation_profile`, or `--proposal-profile` and
+`--validation-profile`. Omitted validation selection inherits the final proposal
+selection; explicit empty or null is invalid. `profiles_dir` defaults to
+`profiles` beside the main config. Complete profiles name a registered backend
+and model. Per-role `--base-url`, `--validation-base-url`, `--model`, and
+`--validation-model` are retired.
+
+Move old `llm.*.api_key_env` to `backends.<id>.credential`, a scalar name of a
+nonempty environment variable, such as `credential: AUDITA_SERVICE_KEY`.
+Credentials on every surviving backend must resolve, including unselected
+backends. Omit or set null for an unauthenticated local backend. Profiles cannot
+perform another key lookup. `--llm-api-key` and `--validation-llm-api-key` are
+retired. `AUDITA_LLM_API_KEY`, `OPENROUTER_API_KEY`, and
+`AUDITA_VALIDATION_LLM_API_KEY` have no implicit meaning, though any of those
+names can be used explicitly as a credential reference.
+
+Replace `concurrency.total_llm`, `.proposal_llm`, and `.validation_llm` plus
+`--total-llm-concurrency`, `--llm-concurrency`,
+`--proposal-llm-concurrency`, and `--validation-llm-concurrency` with
+`backends.<id>.concurrency` (default 1) and optional `queue_capacity` (default
+1024; zero means no waiting admission). One backend capacity is shared by both
+roles and all retry/repair attempts within this Audita process. It does not
+coordinate other processes.
+
+Move old role `timeout` values and `--llm-timeout-seconds` /
+`--validation-llm-timeout-seconds` to profile `timeout_seconds`, a per-attempt
+limit, and `llm.operation_timeout`, a Go duration such as `15m` covering the
+whole logical call. Replace old `max_retries` and `--max-retries` /
+`--validation-max-retries` with common `llm.transient_retries` (additional
+eligible attempts per generation) and `llm.repair_attempts` (additional
+generations for structurally invalid JSON). Explicit zero disables either
+feature. Cancellation and the operation deadline cover capacity waiting,
+backoff, generation, and repairs. PromptKit owns which failures can retry or
+repair; Audita does not retry semantic rejections.
+
+The following domain settings and explicit flags retain their meaning:
+
+| File field | Explicit flag |
+| --- | --- |
+| `validation.max_prompt_tokens` | `--validation-max-prompt-tokens` |
+| `chunking.target_sections` | `--target-sections` |
+| `chunking.max_section_tokens`, `.min_section_tokens` | `--max-section-tokens`, `--min-section-tokens` |
+| `normalization.max_segment_gap`, `.ellipsis_gap`, `.max_segment_duration`, `.max_segment_tokens` | `--normalize-max-segment-gap`, `--normalize-ellipsis-gap`, `--normalize-max-segment-duration`, `--normalize-max-segment-tokens` |
+| `thresholds.glossary`, `.homophones`, `.spoken_word`, `.grammar` | `--glossary-confidence-threshold`, `--homophones-confidence-threshold`, `--spoken-word-confidence-threshold`, `--grammar-confidence-threshold` |
+| `context.description` | `--transcript-description` |
+| `reporting.enabled`, `.stdout_dir` | `--reporting[=true\|false]`, `--report-dir` |
+| `diagnostics.debug`, `.work_dir` | `--debug[=true\|false]`, `--work-dir` |
+
+Normalization gaps and durations are **numeric seconds**, so replace `4s`,
+`3.5s`, and `60s` with `4.0`, `3.5`, and `60.0`. The operation timeout remains a
+Go duration string. `chunking.target_sections` may be omitted or null for
+automatic chunking; zero is invalid. Null is not a general synonym for omission:
+nonoptional scalar settings reject it. Empty `context.description` clears
+context. Legacy `diagnostics.retention` and `--work-dir-retention` are removed.
+Debug is off by default and its directory is retained whenever enabled.
+
+Remove old implicit variables `AUDITA_MODULES`, `AUDITA_MODEL`,
+`AUDITA_VALIDATION_MODEL`, `AUDITA_BASE_URL`, `AUDITA_VALIDATION_BASE_URL`,
+`AUDITA_LLM_TIMEOUT_SECONDS`, `AUDITA_VALIDATION_LLM_TIMEOUT_SECONDS`,
+`AUDITA_MAX_RETRIES`, `AUDITA_VALIDATION_MAX_RETRIES`,
+`AUDITA_TOTAL_LLM_CONCURRENCY`, `AUDITA_LLM_CONCURRENCY`,
+`AUDITA_PROPOSAL_LLM_CONCURRENCY`, `AUDITA_VALIDATION_LLM_CONCURRENCY`,
+`AUDITA_VALIDATION_MAX_PROMPT_TOKENS`, `AUDITA_TARGET_SECTIONS`,
+`AUDITA_MAX_SECTION_TOKENS`, `AUDITA_MIN_SECTION_TOKENS`,
+`AUDITA_NORMALIZE_MAX_SEGMENT_GAP`, `AUDITA_NORMALIZE_ELLIPSIS_GAP`,
+`AUDITA_NORMALIZE_MAX_SEGMENT_DURATION`, `AUDITA_NORMALIZE_MAX_SEGMENT_TOKENS`,
+`AUDITA_GLOSSARY_CONFIDENCE_THRESHOLD`,
+`AUDITA_HOMOPHONES_CONFIDENCE_THRESHOLD`,
+`AUDITA_SPOKEN_WORD_CONFIDENCE_THRESHOLD`,
+`AUDITA_GRAMMAR_CONFIDENCE_THRESHOLD`, `AUDITA_WORK_DIR`, and
+`AUDITA_WORK_DIR_RETENTION`. Put those values in the named fields above or pass
+their surviving explicit flags.
+
+## Inspecting and operating
+
+Run `audita config validate --config <file>` before processing. Use
+`audita config print-effective --config <file> --view settings|profiles|explain`
+to inspect redacted ConfigKit YAML, selected PromptKit profile controls, or
+field origins and derivations. Each view resolves both roles offline without
+calling a provider. `--overlay testing` selects the same overlay for validation,
+inspection, and processing. The old effective-config JSON format is replaced by
+YAML; settings YAML alone does not show profile-owned model controls.
+
+The [ordinary output contract](../integrations/output-files.md) now uses
+version-1 `audita-errors` and `audita-recoveries` YAML documents plus a fixed
+Markdown overview. Compact outcomes distinguish logical operations from
+provider attempts. With debug enabled, utilization remains grouped by backend
+capacity and usage fields retain known/complete flags. Debug runs write
+`effective-config.redacted.yaml`, `selected-profiles.yaml`,
+`config-origins.yaml`, `utilization-diagnostics.json`, and
+`llm-execution.json`. Request and response artifacts remain redacted debugging
+material. See [operations](../operations.md#diagnostics-directory) for placement
+and paths, [configuration](../config.md) for all settings, and the
+[CLI reference](../cli.md) for current flags.
+
+Debug ledger consumers should recognize the deterministic validator key
+`non_empty_resulting_segment` in place of `non_empty_corrected_text`. The
+`empty_resulting_segment` reason is unchanged. Candidate-backed ledger entries
+now include `proposed_corrected_text: ""` for an empty replacement, and applied
+deletions also include `applied_corrected_text: ""`; omitted fields still mean
+there was no candidate or no applied edit, respectively. See the
+[debug artifact contract](../integrations/output-files.md#debug-only-files).

docs/operations.md

@@ -0,0 +1,83 @@
+# Audita Operations
+
+This guide owns physical output placement and failure handling. The
+[CLI reference](cli.md) owns flags and streams, [configuration](config.md) owns
+settings, and the [output contract](integrations/output-files.md) owns schemas.
+
+## Normal Workflow
+
+1. Validate a version-2 configuration and its selected profiles with
+   `audita config validate --config config.yml`.
+2. Prepare a transcript and glossary following the [input contract](integrations/transcript-glossary-files.md).
+3. Run `audita process input.json --config config.yml --glossary glossary.yaml --output corrected.json`.
+4. Check the exit status, stderr summary, corrected transcript, and reports.
+   A completed run can contain recoverable model errors and still apply
+   corrections that passed every required validator.
+
+## Required Output Files
+
+Reporting is enabled by default. For file transcript output, `errors.yml`,
+`recoveries.yml`, and `report.md` go in the transcript output's parent directory.
+For stdout transcript output, they go in `reporting.stdout_dir` (default `.`),
+overridden by `--report-dir`. `--reporting=false` suppresses all three reports
+and removes their path requirements. It does not suppress the stderr summary or
+fatal errors. Stdout contains only transcript JSON; file output leaves it empty.
+
+Existing report directories and parents for regular transcript files must be
+writable. Audita prevents report writes from overwriting transcript, glossary,
+loaded configuration, or loaded profile sources, and rejects conflicting output
+destinations. The final transcript may intentionally replace its own input after
+successful processing; other input/configuration/profile sources remain protected.
+Preparation never truncates an existing transcript. Device and pipe destinations
+(such as `/dev/null` and named pipes) receive ordinary writes at publication time;
+they are not opened during preflight or replaced with regular files. A pipe write
+can wait for its reader, and stream write errors are fatal. With reporting enabled,
+reports still use the output path's parent; disable reporting when that placement
+is unsuitable for a stream destination.
+
+With reporting enabled, all three reports are initialized as `in_progress` before
+input processing. Output is encoded fully before publication. Each regular output
+file is replaced through a temporary file on its destination filesystem:
+transcript first, then `errors.yml`,
+`recoveries.yml`, and `report.md`. This is per-file replacement, not a
+multi-file transaction. New files honor the process umask; replacing an existing
+file preserves its permission bits. A late report error may leave the transcript
+published; the process exits nonzero and attempts failed-status reports. On fatal processing
+or cancellation, no completed transcript is published, but partial reports are
+finalized where possible. An uncatchable termination can leave `in_progress`.
+
+## Diagnostics Directory
+
+`--debug` enables detailed evidence independently of reporting. One directory
+named `run-<id>` is created under `diagnostics.work_dir` (default `/tmp/audita`)
+and retained on success or failure. Without debug, no directory or payload
+files are created. The parent is created when needed only for debug. Failure to
+write optional debug evidence prints a redacted stderr diagnostic but does not
+change correction decisions or the exit status.
+
+Debug evidence includes invocation and effective settings, configuration
+origins, selected profiles, redacted source and normalized snapshots,
+normalization and initial chunk summaries, all captured proposed candidates
+and validator decisions in `correction-ledger.json`, execution/utilization
+records, and model interaction files. Individual artifact writes can fail; a
+directory's existence does not prove every file was written. These files can
+contain transcript text and prompt/response payloads. Configured backend
+credentials are redacted, but the transcript itself is not anonymized. Protect
+and retain the directory according to the sensitivity of the input.
+
+## Failure Inspection and Recovery
+
+Exit `0` means processing and every enabled required output completed. Exit
+`1` means cancellation, fatal processing, or required-output failure. Exit `2`
+identifies usage or configuration/destination selection errors. Recoverable
+model-operation failures appear in `errors.yml` as one record per affected
+work unit; they do not automatically make the run fail. `recoveries.yml`
+records successful operations that needed retry, repair, or proposal filtering.
+`report.md` shows every initial chunk and configured module instance, including
+interrupted and unattempted work. Ordinary reports omit transcript excerpts and
+full decision ledgers; enable debug when those are needed.
+
+Read stderr and report status first. If a required output failed after a
+transcript was published, do not treat the file alone as proof of a completed
+run. Fix the input, service, or destination problem and rerun. Audita has no
+resume/checkpoint API.

docs/policy/agents.md

@@ -0,0 +1,100 @@
+# Agent Policy
+
+**Base Policy Version:** 1.0
+
+**Last Updated**: 2026-09-23
+
+## Purpose
+
+Agents working in this repository must read and adhere to the following house rules.
+
+## Implementation Discipline
+
+**KEY PRINCIPLES:** Write the simplest idiomatic code that satisfies the user's
+request. Do not implicitly strengthen requirements, restrictions, or guarantees
+beyond what the user actually requested.
+
+**Keep implementation machinery proportionate.** Use straightforward,
+idiomatic code and introduce only the machinery needed to support the requested
+behavior and flexibility. Optimize for conceptual simplicity and
+maintainability, not for the fewest lines, fewest options, or narrowest product
+surface.
+
+**Do not invent restrictions.** Do not reject, prohibit, constrain, cap,
+harden, or otherwise narrow behavior unless required by the user, an explicit
+repository contract, or an unavoidable correctness constraint. Preserve
+operator and caller choice where policy has not been established. If an
+additional restriction appears advisable, present it as a decision rather than
+silently implementing it.
+
+**Do not invent guarantees.** Do not strengthen public contracts, guarantees,
+or invariants beyond those expressly requested by the user.
+
+**Use the standard library.** For example: in almost every case, the best
+approach to copying files is to use the standard library. Writing custom code to
+implement power-loss protection at the application level, or to guard against
+bad actors with privileged and concurrent access to the host's filesystem,
+is wasteful, unnecessary, and not idiomatic. Stick to the simple, standard, and
+idiomatic approach unless another path is expressly requested by the user.
+
+**An implementation detail is not repository policy.** Tests and documentation
+must not elevate incidental behavior into a durable guarantee or restriction
+unless that behavior was expressly selected as part of the contract.
+
+## Codebase Discovery
+
+This project uses codebase-memory-mcp to maintain a knowledge graph of the
+codebase. Always prefer its graph tools over grep, glob, or file search for code discovery.
+If the project is not indexed, run `index_repository` first.
+
+Use this order:
+
+1. `search_graph` to find functions, classes, routes, and variables by pattern.
+2. `trace_path` to trace callers and callees.
+3. `get_code_snippet` to read a specific function or class.
+4. `query_graph` for complex graph queries.
+5. `get_architecture` for the high-level project summary.
+
+Fall back to grep or glob for string literals, error messages, configuration
+values, non-code files, or when the graph tools return insufficient results.
+
+Examples:
+
+- Find a handler: `search_graph(name_pattern=".*OrderHandler.*")`.
+- Find callers: `trace_path(function_name="OrderHandler", direction="inbound")`.
+- Read source: `get_code_snippet(qualified_name="pkg/orders.OrderHandler")`.
+
+## Git Commit Policy
+
+By default, create a commit after completing each coherent unit of work, using a
+concise, plain-English message describing what changed. This policy is standing
+authorization to create those local commits; a separate user request is not
+required. An explicit user instruction to leave work uncommitted takes precedence.
+
+A unit of work is a focused, reviewable outcome, including its relevant tests and
+documentation. Do not commit after every intermediate edit or create empty commits.
+Complete the appropriate validation before committing. If work is incomplete or
+validation is blocked, report that state rather than presenting it as completed.
+
+Before committing:
+
+1. Run `git status --short` and review the relevant diff, including any staged
+   changes. Identify changes that predate the current unit of work.
+2. Exclude secrets, local environment files, generated junk, and unrelated changes.
+3. Stage only the files or hunks belonging to the completed unit. Use `git add .`
+   only when every included change belongs to that unit. Do not disturb unrelated
+   staged or unstaged work.
+4. Review `git diff --staged` and run appropriate checks for the change. Do not
+   repeat checks already passed for unchanged content merely to create a commit.
+5. Review `git log --oneline -n 10` and use
+   `git commit -m "<concise plain-English commit message>"`.
+6. Report the commit identifier and any remaining uncommitted work.
+
+If the unit cannot be safely separated from pre-existing work, leave it uncommitted
+and explain the overlap rather than silently including that work. If a commit
+fails, preserve the changes and report the failure; do not bypass hooks.
+
+Default commit authorization does not authorize pushing, publishing, amending,
+resetting, reverting, or rewriting history. Those actions require a separate user
+instruction. Do not use `git commit -a` or `git commit --no-verify` to bypass the
+staging and validation discipline above.

docs/policy/architecture.md

@@ -0,0 +1,111 @@
+# Architecture Policy
+
+**Base Policy Version:** 1.0
+
+**Last Updated**: 2026-09-24
+
+## Purpose and System Shape
+
+Audita is a single-process Go CLI for transcript polishing. Its architecture
+separates deterministic transcript processing from external LLM effects through
+narrow contracts and adapters. Correction modules and validators remain explicit,
+composable components in an ordered processing pipeline.
+
+This policy owns architectural requirements. The [internal overview](../internal/overview.md)
+owns the implemented component inventory and links to runtime mechanics.
+[CLI](../cli.md), [configuration](../config.md), [operations](../operations.md),
+and [integration documentation](../integrations/subprocess.md) own external contracts.
+
+## Ownership and Dependency Direction
+
+- **Core:** owns domain data handling and reusable configuration, schema,
+  normalization, chunking, output, diagnostics, and reporting contracts. Domain
+  behavior must not require CLI or provider-transport knowledge.
+- **Framework:** owns orchestration contracts and shared runtime mechanics for
+  modules, validators, proposals, LLM adapters, and report assembly.
+  It composes domain operations through explicit inputs and outputs.
+- **Modules:** own narrowly scoped correction behavior and module-specific prompt
+  choices. Keep module packages separate rather than collapsing their distinct
+  responsibilities into a broad generic stage abstraction.
+- **Validators:** own validation behavior and composition. Deterministic and
+  LLM-backed validators remain explicit, independently composable elements.
+- **Prompts:** own embedded prompt assets and metadata used by the runtime.
+- **CLI and adapters:** own command parsing, streams, filesystem wiring, and
+  transport. Keep these concerns outside module, validator, and schema logic.
+
+LLM transport belongs behind `StructuredLLMClient` and framework adapter
+boundaries. Prefer small shared helpers over broad rewrites. Keep module and
+validator inputs, outputs, and configuration dependencies explicit, without
+hidden global state.
+
+## Correction and Concurrency Invariants
+
+- Treat LLM responses as nondeterministic inputs. Preserve deterministic proposal
+  indexing, validator ordering, transcript mutation, reports, and serialization.
+- Validate corrections before final mutation, including local proposal checks,
+  decision cardinality, and apply-time safety. Provider-side schema enforcement
+  does not replace these checks.
+- Keep LLM concurrency bounded and context-aware through PromptKit's engine-local
+  per-backend admission. Proposal and validation calls selecting the same backend
+  share that capacity, including retries and repairs. Preserve deterministic
+  collection order. Recoverable model-operation failures stay local to their
+  proposal response or validation batch; fatal invariant/application failure
+  and parent cancellation stop peers and join them.
+- Keep module and validator identities, reason codes, and execution-class
+  distinctions stable where exposed as contracts.
+
+The [pipeline](../internal/pipeline.md), [modules](../internal/modules.md),
+[validators](../internal/validators.md), and [LLM runtime](../internal/llm-runtime.md)
+describe how these requirements are implemented.
+
+## State and Compatibility
+
+Keep transcript mutation and output handling deterministic around external
+calls. Reports and correction ledgers are derived from processing results.
+Preserve the separation between transient processing state and filesystem
+adapters; physical report placement and optional debug retention belong to
+[operations](../operations.md).
+
+Public CLI, configuration, output-schema, diagnostics, report, prompt, module,
+and validator contracts should remain stable unless a change is deliberate and
+tested. The resilience work intentionally replaced the earlier JSON process
+report with versioned errors/recoveries YAML and a fixed Markdown overview;
+future changes should treat those current contracts deliberately. Current implementation
+limits are documented in the [internal overview](../internal/overview.md#current-scope);
+they are not additional prohibitions on future development.
+
+## Errors and Secret Handling
+
+Errors should identify the failed processing phase sufficiently for users and
+subprocess callers. Preserve the stream and exit behavior defined by the
+[CLI contract](../cli.md) and the artifact handling defined by
+[operations](../operations.md).
+
+Diagnostics, reports, and surfaced errors must not leak configured LLM secrets.
+Configuration redaction and LLM payload/error redaction have separate
+responsibilities and should remain separate. Implementation details belong in
+[diagnostics and reporting](../internal/diagnostics-reporting.md).
+
+## Dependencies
+
+Keep Audita dependency-light and prefer the Go standard library for CLI parsing,
+filesystems, synchronization, and tests. ConfigKit owns configuration loading,
+presence, provenance, and secret resolution; PromptKit owns prompt/profile
+loading, OpenAI-compatible transport, retries, repairs, and backend admission.
+Audita owns domain meaning, role selection, and deterministic processing.
+
+Do not introduce broad CLI, dependency-injection, workflow, logging, or plugin
+frameworks without a concrete implemented need and focused tests.
+
+## Testing and Documentation
+
+Follow the [testing policy](testing.md) for risk-based sufficiency and stable test
+boundaries. Keep tests close to the behavior they protect; shared helpers should
+remove duplication without obscuring component-specific behavior. Relevant
+internal docs identify component tests and change recipes; the
+[contributor guide](../development.md) provides baseline validation commands.
+
+Follow the [documentation policy](documentation.md) for ownership, current versus
+future behavior, and the ADR lifecycle. Keep architecture concise and normative;
+use the roadmap for implementation plans rather than maintaining a changelog or
+future-feature specification here.

docs/policy/documentation.md

@@ -0,0 +1,164 @@
+# Documentation Policy
+
+**Base Policy Version:** 1.1
+
+**Last Updated**: 2026-09-23
+
+## Purpose
+
+This policy assigns each documentation topic to one canonical owner. Its goal is
+to keep this application's documentation accurate, concise, discoverable, and resistant
+to drift for users, operators, developers, integrators, and LLM coding agents.
+
+## Core Rules
+
+### One Canonical Owner
+
+Each authoritative fact belongs in one document. A non-owning document may give
+a short, stable summary for orientation, but it must link to the canonical owner
+instead of repeating volatile details.
+
+Volatile details include commands, flags, configuration fields and defaults,
+module keys, schemas, file names, paths, status codes, retry behavior, and
+runtime guarantees. If readers could reasonably treat a statement as a
+contract, maintain it only in the owning document.
+
+### Current And Future Behavior
+
+Outside `docs/roadmap/`, documentation describes implemented behavior only.
+Partial features may be described only to their implemented boundary.
+
+ADRs are the narrow exception: an ADR may record an accepted architectural
+decision before implementation, but acceptance must not be presented as proof
+that the behavior exists. The roadmap owns implementation status and sequencing
+until the decision is implemented. Current architecture, user, operator,
+integration, and internal documentation are updated when the behavior lands.
+
+### Audience And Detail
+
+Write for the document's stated audience and include only the detail needed for
+its owned topic. User and operator docs should not expose implementation detail.
+Developer docs should link to user-facing and external contracts rather than
+restate them.
+
+### Examples
+
+Complete copyable files belong in `examples/`. Documentation may use the
+smallest illustrative snippet needed to explain its owned topic, but should link
+to maintained examples instead of embedding a second complete copy.
+
+Examples must be valid, secret-free, and tested where practical. Commands and
+configuration used in documentation should match the application.
+
+### Security And Privacy
+
+Documentation and examples must not contain real credentials, private keys,
+private environment dumps, sensitive source material, or private infrastructure
+details unless intentionally public. Document secret-handling mechanisms, not
+secret values.
+
+## Canonical Ownership
+
+| Topic | Canonical owner | Owned content | Content owned elsewhere |
+| --- | --- | --- | --- |
+| Product orientation and minimal end-to-end quickstart | `README.md` | What this application is, why it is useful, one shortest successful invocation, and links onward. | Complete command reference, configuration reference, operational procedures, implementation detail. |
+| Contributor entry point | `docs/development.md` | Task-oriented reading guide, minimal contributor orientation, baseline validation commands, and links to canonical docs. | Package inventory, architecture rules, subsystem behavior, detailed change recipes. |
+| Application architecture policy | `docs/policy/architecture.md` | System shape, normative ownership, dependency direction, architectural boundaries, invariants, safety properties, and non-goals. | Concrete package inventory, implementation mechanics, contributor procedures, decision history, future work. |
+| Agent policy | `docs/policy/agents.md` | Implementation discipline for coding agents, including proportionality, idiomatic simplicity, code discovery, Git commit restrictions, and rules against inventing requirements, restrictions, guarantees, or repository policy. | Application contracts, architecture, testing policy, and contributor procedures. |
+| Documentation policy | `docs/policy/documentation.md` | Documentation ownership, audience boundaries, maintenance rules, and ADR/document lifecycle. | Application architecture or product behavior. |
+| Testing policy | `docs/policy/testing.md` | Test philosophy, risk-based sufficiency, test boundaries, doubles, coverage guidance, regression-test policy, and criteria for adding, rewriting, or deleting tests. | Subsystem behavior, application contracts, subsystem-specific test inventories, and implementation plans. |
+| Release policy | `docs/policy/release.md` | Versioning and compatibility rules, release integrity and immutability requirements, release-documentation policy, required gates, publication controls, verification requirements, and exception policy. | Repository-specific commands, product installation guidance, CLI version semantics, historical release summaries, and CI implementation detail. |
+| Repository-specific release procedure | `docs/release.md` | Maintainer prerequisites, candidate preparation, concrete validation and publication commands, repository and registry guards, deliverable publication, consumer-facing verification, and failure recovery. | General release requirements, product installation guidance, historical release summaries, and current compatibility contracts. |
+| CLI contract | `docs/cli.md` | Commands, arguments, flags, invocation semantics, and exit codes. | End-to-end operating procedures, configuration field definitions, runtime filesystem layout, module implementation details. |
+| Configuration contract | `docs/config.md` | Discovery and precedence, file schema, fields, defaults, environment overrides, validation rules, and user-selectable module or validator keys. | Complete example files, CLI syntax, runtime state lifecycle, module implementation details. |
+| Operations | `docs/operations.md` | Runtime workflows, physical filesystem and state layout, output, cache, and debug handling, cleanup, permissions, recovery, and operational limits. | CLI flag syntax, configuration field definitions, logical output schemas, implementation mechanics. |
+| Integration contracts for external components | `docs/integrations/`, including `docs/integrations/output-files.md` for serialized outputs | External file formats and protocols, upstream and downstream contracts, logical output bundle paths and schemas, media types, and compatibility behavior. | Physical runtime placement and lifecycle, internal transformations, CLI syntax, configuration defaults. |
+| Inventory of implemented internal packages | `docs/internal/overview.md` | Current packages and components, their implemented responsibilities, and links to focused internal docs. | Normative architecture, contributor reading policy, external contracts. |
+| Behavior and architecture of internal packages | Other files under `docs/internal/` | Implementation flow, internal collaborators and state transitions, package-local guarantees and failures, and relevant tests. | Global architecture invariants, configuration definitions and defaults, external schemas, operator procedures. |
+| Contributor change recipes | Relevant files under `docs/internal/`, including `docs/internal/cli-config.md`, `docs/internal/modules.md`, and `docs/internal/validators.md` | Concrete editing and validation steps for each component. | Contributor orientation, global policy, and definitions of external contracts. |
+| Architectural decision records | `docs/adr/` | Significant decisions, context, alternatives, rationale, consequences, and supersession history. | Current behavior reference, implementation status, task sequencing. |
+| Future work and implementation status | `docs/roadmap/` | Proposed, accepted, deferred, or rejected work; implementation status; sequencing; and task breakdowns. | Implemented behavior reference and architectural decision rationale. |
+| Release-note history | `docs/releases/` | One checked-in historical summary for each source release made under the procedure. The note at the immutable tag is that release's record. | Current commands, behavior, contracts, and compatibility definitions. |
+| Complete copyable artifacts | `examples/` | Maintained configuration, inputs, and other files intended to be copied or run. | Field-by-field reference, command reference, prose explanation. |
+
+Documents that do not exist are required only when the corresponding interface
+or responsibility exists. Do not create placeholder API, consumer, integration,
+or operations documents for behavior the application does not have. Audita has
+retired its standalone troubleshooting guide; concrete operating and recovery
+procedures belong in `docs/operations.md`.
+
+## Boundary Rules
+
+### Orientation
+
+The README owns product orientation. The developer guide routes contributors.
+Architecture owns normative structure. Internal overview owns the current
+concrete component map. These documents may link to one another but should not
+maintain parallel package or behavior descriptions.
+
+### Commands, Configuration, And Operations
+
+CLI documentation answers how to invoke the application. Configuration
+documentation answers what settings mean. Operations answers what happens to
+runtime state and how to operate or recover the application. When a workflow
+crosses these topics, choose the document that owns the task and link to the
+other contracts.
+
+### Releases
+
+The release policy defines requirements and invariants that every release must
+satisfy. The repository-specific release procedure defines the commands and
+workflow maintainers use to satisfy that policy. Release notes are immutable
+historical summaries, not current-state contract owners: the checked-in note at
+an immutable tag records that release, while current canonical documentation
+must change with the behavior it describes. Do not use a release procedure or
+release note to weaken policy, redefine current behavior, or defer required
+documentation updates.
+
+### Contracts And Implementation
+
+Integration and API documents define externally observable shapes and
+semantics. Internal documents explain how this application implements or consumes those
+contracts. Internal docs may name a field, file, or protocol to identify a
+dependency, but must link to its canonical contract for the definition.
+
+### Security Topics
+
+This policy owns what documentation and examples may contain. Architecture owns
+application security invariants. Configuration owns credential-supply
+mechanisms. Operations owns permissions and handling of sensitive runtime
+artifacts. Internal docs own implementation mechanisms only.
+
+## Architecture Decision Records
+
+Use sequentially numbered ADR filenames such as
+`0001-record-architecture-decisions.md`. Follow the lightweight Nygard format:
+
+1. title;
+2. status;
+3. date;
+4. context;
+5. decision;
+6. alternatives considered;
+7. consequences.
+
+Treat the decision content of an accepted ADR as immutable. When a decision
+changes, create a new ADR and update the earlier ADR's status to superseded.
+Rejected architectural alternatives belong in the ADR; rejected product ideas
+belong in the roadmap.
+
+## Maintenance
+
+When behavior changes, update its canonical owner in the same change. If
+ownership moves, remove the old definition and replace it with a link where
+navigation remains useful.
+
+Before completing documentation work:
+
+- verify affected behavior and examples;
+- check commands, flags, fields, defaults, schemas, and paths against their
+  implementation;
+- keep unimplemented behavior in the roadmap, subject to the ADR exception;
+- remove stale references and validate links;
+- confirm that non-owning documents summarize and link rather than redefine;
+- confirm that no secrets or sensitive private data were added.

docs/policy/release.md

@@ -0,0 +1,168 @@
+# Release Policy
+
+**Base Policy Version:** 1.0
+
+**Last Updated**: 2026-09-23
+
+## Purpose
+
+This policy defines the minimum requirements for selecting, validating,
+publishing, verifying, and correcting releases. Its goal is to make every
+release intentional, reproducible from an exact source revision, documented
+for consumers, and safe from retrospective replacement.
+
+The repository-specific commands that implement these requirements belong in
+`docs/release.md`. Historical release summaries belong in `docs/releases/`.
+This is an Audita-specific adaptation of the base policy, not a requirement to
+implement the standard unchanged. Adapt it further when Audita's needs warrant;
+the canonical procedure must agree with the adopted policy.
+
+## Release Model
+
+A release identifies one exact source commit with an immutable version tag and
+its checked-in release note. The primary release branch is `main`.
+
+Release success means that the intended version tag has been created and
+successfully pushed to the authoritative remote with its source commit. The
+release procedure ends at that point. It must never wait for CI to complete or
+require successful CI status, artifact publication, or consumer installation as
+a completion gate.
+
+CI builds and publishes binaries and checksums asynchronously from the tagged
+source. Delayed or failed automation does not reverse release success. Artifact
+availability is a separate status and must not be claimed merely because the tag
+was pushed. Published artifacts must be traceable to the tagged source.
+
+Automation must not silently change the source candidate, move a published tag,
+or replace an existing published artifact.
+
+## Supported Platforms
+
+| Platform | Support |
+| --- | --- |
+| Linux amd64 | Required production target |
+| Linux arm64 | Required production target |
+| macOS arm64 | Best effort for local development; failures do not block a release or Linux artifact publication |
+
+Windows and macOS amd64 are unsupported. Neither their builds nor tests are
+release requirements. Native macOS testing and access to a Mac are not release
+prerequisites.
+
+## Versioning And Compatibility
+
+Stable releases use Semantic Versioning in the form `vMAJOR.MINOR.PATCH`, with
+no leading zeroes. Each version must be unused locally and on the authoritative
+remote before publication begins.
+
+Unless an adopted policy explicitly supports and defines prereleases,
+prerelease versions are not permitted.
+
+For releases at or after `v1.0.0`:
+
+- a major release may make intentional incompatible changes;
+- a minor release may add backward-compatible behavior; and
+- a patch release may make backward-compatible fixes.
+
+Before `v1.0.0`, a minor release may change a documented public, operational,
+configuration, data, integration, or compatibility contract when the release
+note explains the impact and required consumer action. A patch release must
+not intentionally break those contracts within its minor line.
+
+A repository with additional compatibility promises must define them in its
+adopted policy or canonical contract documentation. A release note may explain
+a compatibility change, but it must not redefine the current compatibility
+contract.
+
+## Release Documentation
+
+Every release made under this policy must include
+`docs/releases/<version>.md` in the tagged source tree. The filename must use
+the complete version tag, and the document must begin with a title containing
+the product name and that exact version.
+
+Release notes must contain a short summary and any relevant compatibility or
+upgrade instructions. Fixed sections and explicit no-action boilerplate are not
+required. Keep notes proportionate to the change.
+
+Release notes are immutable historical summaries. They should link to current
+canonical documentation for exact behavior. Every affected canonical document
+must be updated in the same candidate; a release note must not substitute for
+current documentation or defer required documentation changes.
+
+## Candidate Preparation And Validation
+
+Use the minimum gates needed for confidence in Audita's release. Provide one
+local command for the selected automated checks, following the
+[testing policy](testing.md). Run the deterministic, offline default test suite
+once for the candidate and verify that both required Linux targets build.
+Avoid duplicate test runs solely to satisfy separate release phases.
+
+Additional checks must address a concrete risk not already covered. Do not
+inherit a mandatory linting, documentation audit, platform matrix, or live-service
+checklist solely from the base policy. Documentation migration validation is
+separate work, not automatically a recurring release gate.
+
+Automate mechanical guards. Do not require manual approvals, sign-offs, laptop
+testing, or post-publication inspection as release gates. Version selection and
+a concise release summary remain preparation inputs, not separate approval rounds.
+
+## Publication Guards
+
+The procedure must automate the small set of checks needed to publish the
+intended source under a fresh version:
+
+- valid, unused version on the authoritative remote;
+- clean worktree and index on the primary release branch;
+- the exact candidate that passed the local checks, including its matching note;
+- publication of that source commit and the exact intended tag.
+
+Do not require remote CI success. Avoid an additional manual confirmation for
+each mechanical guard. Publish explicit refs rather than all local tags.
+A successful push is the completion signal; a failed or ambiguous push must be
+resolved against remote ref state before declaring success or retrying.
+
+## Immutability And Correction
+
+Published version tags and published deliverables are immutable. Do not move,
+overwrite, recreate, or delete a published tag. Do not replace an artifact,
+image, package, checksum, signature, provenance statement, or release note
+under an already published version.
+
+If validation fails before publication, correct and commit the candidate, then
+rerun the release command. Validation must apply to the exact candidate. An
+unpublished local tag may be removed after confirming that it was never published
+or used by a release channel.
+
+If a defect is discovered after publication, leave the published release intact. Correct the defect on the primary branch, choose a
+new patch version or other semantically appropriate new version, write a new
+matching release note, and perform the complete release procedure again.
+
+A distribution channel's supported yank, deprecation, or advisory mechanism
+may warn consumers away from a defective release, but it must not replace that
+release's contents or reuse its version.
+
+## Repository Release Procedure
+
+Maintain [docs/release.md](../release.md) as a short, executable maintainer runbook.
+It must identify prerequisites, authoritative remote, version and note preparation,
+one local validation command, automated source/tag guards, exact publication
+commands, and failure handling. It must name the asynchronous automation and link
+to this policy for supported platforms and release-success semantics.
+
+Describe how to diagnose or retry failed artifact delivery separately from release
+completion. Missing artifacts may be published later from the same immutable tag;
+existing published artifacts must not be overwritten. A source defect requires a
+new version. No polling or CI waiting belongs in the release command.
+
+## Audita Adaptations
+
+The maintainer approved these deviations from the base policy on 2026-09-23:
+
+- Audita's needs govern policy adaptation rather than strict template compliance.
+- Only Linux amd64 and arm64 are required; macOS arm64 is best effort.
+- A successful tag push completes the release; asynchronous CI cannot block it.
+- Gates are minimal, automated where practical, and avoid redundant validation
+  and human intervention.
+
+The [release procedure](../release.md) and its linked scripts implement these
+requirements. CI delivery remains asynchronous.

docs/policy/testing.md

@@ -0,0 +1,300 @@
+# Testing Policy
+
+**Base Policy Version:** 1.0
+
+**Last Updated**: 2026-09-23
+
+## Purpose
+
+Our tests exist to make **incorrect changes expensive and correct changes cheap**.
+
+We do not optimize for test count, line coverage, exhaustive isolation, or the fewest possible tests. We optimize for sufficient confidence in important behavior while imposing as little unnecessary friction as possible on future development.
+
+## Every test has a cost
+
+Testing is not an unqualified good. Every test imposes both an immediate cost and a continuing lifetime cost.
+
+A test must be:
+
+- written and reviewed;
+- understood by future maintainers and coding agents;
+- executed in local and CI workflows;
+- diagnosed when it fails;
+- updated when legitimate behavior changes;
+- maintained as fixtures, APIs, and dependencies evolve; and
+- removed or rewritten when it becomes redundant, brittle, misleading, or obsolete.
+
+Tests also create cognitive and architectural friction. They can constrain refactoring, duplicate policy, slow feedback loops, add noise to failures, and cause harmless implementation changes to require unrelated edits across the suite.
+
+A test is warranted only when the confidence it provides justifies these costs.
+
+Apply this cost-benefit analysis at two levels:
+
+1. **Per test:** What realistic defect does this test detect, how consequential would that defect be, and is that protection worth the test's lifetime cost?
+2. **Across the suite:** Does this collection provide materially more confidence than a smaller, simpler suite would?
+
+The preferred test suite is a **lean suite that provides sufficient confidence in the risks that matter, without redundant or low-value tests**. We seek sufficient confidence with the least unnecessary testing friction, not the fewest possible tests.
+
+Some friction is intentional. Tests should make dangerous changes—such as breaking compatibility, corrupting data, violating security boundaries, or reintroducing subtle bugs—require deliberate review. They should not make ordinary internal changes needlessly expensive.
+
+The cost of a test is not a reason to omit testing by default. Do not cite maintenance cost abstractly. When omitting a plausible test, be able to state why the protected failure is low-risk, already covered, obvious, reversible, or cheaper to detect elsewhere. For consequential, subtle, or difficult-to-observe behavior, the presumption should favor testing.
+
+## Default testing style
+
+Use a **classical/Detroit-style** approach:
+
+- Test observable behavior, resulting state, contracts, and invariants.
+- Use real internal collaborators when they are fast and deterministic.
+- Use fakes, stubs, or mocks primarily at expensive, nondeterministic, destructive, or external boundaries.
+- Prefer package-level behavioral tests over tests coupled to private helpers or internal call sequences.
+- Treat exact collaborator interactions as testable behavior only when the interaction itself is a requirement.
+
+Examples of appropriate seams include clocks, randomness, subprocesses, remote APIs, object storage, email, and paid LLM calls.
+
+## Test execution requirements
+
+Tests in the default suite must be deterministic, offline, and independent of real credentials. They must not invoke paid APIs or depend on mutable external services. Tests that require live infrastructure must be explicitly opt-in and clearly separated from the default suite.
+
+Control clocks, randomness, environment variables, and other process-global or machine-specific state when they affect behavior. Tests should be safe to run repeatedly and alongside other tests without depending on execution order or state left by an earlier test.
+
+## What deserves tests
+
+To the extent the following features are present in the application, prioritize tests for:
+
+1. Public and package-level contracts.
+2. Domain rules and important invariants.
+3. Boundary conditions and malformed input.
+4. Failure handling, cancellation, retries, recovery, and partial success.
+5. Serialization, schemas, compatibility, and round trips.
+6. Previously observed or plausible regressions.
+7. Representative integration and end-to-end workflows.
+
+A package-level contract is behavior relied upon by another package or major collaborator, not every observable detail of a package implementation.
+
+For behavior involving **data integrity, destructive operations, compatibility, security, concurrency, idempotency, or recovery**, presume that durable tests are required unless the behavior is already credibly protected at another layer.
+
+Do not add tests merely because a function, branch, or line exists. Do not add a test when the same meaningful risk is already adequately protected elsewhere.
+
+## Choose the right test boundary
+
+Test through the narrowest stable boundary that expresses the behavior clearly.
+
+This is often the package API, but it may instead be:
+
+- a smaller pure function when dense domain logic is most clearly isolated there;
+- a package-level operation when several internal collaborators jointly produce the behavior; or
+- a larger integration boundary when correctness emerges from interaction with a real dependency.
+
+Do not force all behavior through oversized end-to-end tests. Do not test every private helper merely because it exists. Choose the boundary that gives durable confidence with the least incidental coupling.
+
+## Test behavior, not implementation
+
+A test should protect a decision, contract, or invariant. A test must not memorialize the current implementation.
+
+Before adding or retaining a test, ask:
+
+> What realistic defect would this test catch?
+
+A test is suspect when its main purpose is to detect that someone:
+
+- changed an internal constant;
+- renamed or split a private helper;
+- reordered equivalent internal operations;
+- changed incidental formatting;
+- replaced one correct algorithm with another; or
+- refactored internal object structure without changing behavior.
+
+Refactoring should normally require no test edits unless the refactored structure is itself part of the contract.
+
+A test can be factually correct and still have negative value. Accurately describing current behavior is not sufficient; the protected behavior must be important enough to justify the future friction.
+
+## Expected effects of different changes
+
+Use the following expectations when evaluating test failures and test maintenance:
+
+| Change | Expected effect on tests |
+|---|---|
+| Internal refactor that preserves behavior | Existing tests should normally remain unchanged and continue to pass. |
+| Change to an internal default with no contractual significance | Behavioral tests should normally remain unchanged; tests should derive expectations from configuration or relationships rather than duplicate the old value. |
+| Intentional change to public behavior, policy, schema, or compatibility guarantees | The relevant tests should be reviewed and changed deliberately. |
+| Accidental violation of a contract or invariant | Tests should fail; fix the production code rather than rewriting the tests to accept the defect. |
+
+A test failing is not the same as a test needing to be edited. Many tests may correctly fail because of one production defect. The maintenance smell is a correct internal change that requires unrelated expectation updates throughout the suite.
+
+## Separate mechanism from policy
+
+Configurable thresholds and defaults must not be duplicated throughout the test suite.
+
+For example, do not encode an internal concurrency limit indirectly:
+
+```go
+// Production policy:
+const maxConcurrency = 4
+
+// Brittle test:
+err := startProcesses(5)
+require.Error(t, err)
+```
+
+Instead, test the mechanism relationally:
+
+```go
+const limit = 2
+runner := NewRunner(limit)
+
+require.NoError(t, runner.Start(limit))
+require.ErrorIs(t, runner.Start(limit+1), ErrTooMuchConcurrency)
+```
+
+The test should prove:
+
+- the configured limit is accepted; and
+- one beyond the configured limit is rejected.
+
+The production default should be tested exactly only when its literal value is itself a public, operational, safety, protocol, or compatibility requirement.
+
+Apply the same rule to limits, timeouts, capacities, retry counts, and ranges: test relationships and behavior, not duplicated literals.
+
+For concurrency limits, test both kinds of behavior when relevant:
+
+1. **Configuration enforcement:** invalid or excessive requested values are handled correctly.
+2. **Runtime enforcement:** observed peak concurrency never exceeds the configured limit.
+
+Use a test-controlled limit and measure the behavior relative to that limit. Do not merely assert today's default value.
+
+## Avoid semantic duplication across layers
+
+Each behavior should have a clear test owner.
+
+- Parser tests own parsing cases.
+- Validator tests own validation rules.
+- Domain tests own transformations and invariants.
+- Adapter tests own external integration behavior.
+- Orchestrator tests own coordination and failure propagation.
+- CLI tests own argument and configuration mapping.
+- End-to-end tests prove that representative assembled workflows work.
+
+Higher-level tests should not repeat every lower-level case. A single intentional policy change should not require unrelated edits across many test files.
+
+Tests that are individually reasonable may still be collectively redundant. Evaluate the marginal value of each additional test in light of the protection already provided by the rest of the suite.
+
+## Use test doubles deliberately
+
+Choose the least elaborate test double that provides the required control or observation.
+
+As a default:
+
+1. Prefer real collaborators when they are fast and deterministic.
+2. Use small in-memory fakes when realistic stateful behavior is helpful.
+3. Use stubs when a dependency only needs to provide controlled responses.
+4. Use mocks when the interaction itself is contractual.
+
+Mocks are appropriate when the contract includes facts such as:
+
+- a notification is sent exactly once;
+- a transaction is committed only after successful writes;
+- cancellation reaches a subprocess;
+- an expensive API is called no more than once; or
+- a security audit event is emitted.
+
+Do not use mocks merely to isolate every object or reproduce the implementation's call graph.
+
+## Go-specific guidance
+
+Use:
+
+- table-driven tests for meaningful behavioral categories and boundaries;
+- `t.TempDir()` for real filesystem behavior;
+- `httptest.Server` for realistic HTTP interactions;
+- fuzz tests for parsers, normalization, path handling, and broad input spaces;
+- golden files only when the complete output is intentionally stable;
+- integration tests where correctness depends on component interaction; and
+- a small number of representative end-to-end tests.
+
+Avoid exact error-string assertions unless the wording is itself contractual. Prefer `errors.Is`, `errors.As`, typed errors, or structured error fields.
+
+At CLI boundaries, prefer exit classifications, structured output, and the smallest stable semantic fragment needed to identify the error. Do not snapshot complete diagnostic wording unless it is contractual.
+
+Golden-file updates must require an explicit local flag. CI must not update golden files automatically, and reviewers must inspect the semantic diff before accepting an update.
+
+Keep tests readable and direct. Test helpers and fixture frameworks must earn their own maintenance cost; do not build elaborate test infrastructure for small or isolated needs.
+
+## Coverage
+
+Coverage is a diagnostic, not a target.
+
+Use it to find untested critical branches and unexpectedly weak packages. Do not write low-value tests solely to increase a percentage, and do not infer test quality from coverage alone.
+
+Pure domain logic will often warrant higher coverage than CLI wiring or external adapters. Uneven coverage is acceptable when it reflects risk.
+
+Increasing coverage is valuable only when the newly covered behavior protects a meaningful risk at an acceptable cost.
+
+## Regression tests
+
+A bug fix should normally include a regression test that fails before the fix and passes afterward.
+
+Retain the test when the defect could realistically recur and its consequences justify the ongoing cost. Prefer the narrowest durable test of the violated contract or invariant; do not preserve accidental implementation details from the original bug.
+
+Not every historical bug requires a permanent test. If the underlying design has made recurrence impossible, the test has become redundant, or a stronger invariant test now subsumes it, remove or consolidate it.
+
+## Deleting or rewriting tests
+
+Tests are maintained code, not permanent historical artifacts.
+
+Delete or rewrite a test when its maintenance cost exceeds the confidence it provides.
+
+Strong candidates include tests that:
+
+- require updates after harmless internal changes;
+- directly assert private constants without protecting a real contract;
+- duplicate the same policy across several layers;
+- verify mock choreography rather than outcomes;
+- snapshot large amounts of incidental output;
+- test trivial private helpers already exercised through stable package behavior;
+- protect risks already covered more effectively elsewhere;
+- are flaky, misleading, obsolete, or disproportionately expensive to diagnose; or
+- no longer correspond to a plausible failure mode.
+
+Several brittle tests may encode one genuine requirement. Replace them with one durable behavior-level or invariant test rather than preserving all of them.
+
+Deleting a low-value test can improve the quality of the suite by reducing noise, maintenance burden, and friction around legitimate change.
+
+## Reviewing a proposed test
+
+Use the following questions when the value, boundary, or durability of a proposed test is not self-evident. Significant test additions should be reviewable against them, but written answers are not required for every routine test.
+
+1. What realistic defect would it catch?
+2. How likely is that defect?
+3. How consequential would it be?
+4. Is the behavior already protected elsewhere?
+5. At which layer should this behavior be owned?
+6. Does the test assert a durable contract or an incidental implementation detail?
+7. Could the implementation be refactored without changing the behavior and without editing this test?
+8. What should cause this test to fail?
+9. What legitimate changes should not cause this test to fail?
+10. What ongoing maintenance, execution, and diagnostic cost will the test impose?
+11. Is there a smaller or more direct test that protects the same risk?
+
+Do not add the test when its expected lifetime cost exceeds its expected protective value.
+
+When deciding not to test plausible behavior, record or be able to explain why the risk is low, already protected, obvious, reversible, or cheaper to detect elsewhere.
+
+## Definition of sufficient
+
+A test suite is sufficient when:
+
+- important contracts and invariants are protected;
+- meaningful boundaries and failure modes are exercised;
+- realistic and consequential regressions are credibly protected against silent recurrence;
+- behavior involving data integrity, destructive operations, compatibility, security, concurrency, idempotency, and recovery is credibly protected;
+- important external boundaries have realistic integration coverage;
+- representative complete workflows are tested;
+- failures provide useful signal rather than redundant noise;
+- legitimate internal changes usually do not require test edits; and
+- additional tests would mostly repeat existing protection or preserve inconsequential implementation details.
+
+Sufficiency is a risk judgment, not a coverage percentage or test count. Reassess it as the application, its users, and the consequences of failure evolve.
+
+The governing rule is:
+
+> Test heavily where failure is consequential, subtle, or difficult to detect after the fact. Test lightly where failure is obvious, reversible, and inexpensive—and retain no test whose lifetime cost exceeds the confidence it provides.

docs/release-checklist.md

@@ -1,124 +0,0 @@
-# Audita Release Checklist
-
-Use this checklist before cutting a pre-1.0 or 1.0 release candidate.
-
-## Core test pass
-
-- Run:
-  - `go test ./...`
-- Confirm tests pass without live LLM credentials and without Python dependencies.
-
-## Config validation and precedence
-
-- Validate a representative config:
-  - `audita config validate --config <path>`
-- Inspect redacted effective config:
-  - `audita config print-effective --config <path>`
-- Confirm precedence behavior:
-  - defaults -> file config -> environment -> CLI.
-- Confirm default config search order:
-  - `/usr/local/etc/audita/config.yml` first, then `/etc/audita/config.yml`.
-- Confirm missing both default-path config files is non-fatal when `--config`/`AUDITA_CONFIG` are unset.
-
-## Output schema checks
-
-- Verify default output schema remains `bare-segments`.
-- Verify `--output-schema audita-v1` emits object payload with `schema` and `version`.
-- Verify unknown schema (for example `seriatim-intermediate`) fails clearly.
-
-## Subprocess contract checks
-
-- With `--output`, verify stdout is empty on success.
-- Without `--output`, verify stdout contains transcript JSON only.
-- Verify `--report-json` writes file output and does not write report JSON to stdout.
-- Verify failure stderr remains human-readable and includes diagnostics path when available.
-- Verify nonzero exit on failures.
-
-## Structured LLM checks
-
-- Verify runtime uses the Audita-owned OpenAI-compatible adapter.
-- Verify structured response schemas are attached via `response_format.type=json_schema`.
-- Verify diagnostics metadata includes structured schema `id/version/name/sha256`.
-- Verify provider output is still locally decoded/validated before use.
-
-## Report and diagnostics schema checks
-
-- Verify report metadata fields:
-  - `report_schema_name`
-  - `report_schema_version`
-  - `output_schema`
-  - `config_version` when file config is used.
-- Verify diagnostics artifact references exist in reports:
-  - transcript/normalization/chunking/invocation/effective-config artifacts
-  - utilization diagnostics artifact
-  - correction ledger artifact
-  - error log on failures.
-
-## Redaction checks
-
-- Verify secrets are redacted from:
-  - `effective-config.json`
-  - run-dir and `--report-json` reports
-  - LLM request/response/error diagnostics payloads.
-- Verify no API keys/bearer tokens leak into fixtures or outputs.
-
-## Prompt and validator metadata checks
-
-- Verify prompt metadata appears in LLM request metadata diagnostics:
-  - `prompt_id`, `prompt_version`, `prompt_source`, `embedded_path`, `sha256`.
-- Verify stable validator keys appear in report decisions/rejections.
-- Verify built-in validator chains resolve and execute for default and explicit module runs.
-
-## Utilization diagnostics checks
-
-- Verify `utilization-diagnostics.json` exists on successful runs.
-- Verify partial utilization artifact behavior on controlled failure paths.
-- Verify utilization fields are structurally present and nonnegative:
-  - effective concurrency
-  - run timing
-  - module timing summaries
-  - per-validator timing summaries.
-
-## Correction ledger checks
-
-- Verify `correction-ledger.json` exists on successful runs.
-- Verify report references ledger artifact path.
-- Verify ledger dispositions include applied/rejected and skipped/failed where exercised.
-- Verify validator rejection and proposal-application skip remain distinct.
-
-## Pipeline behavior checks
-
-- Verify default full pipeline run remains:
-  - `glossary`, `homophones`, `glossary`, `spoken_word`, `grammar`
-  - with deterministic repeated instance naming (`glossary_1`, `glossary_2`).
-- Verify explicit module runs (`--modules`) still work.
-
-## Failure and cancellation checks
-
-- Verify controlled failure paths retain diagnostics and produce best-effort failure reports.
-- Verify timeout/cancellation paths exit nonzero, do not hang, and retain failure diagnostics when initialized.
-
-## Release fixture/idempotence checks
-
-- Run release fixtures (`internal/cli/testdata/release`) through `go test ./...`.
-- Confirm fixture checks cover:
-  - must-apply and must-not-apply expectations
-  - protected-term survival
-  - report and diagnostics contracts
-  - output-schema checks
-  - prompt/schema metadata diagnostics
-  - utilization/ledger artifacts
-  - idempotence-oriented second pass no-op behavior with deterministic fake responses.
-
-## Deferred-feature guardrail
-
-- Confirm release docs do not claim support for deferred items:
-  - filesystem prompt overrides
-  - user-configurable validator chains
-  - arbitrary user-supplied output schemas
-  - resume/start-at/stop-after execution
-  - diff/check/propose-only modes
-  - generated transcript descriptions enabled by default
-  - interactive review UI
-  - UI/server wrapper
-  - provider benchmarking harness.

docs/release.md

@@ -0,0 +1,130 @@
+# Releasing Audita
+
+The [release policy](policy/release.md) defines versioning, supported platforms,
+immutability, and success. This runbook implements it with
+[scripts/release.sh](../scripts/release.sh). Publication ends after a successful
+source/tag push; binary delivery runs asynchronously.
+
+## Prerequisites
+
+Use Bash, Git, and Go 1.25.5 or newer (CI uses Go 1.25.5 or newer), with dependencies available
+locally or downloadable. Tests themselves are offline. Work from a
+checkout of `main` with permission to push that branch and tags to `origin`.
+For this repository, `origin` is `https://gitea.maximumdirect.net/eric/audita.git`.
+The script uses the checkout's configured `origin`; set it correctly before use.
+The remote must support atomic pushes. No CI credentials or Mac are needed locally.
+
+CI uses the existing `GITEA_RELEASE_TOKEN` secret for asynchronous asset delivery.
+macOS arm64 is a best-effort development environment; no macOS binary is published.
+
+## Prepare the Candidate
+
+Choose an unused version following the policy. Write
+`docs/releases/<version>.md`, beginning with `# Audita <version>`, followed by a
+short summary and any compatibility or upgrade instructions relevant to the
+release. There are no mandatory boilerplate sections. Update affected canonical
+documentation with behavior changes.
+
+Commit the intended changes, including the note, on `main`. The release script
+requires a clean index and worktree and does not create commits. There is no
+separate approval round. Published notes describe their tagged release; existing
+release history does not need to be reconstructed.
+
+## Validate and Publish
+
+From the repository root, substitute the chosen version:
+
+```sh
+bash scripts/release.sh --publish v0.2.0
+```
+
+This single invocation:
+
+1. Checks the version, branch, clean candidate, matching note title, and tag absence.
+2. Runs the default offline test suite once, then builds both required Linux targets
+   in a temporary directory using the same build script as CI.
+3. Checks that the candidate and worktree are unchanged and that it includes the
+   current `origin/main` history.
+4. Creates the version tag at that commit and atomically pushes the exact source
+   commit to `main` and the exact tag to `origin`.
+5. Reports release success and exits without polling CI or checking asset availability.
+
+To validate a prepared candidate without remote access, creating a tag, or pushing:
+
+```sh
+bash scripts/release.sh --check v0.2.0
+```
+
+This is an optional diagnostic command, not a required step before `--publish`.
+Publication validates within its own invocation; it does not use persistent cached
+results. Both modes require the candidate note and a clean `main` checkout.
+
+## Artifact Delivery
+
+[Woodpecker](../.woodpecker/release.yml) runs on tag events. It invokes
+[scripts/build-release.sh](../scripts/build-release.sh), then publishes:
+
+- `audita-<version>-linux-amd64`
+- `audita-<version>-linux-arm64`
+- `SHA256SUMS`
+
+Builds disable CGO and use the shared script's release flags. The version is in the
+asset name and source tag; Audita has no CLI version command. Cross-compilation
+checks compilation, not native execution on each architecture.
+
+The CI job status and logs show delivery progress or failure. Assets appear on
+the matching Gitea release page when delivery succeeds. A successful tag push
+must not be reported as confirmation that these files are available.
+
+Consumers can download the binary for their Linux architecture and `SHA256SUMS`
+from that page, check its listed SHA-256 digest with `sha256sum`, and make it
+executable. Optional consumer checks are not maintainer release gates. Digests
+detect file mismatches; they are not an independent signature of authenticity.
+
+## Failure and Retry
+
+**Local validation fails:** correct the issue, commit any changes, and rerun the
+publication command. A failure before tag creation publishes nothing. Temporary
+local build files are removed on exit.
+
+**Push fails or the connection is interrupted:** the script retains the local tag
+and reports failure rather than guessing whether publication succeeded. Inspect
+remote state with the actual version:
+
+```sh
+git rev-parse 'v0.2.0^{commit}'
+git ls-remote origin refs/heads/main refs/tags/v0.2.0
+```
+
+If the remote tag identifies the intended candidate and its source is on remote
+`main`, the release succeeded; no CI waiting is needed. An atomic push normally
+updates both refs or neither. If `main` has advanced since publication, fetch it
+and check candidate ancestry rather than requiring tip equality.
+
+If the tag is absent remotely and was never published or used by a release
+channel, remove only that unpublished local tag (`git tag -d v0.2.0`), resolve
+any rejected branch update, and rerun the command. If the remote tag has a
+different target, do not overwrite it; choose a fresh version. Never force a
+branch or tag update to bypass the guards.
+
+**CI delivery fails:** the release remains successful. Inspect the failed tag
+job's logs and retry that same tagged job for a transient failure. Publication is
+configured to skip existing filenames and not overwrite assets. Keep the same
+source and build environment when retrying; inspect any existing assets/checksum
+manifest before filling gaps. Do not regenerate and replace a published manifest
+or binary. If an existing manifest cannot describe the missing rebuilt assets,
+use a new version instead. A pipeline or source defect also requires a corrected
+commit and new version. An outage does not require moving the tag.
+
+**A released source defect is discovered:** fix it on `main` and publish a new
+appropriate version with a note explaining the correction. Leave the old tag and
+published files intact.
+
+## Maintaining the Release Workflow
+
+The shared build script keeps local and CI build settings aligned. Tests in
+[scripts/release_test.go](../scripts/release_test.go) exercise publication against
+temporary local Git remotes, including guard failures and rejected pushes, with
+Go execution stubbed to avoid recursive builds/tests. They run in `go test ./...`.
+Changes to CI delivery should preserve explicit asset names, checksums, and
+non-overwriting publication. CI delivery is separate from the local success path.

docs/releases/README.md

@@ -0,0 +1,9 @@
+# Release Notes
+
+- [Audita v2.0.0](v2.0.0.md)
+- [Audita v1.1.0](v1.1.0.md)
+
+See the [release documentation requirements](../policy/release.md#release-documentation).
+Use the [release procedure](../release.md) to prepare and publish new releases.
+
+This index does not reconstruct or replace records for existing tags.

docs/releases/v1.1.0.md

@@ -0,0 +1,22 @@
+# Audita v1.1.0
+
+This release refreshes Audita's documentation and repository policies and adds a
+small, automated release workflow. Transcript-processing behavior, CLI syntax,
+configuration, and serialized output formats are unchanged from v1.0.0.
+
+- Consolidated CLI, configuration, operations, and integration references, including
+  a dedicated reference for output files and reports.
+- Added contributor navigation and agent, testing, and release policies; retired
+  speculative troubleshooting documentation and the completed migration roadmap.
+- Added local release validation and publication scripts, with publication complete
+  after the source and version tag are pushed. CI delivers binaries asynchronously.
+- Isolated CLI tests from host configuration for a reliable local test baseline.
+
+Release binaries are now provided only for Linux amd64 and Linux arm64. Windows
+and Intel macOS are no longer supported. macOS arm64 remains a best-effort local
+build/development environment; macOS binaries are no longer published.
+
+Linux users need no configuration or data migration. Users of previously published
+Windows or macOS binaries should account for the distribution changes above before
+upgrading. See the [release procedure](../release.md) for artifact delivery and the
+[contributor guide](../development.md) for local build instructions.

docs/releases/v2.0.0.md

@@ -0,0 +1,38 @@
+# Audita v2.0.0
+
+Audita v2 adopts ConfigKit for configuration and PromptKit for LLM execution,
+adds resilient processing and focused reports, and improves correction prompts
+and validation. This is an intentional breaking release; existing installations
+must migrate their configuration and invocation settings.
+
+- Configuration uses `version: 2` YAML, named backends, PromptKit model profiles,
+  and optional named overlays. Offline configuration validation and inspection
+  expose effective settings, selected profiles, and provenance.
+- Proposal and validation calls share per-backend capacity. PromptKit handles
+  transport retries and structural response repair within bounded operation
+  timeouts.
+- Transcript chunks remain fixed through the LLM stages. Recoverable model
+  failures preserve progress and allow other work to continue. If every LLM
+  operation fails recoverably, processing still produces the deterministically
+  normalized transcript without LLM corrections. Cancellation, application
+  failures, and required output failures remain fatal.
+- Ordinary output includes `errors.yml`, `recoveries.yml`, and `report.md`, plus
+  a brief completion summary. Detailed correction and LLM diagnostics are opt-in
+  with `--debug`; `--reporting=false` suppresses all three ordinary reports.
+- Empty replacement text supports validated span deletion while preserving a
+  nonblank segment. Validator responses retain usable decisions despite missing,
+  duplicate, or extraneous entries; unavailable decisions never approve an edit.
+- Prompts use short request-local references, consistent shared instructions,
+  and relevant neighboring/glossary context for validation. Spoken-word cleanup
+  can reduce accidental protected-name repetition subject to validation.
+
+Follow the [v2 migration guide](../migrations/v2.md) before upgrading. Version-1
+configuration, implicit environment-setting overlays, retired model/credential/
+concurrency flags, and the former JSON report interface have no compatibility
+shim. Start from the maintained examples linked in that guide, configure profiles
+and credential references, and validate the new configuration. Update report and
+debug consumers for the [current output contract](../integrations/output-files.md).
+
+Linux amd64 and arm64 remain the production targets; macOS arm64 is best effort
+for development. Binary and checksum delivery runs asynchronously after source/tag
+publication, as described in the [release procedure](../release.md).

docs/roadmap/configkit.md

@@ -0,0 +1,171 @@
+# ConfigKit Adoption Roadmap
+
+**Last updated:** 2026-09-24
+
+**Status:** Production implementation complete. The selected configuration and
+overlay model is implemented; [implementation evidence](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md)
+records verification. No release has been published.
+
+## Purpose and Direction
+
+Adopt ConfigKit to harmonize Audita's configuration policy with the other
+applications in the suite. Consistent configuration conventions are the primary
+benefit; reducing application-owned configuration infrastructure is a secondary
+benefit and a useful measure of integration quality.
+
+This is an intentional breaking migration. There is no goal of backward
+compatibility with existing installations, configuration files, CLI flags,
+environment variables, or configuration-inspection output. Provide a migration
+guide instead of legacy loaders, aliases, dual operating modes, or an ongoing
+compatibility layer. Retain existing syntax only when it independently suits the
+new model, not because compatibility requires it.
+
+The initial review and prototype used ConfigKit v0.3.1, then verified v0.4.0
+selected overlays. Those findings remain historical; the prototype was removed
+after production integration. The application depends on a versioned module,
+without a sibling checkout or local module replacement.
+
+The shared boundary/prototype unit is complete. Its [selected contract, inventory,
+and evidence](library-boundary-findings.md) define the detailed target alongside
+this roadmap. Production configuration migration is complete.
+
+## Target End State
+
+- ConfigKit owns strict YAML loading, input presence, default application, explicit
+  override mechanics, recorded resolution, secret references, and safe inspection.
+- Audita owns its configuration types, field meanings, defaults, schema-version
+  policy, file selection, CLI parsing, domain validation, and runtime initialization.
+- Adopt ConfigKit's precedence and validation lifecycle rather than reproduce
+  Audita's current per-layer validation and environment-overlay behavior.
+- General configuration environment overrides are retired. Credentials use typed
+  environment-name secret references; supplied references must resolve according
+  to ConfigKit's contract. File selection follows the application-owned
+  discovery rules in the selected boundary.
+- Standardize duration representations, explicit empty/null/zero handling, and
+  inheritance on the shared conventions. Express genuine Audita domain rules
+  through the supported resolution and validation APIs.
+- Retained domain-setting flags and proposal/validation profile flags become
+  explicit ConfigKit overrides according to the migration inventory. Remove old
+  provider/key/retry/concurrency flags and aliases. Credentials use secret references.
+- Use ConfigKit's redacted effective output and provide source/explanation access
+  through the selected views below. Legacy JSON inspection is replaced.
+  Diagnostic configuration snapshots use the same new model.
+- Delete superseded loading and override machinery. Do not hide the old model
+  behind adapters that defeat the adoption's consistency benefit.
+
+Keep Audita's domain capabilities, such as ordered and repeated correction modules,
+expressible in the new model. Their existing configuration spelling is not binding.
+Transcript/glossary formats, processing algorithms, and output schemas are outside
+this migration unless later planning identifies a concrete dependency or an
+explicitly coordinated change.
+
+## Selected Overlays for Testing
+
+Enable ConfigKit's optional selected-overlay feature in the migration. A testing
+configuration that selects cheaper, faster LLMs is the initial use case. ConfigKit
+**overlays** modify application settings; PromptKit **execution profiles** still
+own model and generation settings. Keep `profiles_dir`, the PromptKit `profiles/`
+directory, and proposal/validation profile names unchanged by this terminology
+change.
+
+The planned interface is `--overlay <name>` on `process`, `config validate`, and
+`config print-effective`. Treat selection as invocation state passed to
+`Options.Overlay`, not as another persistent setting or environment overlay.
+Follow ConfigKit v0.4.0's selection contract: omission, an empty name, or `default`
+selects no overlay; a named selection loads `overlays/<name>.yml` beneath the selected
+configuration root. A missing selected overlay is an error. Use the library's
+single-overlay semantics without adding stacking, inheritance, or legacy-directory
+fallback.
+
+Precedence is application defaults, main configuration, selected overlay, then
+explicit setting flags. Perform Audita's role inheritance and semantic validation
+on the resolved configuration. ConfigKit's base parsing and schema-version checks
+still precede overlay application; overlays cannot replace the schema version or
+contain ConfigKit directives. General environment-variable overrides remain retired.
+Imports and secret-directory bootstrap remain disabled.
+
+Provide a maintained `overlays/testing.yml` example that selects testing execution
+profiles through `llm.proposal_profile` and `llm.validation_profile`. Model names and
+generation controls belong in the corresponding PromptKit profile files. Set both
+role selections in this example so a base configuration's explicit validation
+profile cannot accidentally retain the production model. Normal omission-based
+validation-profile inheritance continues to apply to the final effective settings.
+The example must make its backend/credential requirements clear and use no real
+secrets. Model choice is operator-controlled; “testing” does not promise a price,
+performance level, or equivalent correction quality.
+
+Overlays may modify other ordinary application settings supported by the schema;
+they are not restricted to profile selection. The resulting configuration must
+still satisfy shared backend capacity, credential, and domain rules. Validation,
+effective-config inspection, selected-profile inspection, and processing must all
+use the same selected overlay. Preserve overlay provenance in configuration
+explanations and diagnostics while retaining secret redaction.
+
+The [recorded overlay results](library-boundary-findings.md#selected-overlay-extension-2026-09-24)
+prove the representative load, inspection, and HTTP path. Production acceptance
+requires the same selection to reach both actual LLM roles, with explicit flags
+winning over overlays and surviving secret references resolved/redacted correctly.
+The isolated prototype is evidence, not a second supported runtime.
+
+## Inspection and Invocation Contract
+
+`process`, `config validate`, and `config print-effective` share file/overlay
+selection and retained setting overrides. Invocation I/O paths remain process-only.
+Only explicitly supplied flags override settings. Validation requires an explicit
+config file; processing and inspection use the selected discovery order. Help is
+available before configuration loading. All three commands validate selected
+profiles without requiring a provider call during setup.
+
+`config print-effective --view settings|profiles|explain` exposes three YAML views:
+
+- `settings` (default): ConfigKit's redacted effective configuration.
+- `profiles`: an Audita-owned summary of each selected role's profile, backend,
+  model, temperature, token limit, top-p, timeout, service tier, and reasoning
+  effort where present. Arbitrary extra provider parameters and credentials are
+  omitted; this is a selected-controls view, not every execution input.
+- `explain`: field paths, sources, and derivation records from ConfigKit, without
+  resolved secret values.
+
+A ConfigKit snapshot alone does not describe profile-owned controls. Profile
+inspection must enforce the registered-backend and credential boundary. Diagnostic
+snapshots use these safe views and preserve overlay/explicit-override provenance.
+The migration guide explains the intentional replacement of legacy inspection JSON.
+
+## Integration and Completion Criteria
+
+ConfigKit and PromptKit form one intentional breaking refactor. Intermediate
+commits need not build or run. The target contains no temporary old-runtime adapter,
+compatibility shim, processing gate, dual loader, or automatic legacy conversion.
+The [combined feature roadmap](library-adoption.md) owns execution/reporting intent;
+[implementation.md](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md) alone owns stage order and implementation tasks.
+
+The final application loads all retained domain settings, assembles both selected
+profiles, and processes through PromptKit. Controlled HTTP evidence must connect
+configuration selection to actual requests, domain outcomes, safe diagnostics,
+and shared finite backend capacity. Maintained examples and one migration guide
+cover the complete file/flag/environment transition. Obsolete application-owned
+loading, transport, rendering, and scheduler machinery is removed.
+
+The combined target is Audita v2.0.0. Release compliance and publication remain
+last and separate from implementing the refactor. Retention and transcript timestamp
+changes are not included; the selected retry/repair contract resolves the retry
+follow-up in [runtime behavior](runtime-behavior.md).
+
+## Scope Boundaries
+
+- Pin ConfigKit v0.4.0 in production and align Go/toolchain requirements. ConfigKit requires
+  Go 1.25; the combined PromptKit prototype requires Go 1.25.5. Adopting these libraries
+  does not automatically remove Audita's existing YAML dependency used for glossary parsing.
+- Implement the selected configuration and final PromptKit runtime directly,
+  without temporary runtime adapters, compatibility shims, or process gates.
+- Enable selected overlays as specified above. Keep imports and secret-directory
+  bootstrap disabled unless a separate concrete requirement justifies them.
+- Validation and inspection resolve surviving credential references and fail on
+  missing or empty values; no credential-free bypass is planned.
+- Inspection and diagnostics use the safe selected-profile view and ConfigKit
+  provenance described above. Changed formats are covered by the migration guide.
+
+The eventual completion test is that Audita follows the suite's ConfigKit
+conventions while retaining its intended domain capabilities, with a clear
+migration guide and no ongoing obligation to support the previous configuration
+model.

docs/roadmap/decision-set-reconciliation.md

@@ -0,0 +1,162 @@
+# Partial Success for Validator Decision Sets
+
+**Last updated:** 2026-09-25
+
+**Status:** Implemented. This document owns the target behavior and user intent.
+[implementation.md](implementation.md) owns concrete decisions, ordered work,
+and verification.
+
+## Purpose and User Intent
+
+Preserve usable validator decisions when other entries in the same model response
+have missing, duplicate, or unexpected correction indices. One defective entry
+should not prevent unrelated candidates from completing their required validation.
+This extends Audita's existing partial-success approach without another model call.
+
+Implement this simpler improvement before reconsidering guided domain recovery.
+Missing validation must never authorize applying a correction, but neither should
+it invalidate a different candidate's unambiguous approval or editorial rejection.
+The existing [resilience design](resilience.md) remains the foundation.
+
+## Historical Baseline
+
+The reviewed application baseline is `d2a6fcf`:
+
+- PromptKit validates the complete model response against Audita's embedded JSON
+  Schema and performs its configured structural repairs. Only structurally valid,
+  decoded responses reach domain decision mapping.
+- `mapLLMResponseToDecisions` in
+  [`llm_validators.go`](../../internal/framework/validators/llm_validators.go)
+  requires exactly one decision for every requested correction index. An empty
+  decision list, missing index, duplicate index, or unexpected index returns an
+  error instead of retaining usable decisions.
+- `LLMBackedValidator.Validate` consequently marks the entire batch unavailable
+  with `validation_decisions_invalid`. Other batches continue, but usable decisions
+  within that batch are lost.
+- The runner's cardinality checks enforce an internal contract: each candidate
+  receives one outcome from each validator it reaches. These checks apply to
+  Audita-owned results, not directly to raw model entries.
+- Deterministic validators already filter individual proposals. Unpreviewable
+  candidates and oversized validation inputs are isolated; recoverable provider
+  failures remain batch-local; application skips individual conflicts.
+- Ordinary reports distinguish failed operations from successful operations with
+  recoveries. Several failed operations in one stage/chunk share one work-unit
+  error record. The [output contract](../integrations/output-files.md) owns the
+  current counts and serialized representation.
+
+The defect addressed here is the all-or-nothing domain mapping boundary, not
+validation itself or the runner's internal consistency checks.
+
+## Target Decision Semantics
+
+Reconcile each structurally valid response against the indices actually requested
+in that particular validator batch. Correction indices, rather than segment IDs
+or positions in the response array, identify candidates. Several corrections may
+legitimately target the same segment.
+
+| Response condition | Target behavior |
+| --- | --- |
+| Exactly one valid decision for a requested index | Preserve its approval or rejection and associated decision evidence. |
+| Requested index omitted | Produce an explicit validation-unavailable outcome for that candidate. |
+| Requested index appears more than once | Treat that candidate's validation as unavailable; preserve unrelated decisions. |
+| Index was not requested in this batch | Ignore the extraneous entry and record the handled defect. |
+| Empty decision list for a nonempty request | Mark all requested candidates unavailable; do not fail the run. |
+| Decisions appear in a different order | Match by identity and retain deterministic internal results. |
+
+Duplicate handling does not choose the first or last answer, nor infer a verdict
+by voting. Even identical duplicates follow the agreed duplicate-index handling;
+accepting them specially is not needed for this feature. Duplicate extraneous
+indices remain extraneous entries, not reasons to invalidate requested candidates.
+
+An unavailable candidate remains unapplied and does not proceed as approved to
+later validators. It is distinct from a valid `approved: false` decision, which
+is ordinary editorial processing. Valid approvals continue through the remaining
+required validators and apply-time checks; this feature grants no bypass.
+
+The mapper returns one internal outcome for every requested candidate and no
+outcome for an extraneous index. Preserve strict runner cardinality checks so
+application bugs are still detected. Do not convert genuine internal invariant
+failures, cancellation, or fatal setup failures into tolerated model defects.
+
+## Outcomes and Reporting
+
+Keep existing work-unit aggregation and operation accounting:
+
+- If any requested candidate lacks an unambiguous usable decision, record a
+  validation-operation error identifying the affected requested indices. Other
+  candidates' usable decisions from that operation remain effective. An operation
+  error does not mean every candidate in its batch failed.
+- If all requested candidates have usable decisions and extra entries were
+  ignored, record a successful operation with a recovery for that handled defect.
+  This remains a recovery even if all legitimate decisions are editorial rejections.
+- If missing/duplicate requested decisions and ignored extras coexist, retain the
+  operation error. Follow the existing rule that failed operations do not also
+  produce recovery records; retain extra-entry details in debug evidence.
+- A valid complete response with no handled defect adds neither error nor recovery.
+  Valid editorial rejection alone never adds either category.
+- Preserve actual structural repair and service-retry counts. On successful
+  operations, consolidate applicable recovery reasons in one recovery record.
+  Do not count reconciliation as another LLM operation or structural repair.
+
+Error attribution must distinguish affected candidates from the rest of the batch.
+Debug evidence should preserve supplied decisions, ignored entries, and the reason
+an expected candidate became unavailable. Do not fabricate candidate ledger entries
+for unknown indices or mislabel missing decisions as deliberate editorial rejection.
+Ordinary reports must retain safe metadata without raw transcript or response text.
+
+Use a decision-specific recovery reason rather than treating ignored validator
+entries as defective correction proposals. Concrete reason names and minimal data
+model changes belong in the implementation plan. Keep YAML, Markdown, and CLI
+counts consistent, with the existing reporting-disable and independent debug options.
+
+## Ownership and Target Architecture
+
+Audita's validator layer reconciles structurally valid model responses into
+complete internal candidate outcomes. The runner preserves strict internal
+cardinality and advances only approved candidates. Reporting consumes operation
+facts without treating partial validation failure as loss of every decision.
+Debug evidence distinguishes unavailable validation from editorial rejection.
+
+The schema/transport boundary remains with PromptKit; reconciliation needs no
+additional model call. Concrete helper structures, reason names, mapping changes,
+and verification steps belong in the [implementation plan](implementation.md).
+
+## Scope Boundaries
+
+PromptKit's whole-response structural validation and repair remain unchanged.
+A malformed entry, including an extraneous entry that violates the schema, may
+still invalidate the response and exhaust structural repairs before reconciliation
+can run. Ignoring schema-valid extras does not promise salvage from malformed JSON,
+missing required fields, wrong types, or out-of-range values.
+
+General per-entry structural salvage is a separate design question, not part of
+this fix. Likewise, guided regeneration remains deferred for reconsideration after
+this change; no new retries, feedback prompts, budgets, upstream dependency changes,
+or operator controls are required here.
+
+Preserve module scope, literal replacement behavior, fixed chunks, stage barriers,
+span-deletion safeguards, backend capacity limits, and fatal cancellation behavior.
+No additional restrictions on valid corrections or guarantees about model judgment
+are introduced. This work does not itself authorize tagging or publishing v2.0.0.
+
+## Acceptance Criteria
+
+- A schema-valid response can retain usable decisions despite missing, duplicate,
+  or unexpected indices elsewhere in the same batch.
+- Only the affected requested candidates become unavailable; extras never create
+  candidates or attach decisions to a different batch's candidates.
+- Valid negative decisions remain editorial outcomes, and every applied correction
+  completes all required validation.
+- Internal validator results satisfy existing strict cardinality checks.
+- Missing/duplicate requested decisions yield accurately attributed errors without
+  losing valid work. Extra-only defects yield a consolidated recovery; mixed
+  defects follow existing failed-operation reporting semantics.
+- Final transcript, debug evidence, YAML reports, Markdown counts, and CLI summary
+  agree about applied work and failures/recoveries.
+- Reconciliation makes no additional provider calls. Existing structural failures,
+  provider failures, and cancellation retain their documented handling.
+
+## Open Questions
+
+No unresolved product-policy questions block implementation. Guided recovery and
+structural salvage remain separate future decisions.

docs/roadmap/future.md

@@ -0,0 +1,80 @@
+# Deferred Feature Work
+
+**Last updated:** 2026-09-25
+
+**Status:** Deferred or exploratory work; not implemented or scheduled. These
+entries preserve scope outside the [resilience roadmap](resilience.md) and
+[span-deletion consistency roadmap](span-deletion.md). Each entry states its
+intent; inclusion does not commit us to implementing every capability.
+
+## Domain-Aware Response Recovery
+
+The [partial-success decision reconciliation](decision-set-reconciliation.md)
+preserves usable decisions without another model call. Guided recovery remains
+deferred pending evidence from failures that remain. General per-entry
+salvage from structurally invalid responses is also a separate design question;
+the accepted reconciliation feature retains whole-response structural validation.
+
+Consider bounded feedback and regeneration for model responses that satisfy the
+structural schema but leave requested decisions unavailable after reconciliation.
+The goal would be to recover more usable work before excluding affected
+candidates, when doing so materially improves outcomes.
+
+This capability is explicitly deferred. Current behavior retains PromptKit
+structural repair and per-candidate reconciliation; unavailable requested
+decisions produce errors while other candidates can succeed.
+
+Before implementation, define the Audita/PromptKit ownership boundary, which
+checks are eligible for feedback, how feedback identifies defects without exposing
+secrets, and how regeneration fits bounded budgets and operation deadlines. Reuse
+shared library capabilities where appropriate; do not introduce an uncoordinated
+outer retry loop that multiplies existing retry/repair budgets.
+
+Ordinary editorial rejection is a valid decision, not a reason to regenerate until
+an edit is approved. Recovery must preserve required validation, candidate identity,
+partial-success safety, and truthful error/recovery reporting. Validate with
+controlled responses that distinguish recoverable protocol defects from valid
+negative decisions. Exact APIs, budgets, and sequencing remain undecided.
+
+## Independent Whole-Pipeline Traversal by Chunk
+
+Explore allowing each fixed chunk to advance through its ordered stages without
+waiting for other chunks to finish the current stage. This could reduce waiting
+behind a slow chunk and improve utilization while preserving stage order within
+each chunk and deterministic final recombination.
+
+The resilience feature deliberately retains stage barriers: finish a stage across
+chunks before advancing to the next stage. Fixed boundaries alone do not authorize
+independent traversal.
+
+A future design must establish that stages do not depend on other chunks' evolving
+state, preserve stable identities and partial-success rules, and continue to share
+PromptKit's per-backend capacity across all roles and stages. Reporting must remain
+deterministic despite different completion orders, and fatal cancellation must
+still stop and join active work. Assess whether the scheduling change materially
+improves real workloads before introducing additional orchestration machinery.
+No new scheduler, global limiter, or execution mode is selected by this entry.
+
+## Whole-Segment Removal
+
+Potential future work, subject first to evaluating whether it is worth implementing
+at all. Empty replacement text and deletion of a span within a surviving segment
+are covered by the accepted [span-deletion consistency roadmap](span-deletion.md).
+That feature does not remove segment records or permit blank resulting segments.
+
+Before designing segment removal, review representative production examples to
+establish whether whole-segment deletion is a meaningful unmet need. An empty
+replacement alone does not establish that intent: it can express ordinary filler
+or punctuation deletion within a larger segment. Assess how often genuinely
+expendable whole segments occur after deterministic normalization, the transcript
+quality benefit of removing them, and the risk of losing substantive speech.
+Compare those benefits with the implementation and ongoing maintenance costs.
+Retaining the segment and declining the edit is a valid outcome if the benefit is
+small or uncertain. No numerical adoption threshold is selected.
+
+Only if that evaluation supports proceeding should a separate feature design
+choose a representation, authorization and validation rules, and reporting
+semantics. It must address meaning preservation and the effects on segment
+identity, timestamps, fixed chunks, subsequent stages, recombination, and output
+consumers. Do not infer authorization to remove a segment from an empty string
+or introduce segment-removal machinery as part of span-deletion consistency.

docs/roadmap/implementation.md

@@ -0,0 +1,287 @@
+# Implementation Plan: Partial Validator Decision Sets
+
+**Last updated:** 2026-09-25
+
+**Status:** Implemented; all stages completed.
+**Source baseline:** `60e2fe7`.
+**Audience:** gpt-5.6-terra coding agent.
+
+## Scope and Execution
+
+Implement [partial success for validator decision sets](decision-set-reconciliation.md).
+That roadmap owns product intent and target semantics; this document selects the
+implementation and verification work. Follow Stage 1 through Stage 3 in order and
+record completion evidence beneath each stage. Each stage is a focused agent task.
+
+Read [development guidance](../development.md) and all policies under
+[docs/policy](../policy/): [agents](../policy/agents.md),
+[architecture](../policy/architecture.md), [documentation](../policy/documentation.md),
+[testing](../policy/testing.md), and [release](../policy/release.md). Use graph-first
+code discovery, preserve unrelated work, and commit completed coherent units with
+plain English messages. Do not introduce approval rounds between stages.
+
+Aim directly at the final implementation. No temporary modes, compatibility shims,
+new configuration, extra model calls, or upstream dependency work is required.
+Keep PromptKit structural validation/repair and the runner's internal cardinality
+checks. Implement the agreed outcomes, not extra restrictions suggested by existing
+tests or incidental internal structure. Internal helper signatures and layout can
+be adjusted idiomatically while preserving the decisions below.
+
+This replaces the completed span-deletion plan, available in Git at
+`60e2fe7:docs/roadmap/implementation.md`. It does not reopen that feature. The final
+stage includes correcting historical links that otherwise point to this new plan.
+No release preparation, tag, push, CI wait, or paid-provider evaluation is included.
+
+## Selected Implementation Decisions
+
+### Reconciliation boundary
+
+Keep reconciliation in `internal/framework/validators`, after successful structured
+completion and before converting the batch into runner-facing results. Replace
+`mapLLMResponseToDecisions` with a small pure reconciliation helper; use an ordinary
+private result struct containing:
+
+- `Decisions []Decision`: one outcome per requested item, in request order;
+- `UnavailableIndices []int`: requested indices without a usable verdict;
+- `IgnoredIndices []int`: unexpected indices encountered in response order, including
+  repeated unexpected entries, for debug attribution.
+
+No generic reconciliation framework or new shared abstraction is needed. Requested
+identities come from `batch.Items`, not transcript segment IDs, the whole section,
+or any other batch. Expected input identities remain an Audita internal contract.
+
+First collect response entries by requested index and count occurrences, separating
+unexpected indices. Then walk requested items once to construct outcomes. This
+avoids accepting an early duplicate before seeing a conflicting later entry.
+
+Use these candidate-level reason codes in `validators/models.go`:
+
+| Condition | Internal outcome/reason |
+| --- | --- |
+| Missing requested entry | `DispositionUnavailable`, `missing_validation_decision` |
+| More than one entry for a requested index | `DispositionUnavailable`, `duplicate_validation_decision` |
+| Single requested entry failing the existing confidence range check | `DispositionUnavailable`, `invalid_validation_decision` |
+| Single usable entry | Existing approved or `llm_rejected` decision with its model reason. |
+
+Unavailable outcomes have `Approved: false` and fixed safe explanatory messages.
+Duplicate classification takes precedence over inspecting individual duplicate
+values; identical duplicates follow the same rule. Unexpected entries are ignored
+before inspecting confidence. Retain the existing confidence check as a local
+check for decoded callers/test doubles, scoped to that candidate; do not broaden
+this into structural salvage, confidence thresholds, or new semantic validation.
+The production JSON Schema still rejects out-of-range confidence before mapping.
+
+An empty response list makes every requested item unavailable. Response ordering
+must not change verdicts or which candidate is affected. Preserve the existing
+handling of valid negative decisions, including reason text. No candidate may
+advance merely because another candidate was approved.
+
+### Operation accounting and ordinary reports
+
+Each actual validation batch remains one `outcomes.Operation`. Use existing
+`ReasonValidationDecisionsInvalid` (`validation_decisions_invalid`) for a batch
+with unavailable requested candidates. Set `Succeeded: false`, set `ProposalIndices`
+to only those unavailable indices, retain execution counts, and leave recovery
+reasons empty. Still return usable decisions from that batch alongside explicit
+unavailable outcomes, with no Go error for the reconciled model defect.
+
+If no requested candidates are unavailable, set `Succeeded: true`. Add
+`outcomes.RecoveryValidationDecisionsFiltered` with serialized value
+`validation_decisions_filtered` when extras were ignored. Merge it with actual
+`output_repaired` and `service_recovered` reasons in the same recovery operation.
+Valid negative decisions do not prevent operation success. A clean response creates
+no new recovery reason.
+
+When unavailable candidates and extras coexist, record the failed operation only;
+keep ignored-entry evidence in debug. Do not create both an error and a recovery
+for that operation. Successful survivors must continue even though the operation
+itself records a failure. Retain actual execution counts on all paths.
+
+Reuse existing ordinary YAML structures and version 1. The only new public report
+value is the decision-specific recovery reason; affected error indices already fit
+`Failure.ProposalIndices`. Do not add ignored entries to `Operation.Filtered` or
+`filtered_proposals`, which describe correction proposals, and do not increment
+`proposal_responses_with_filtering`. The recovery reason and existing operation
+identity suffice for ordinary reporting; detailed extra indices belong in debug.
+
+Update `processreport.safeMessage` for `validation_decisions_invalid` to explain
+that some required decisions were unavailable without claiming all batch decisions
+were discarded. Generalize the recovery message to include response filtering,
+rather than only proposal filtering. Exact prose is not a test contract. Preserve
+work-unit error grouping, operation counts, and existing YAML/Markdown/CLI totals.
+
+### Debug evidence and runner records
+
+Existing interaction diagnostics already retain the full decoded response and final
+content. Add the reconciliation result to the response diagnostic payload for a
+successfully decoded validation response. Perform reconciliation before writing
+that diagnostic; retain existing call-error diagnostics and diagnostic-write failure
+handling. All reconciled decisions, including unavailable outcomes, receive the
+returned response artifact path. This records unexpected indices and duplicates
+without inventing candidate ledger entries or an additional debug file.
+
+Carry `validators.DecisionDisposition` through `runner.ValidatorDecisionRecord`
+and `runner.ValidatorRejectedChange` in an additive `disposition` field. Populate
+it from each actual `Decision`, including partial decision evidence returned on
+fatal paths. Keep the existing collection names and application flow; no parallel
+unavailable-candidate collection is necessary.
+
+In `processreport`, add `disposition` to `LedgerValidatorDecisionRecord` and map
+unavailable candidates to a top-level ledger disposition `unavailable`, instead
+of `rejected`. Ordinary editorial/deterministic rejections remain `rejected`.
+Use the explicit disposition, not a reason-code allowlist, so existing timeout,
+provider-failure, and oversized-input outcomes retain the same distinction too.
+Update affected test fixtures to supply dispositions where they represent real
+validator outputs. Do not invent a fatal check or compatibility subsystem for a
+zero-value disposition in handcrafted fixtures.
+
+Retain candidate identity, literal proposed replacements (including explicit empty
+strings), and omission of applied replacements for unapplied candidates. Existing
+fatal/interrupted/module-failure handling and debug-off behavior remain. Supplied
+model decisions remain visible in interaction evidence even when reconciliation
+cannot accept them as authoritative verdicts.
+
+## Stage 1 — Reconcile Model Decisions and Preserve Valid Work
+
+**Completed:** 2026-09-25. Reconciliation preserves usable decisions, attributes
+unavailable requested indices and ignored extras, and keeps strict runner
+cardinality. Validator and runner boundary checks passed with
+`go test ./internal/framework/validators ./internal/framework/runner`.
+
+**Work**
+
+1. Read the validator models, `llm_validators.go`, existing partial-batch tests,
+   `outcomes/outcomes.go`, and `runner.validateSectionCandidates`.
+2. Implement the reconciliation result/helper and candidate reason codes selected
+   above. Replace the whole-batch mapping-error branch with outcome construction
+   and correct batch accounting. Add the decision-specific recovery constant.
+3. Move successful-response diagnostic assembly after reconciliation and include
+   its result. Preserve raw/decoded response evidence and diagnostic paths; keep
+   recoverable call errors, fatal errors, and cancellation on their existing paths.
+4. Keep `EnforceDecisionCardinality` and `EnforcePartialDecisionCardinality` strict.
+   Ensure final validator decisions are complete and unique internally. The runner
+   should continue passing only approved candidates onward without treating an
+   operation's failure metadata as a reason to discard valid decisions.
+5. Replace obsolete batch-rejection expectations in `llm_validators_test.go` and
+   relevant runner tests. Preserve cardinality tests that exercise broken internal
+   validators. Update the repaired-but-domain-invalid tests so actual unavailable
+   cases remain errors with counts retained and no recovery record.
+
+**Focused verification**
+
+Use a small table at the validator boundary for mixed missing/extra indices,
+conflicting and identical requested duplicates, extra-only responses, empty lists,
+reordered responses, and a single invalid-confidence decoded test-double entry.
+Include noncontiguous expected indices and valid negative decisions. Assert semantic
+outcomes and affected indices, not helper layout or exact messages. Include one
+controlled case combining successful structural/service recovery facts with ignored
+extras and one failed reconciliation retaining counts without recovery reasons.
+
+At the runner boundary, use the real LLM-backed validator with a controlled client
+and a later validator: prove a survivor reaches later required validation, a later
+rejection prevents application, and unavailable candidates never reach that later
+validator as approved. Use successful proposal operation facts where accounting is
+asserted; do not check an empty operation list. Reuse existing fixtures/tests rather
+than creating a second framework or duplicating every table case.
+
+Run `go test ./internal/framework/validators ./internal/framework/runner`.
+
+**End state:** Usable decisions survive model indexing defects, strict internal
+cardinality remains satisfied, and operation facts explain partial failure or
+successful extra filtering. No extra model generation is introduced.
+
+## Stage 2 — Carry Dispositions Through Debug and Report Partial Outcomes
+
+**Completed:** 2026-09-25. Runner and ledger records carry explicit
+dispositions; ordinary reports distinguish unavailable decision errors from
+extra-index filtering recoveries. Focused report, ledger, runner, and YAML
+checks passed with `go test ./internal/framework/runner
+./internal/framework/processreport ./internal/core/reporting`.
+
+**Work**
+
+1. Add and populate disposition fields through the runner and ledger as selected
+   above. Map unavailable candidates distinctly and preserve empty replacement
+   serialization. Check both normal and fatal partial-evidence paths.
+2. Update `processreport` messages and review `Build`, `mapFailure`, and `mapRecovery`.
+   Reuse their existing aggregation; change only assumptions that would misreport
+   the new operation facts. Core YAML structures and Markdown table layout remain.
+3. Extend existing report/ledger tests for an operation with surviving work plus
+   unavailable indices, extra-only recovery, and a mixed defect operation. Verify
+   error indices exclude survivors, extra filtering does not count as proposal
+   filtering, and failed operations do not also generate recovery records.
+4. Verify debug ledger entries distinguish genuine rejection from unavailability,
+   and interaction diagnostics preserve ignored entries and reconciliation reasons.
+   Unknown response indices must not create fake ledger candidates. Retain existing
+   redaction and optional-debug-failure tests rather than reproducing them all.
+5. Update `docs/integrations/output-files.md` for the new recovery reason and debug
+   disposition, and `docs/internal/diagnostics-reporting.md` for the mapping changes.
+   Explain that validation failure indices identify unavailable candidates, not
+   necessarily every candidate submitted to that batch.
+
+**Validation**
+
+Run `go test ./internal/framework/runner ./internal/framework/processreport ./internal/core/reporting`.
+Use semantic YAML/summary assertions; no whole-report prose golden, exact message
+snapshot, new schema version, or new rendering mode is needed.
+
+**End state:** Reports and debug evidence accurately describe partial validation
+success and handled extras through existing reporting mechanisms.
+
+## Stage 3 — Verify the Assembled Workflow and Finalize Documentation
+
+**Completed:** 2026-09-25. Controlled subprocess runs covered a mixed decision
+response and an extra-only response through the CLI, transcript, YAML, Markdown,
+debug ledger, and interaction evidence. Fixture execution facts showed one
+provider attempt and zero structural repairs per logical call. `go test
+./cmd/audita` and `go test ./...` passed. Canonical validator guidance and
+roadmap status were updated; the two span-deletion plan links now target the
+historical plan at `60e2fe7`.
+
+**Work**
+
+1. Extend the existing local-provider subprocess fixtures in
+   `cmd/audita/main_integration_test.go`. Add a representative mixed-decision run
+   with an applied survivor, a legitimate rejection, a missing requested decision,
+   a duplicate requested decision, and an extra index. Return responses by the
+   requested correction identities; do not freeze incidental validator call order.
+   Ensure the survivor passes all remaining required validation. Assert successful
+   CLI completion, correct transcript, one work-unit error with only unavailable
+   indices, no reconciliation recovery for the failed operation, and debug
+   evidence distinguishing rejection/unavailability without unknown candidates.
+2. Add or adapt a compact extra-only case proving usable work survives and one
+   `validation_decisions_filtered` recovery appears. Verify YAML, Markdown, and CLI
+   agree on totals using semantic assertions. Keep proposal-filtering counts separate.
+   Existing reporting-disable/debug selection tests remain the owners of those options.
+3. In these fixtures give PromptKit its normal configured structural repair allowance,
+   but supply schema-valid responses. Verify reconciliation adds no generations or
+   structural repairs using fixture requests/execution facts, without encoding today's
+   total validator count as a permanent contract. Keep structural-malformation tests
+   and recoverable batch-failure tests; do not relax the schema to make fixtures pass.
+4. Update `docs/internal/validators.md` and any affected canonical descriptions found
+   by searching for whole-batch rejection claims. Explain tolerant external response
+   reconciliation versus strict internal cardinality. The structural-validation
+   boundary remains explicit, and guided recovery remains deferred in `future.md`.
+5. Mark the feature roadmap implemented only after the assembled checks pass, and
+   record actual completion evidence in this plan. Repair the two span-deletion
+   roadmap links to the old plan with the historical URL
+   `https://gitea.maximumdirect.net/eric/audita/src/commit/60e2fe7/docs/roadmap/implementation.md`.
+   Preserve historical release notes and unrelated roadmaps.
+
+**Validation**
+
+Run `go test ./cmd/audita` while developing the fixture, then `go test ./...` once
+for the completed implementation. Run `git diff --check` and verify affected
+Markdown links. Do not add live-provider checks, a platform matrix, repeated full
+suite runs, or release gates solely for this feature. No code is implemented by
+this planning pass; these commands are instructions for the implementation sprint.
+
+**End state:** The accepted partial-success behavior is exercised through the real
+CLI/library boundary, canonical documentation matches it, and coherent changes
+are committed for subsequent release consideration.
+
+## Open Questions
+
+None required before implementation. The implementation decisions above preserve
+the agreed product scope. General structural salvage and guided regeneration remain
+separate future decisions, not unresolved prerequisites for this work.

docs/roadmap/library-adoption.md

@@ -0,0 +1,129 @@
+# Coordinated ConfigKit and PromptKit Adoption
+
+**Last updated:** 2026-09-24
+
+**Status:** Production implementation complete. ConfigKit v0.4.0 and PromptKit
+v0.11.0 are adopted; [historical implementation evidence](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md) records the
+completed checks. The current `implementation.md` covers span deletion instead.
+Release compliance and publication remain separate work.
+
+## Goal and Ownership
+
+Adopt ConfigKit and PromptKit to align Audita with the application suite's shared
+configuration and LLM execution conventions. Consistency is the primary benefit;
+removing overlapping application infrastructure is a secondary benefit. Preserve
+Audita's transcript-correction capabilities and domain protections.
+
+| Owner | Responsibilities |
+| --- | --- |
+| ConfigKit | Strict application configuration loading, presence/default semantics, explicit overrides, recorded resolution, secret references, and safe inspection |
+| PromptKit | Prompt/profile formats and rendering, provider calls, structural output validation and repair, transient retries, engine-local backend admission, and execution observations |
+| Audita | Setting meanings, file/CLI selection, role/profile association, normalization/chunking, ordered modules, domain validation, correction application, reports, diagnostic redaction, and artifact lifecycle |
+
+The [configuration roadmap](configkit.md) owns configuration and inspection policy.
+The [boundary findings](library-boundary-findings.md) own the selected defaults,
+full migration inventory, failure mapping, and prototype evidence. The selected
+production versions are ConfigKit v0.4.0 and PromptKit v0.11.0, with Go 1.25.5 as
+the minimum. Use published modules, never sibling checkouts or local replacements.
+The existing YAML dependency remains while glossary parsing needs it.
+
+## Migration Posture
+
+This is an intentional breaking transition targeting Audita v2.0.0. No backward
+compatibility is required for existing installations, config files, CLI flags,
+environment overlays, or inspection output. Provide one migration guide, without
+legacy loaders, aliases, dual modes, or automatic conversion tooling.
+
+Both libraries belong to one refactor. Intermediate commits need not build or run;
+there is no supported intermediate product state. No temporary adapters, compatibility
+shims, or processing gates are needed. The complete application must work and pass
+its checks before the refactor is finished. Release compliance and publication
+come last, with no intermediate public release.
+
+## Configuration and Execution Target
+
+ConfigKit selected overlays enable a testing configuration that chooses cheaper or
+faster models. The example selects both proposal and validation profiles. ConfigKit
+application overlays and PromptKit execution profiles remain separate concepts.
+Complete operator-maintained profiles own model/generation controls; embedded
+application prompts and schemas are not operator-selected files. No catalog wiring,
+profile-inheritance feature, ConfigKit imports, or secret-directory bootstrap is
+part of this migration.
+
+Capacity controls serve two needs: respecting a local service's hard simultaneous-
+query limit and bounding concurrent paid API activity. One finite limit per backend
+is shared across both roles and every initial, retry, and repair attempt. Limits
+are engine-local, not coordinated across processes. No additional global or
+per-role limiter or fine-grained scheduling policy is required. Profiles must use
+registered backends and ConfigKit-resolved credentials; endpoint-only bypass and
+competing profile credential lookup are excluded.
+
+Structural repair is enabled by default for both roles because additional attempts
+to obtain valid output are desired. Transient retries and structural repairs have
+separate budgets; explicit zero disables the respective behavior. Each generation,
+including a repair, receives its own transient-retry budget. The Audita logical-call
+deadline covers preparation, waiting, backoff, all generations, and repairs; profile
+HTTP timeouts apply per attempt. Exact defaults and bounds are in the
+[selected contract](library-boundary-findings.md#configuration-and-profile-contract).
+No Audita outer retry loop restores the old broader retry policy.
+
+Audita retains all four correction modules, five LLM-validator prompts,
+deterministic validators, confidence/applicability checks, glossary protection,
+validation batching/cardinality, deterministic proposal and ledger ordering,
+transcript encoders, and artifact retention. PromptKit-valid JSON still passes
+Audita's domain checks. Repair exhaustion, provider failure, and domain rejection
+follow the [selected failure table](library-boundary-findings.md#execution-and-failure-mapping).
+Retention and transcript timestamp changes remain separate work; this migration
+resolves the [retry-count follow-up](runtime-behavior.md).
+
+## Reporting and Diagnostic Target
+
+Keep stable Audita-owned serialized records; do not expose raw library structs.
+Use process report schema `v2` with the existing domain outcomes and a new
+`llm_summary` separating logical calls, dispatched provider attempts, transient
+retries, structural repairs, and usage. Transcript schema versions remain unchanged.
+The implementation plan specifies the concrete new fields and artifact wiring.
+
+Usage distinguishes known zero, unknown, and incomplete totals. Preserve observations
+from failed calls even when no final result exists. Count terminal totals once;
+attempt observations provide diagnosis without being added to those totals again.
+This is operational evidence, not a complete billing ledger.
+
+Replace global/role scheduler utilization with per-backend configured capacity and
+observed queue/generation/retry/logical-call timing. Do not fabricate measurements
+that the event API does not provide. Keep domain timing and deterministic correction
+ledger ordering. Record prompt definition/rendered hashes and selected profile/model
+without promising the old hash values. Callbacks are short, concurrency-safe, and
+free of blocking I/O or waits on their own execution.
+
+Effective configuration and selected-profile snapshots remain distinct safe views.
+Secret values must not escape through inspection, logs, reports, runtime errors,
+or retained diagnostic files. Observation records exclude payloads, URLs, raw
+provider messages, prompts, and credentials. Existing artifact lifecycle and
+partial-failure handling remain Audita responsibilities.
+
+## Completion Criteria and Release Boundary
+
+- Both libraries own their selected responsibilities in the production application.
+  Superseded loaders, transport, renderers, schedulers, aliases, and the isolated
+  prototype are gone once its useful evidence is represented in production tests.
+- Offline tests prove configuration-to-execution routing for both roles, selected
+  overlays and explicit overrides, shared backend capacity, retry/repair outcomes,
+  cancellation/deadlines, domain decisions, redaction, and truthful partial reports.
+  Tests exercise real internal collaborators and controlled HTTP boundaries.
+- Maintained examples and canonical docs describe the implemented final model.
+  One migration guide covers files, flags, environment inputs, profiles, credentials,
+  overlays, inspection, and changed runtime/report/diagnostic semantics.
+- The assembled application passes its required checks; Linux amd64 and arm64 build.
+  macOS arm64 remains best effort. No live-provider or CI-success gate is added.
+
+PromptKit v0.11.0's reviewed retry, observation, and capacity APIs remove the earlier
+upstream feature prerequisites; no upstream edit is required for adoption. The
+prototype proved representative behavior, not release compliance or model quality.
+
+The selected PromptKit release uses LGPL-3.0-or-later. A later release unit must
+review the actual production dependency graph and distribution requirements,
+including notices and source/rebuild materials, before binary publication. This
+implementation plan does not claim that review is complete. Release publication
+uses the existing [procedure](../release.md) and completes on successful source/tag
+push, without waiting for CI or artifact delivery.

docs/roadmap/library-boundary-findings.md

@@ -0,0 +1,392 @@
+# Shared Library Boundary: Decisions and Evidence
+
+**Status:** Boundary and offline prototype completed, 2026-09-23. Selected prototype contract and proposed
+production migration target; not the current Audita configuration contract.
+The [bounded roadmap](library-boundary.md) owns scope and completion criteria.
+
+**Subsequent scope decision:** The [ConfigKit overlay plan](configkit.md#selected-overlays-for-testing)
+adds optional selected overlays, with v0.4.0 as the migration baseline. It supersedes
+the original no-overlay decision and adds an input layer between the main file and
+explicit setting overrides. The v0.3.1 prototype results below remain historical
+evidence; they do not establish overlay integration.
+
+## Selected Boundary
+
+The former prototype is retained in Git history at `590604b`; it was a separate
+Go module using published ConfigKit v0.4.0, PromptKit v0.11.0, and
+Audita v1.1.0. The Audita dependency lets this consumer reuse actual internal domain
+helpers through its Audita module namespace without a local replacement. It does
+not make those internal packages a new public API. Remove this dependency along
+with the prototype after equivalent production integration tests landed.
+
+The initial prototype results below were recorded with ConfigKit v0.3.1. On
+2026-09-24, upgrading only its ConfigKit pin and sums to v0.4.0 required no source
+changes. With Go 1.26.8 on linux/amd64, the nested offline `go test ./...` passed
+and `go run .` still reported one accepted correction, four attempts, zero
+transient retries, and two structural repairs. This verifies the dependency
+upgrade, not selected-overlay behavior.
+
+ConfigKit owns loading, presence, explicit overrides, secret resolution, and
+redacted effective settings. PromptKit owns profile parsing, rendering, provider
+execution, transient retries, structural repair, and backend admission. Audita
+owns file selection, setting meanings, role selection, domain acceptance, and
+mapping execution facts into reports. Production CLI and runtime remain unchanged.
+
+## Configuration and Profile Contract
+
+Selected for the prototype and recommended for production:
+
+- Main YAML requires integer `version: 2`; absent, null, unknown fields, and older
+  versions fail. No legacy parser, aliases, environment overlays, or automatic
+  conversion. Defaults precede the selected file; explicitly supplied CLI settings
+  become ConfigKit overrides. Validation runs on the resolved result, so an override
+  may repair an invalid lower-precedence setting when decoding succeeds.
+- Production file selection: explicit `--config`, then
+  `/usr/local/etc/audita/config.yml`, then `/etc/audita/config.yml`. Select one file;
+  do not merge them. Missing system files are skipped; explicit missing or empty
+  paths and inaccessible selected files fail. No selected file fails with a request
+  to configure backend/profile selection: no hidden provider or model default.
+  Remove `AUDITA_CONFIG`. The prototype loads its explicit fixture only; discovery
+  and command parsing belong to the ConfigKit implementation stage.
+- `profiles_dir` defaults to `profiles`, resolved relative to the selected main
+  configuration file, with absolute paths accepted. PromptKit reads that directory.
+  Prompts and schemas are application-embedded, not operator-selected prompt files.
+  The prototype uses complete application-owned profile files. Catalog profiles and
+  profile inheritance are not needed for this migration; do not add catalog wiring.
+- `backends.<id>.endpoint` is required. Use one application backend identity per
+  service capacity pool. Do not split proposal and validation into different backend
+  IDs when they use the same service. Limits are engine-local, not shared across
+  multiple Audita processes or other applications.
+- `backends.<id>.concurrency` defaults to **1**, must be positive, and is shared by
+  both roles and all initial/retry/repair attempts. Operators can select **4** for
+  the described local backend or a suitable finite paid API limit. There is no
+  additional global or per-role limiter. `queue_capacity` defaults to **1024**,
+  following PromptKit; non-negative, with zero meaning no waiting admission.
+  This is an operational escape hatch rather than a second scheduling hierarchy.
+- `backends.<id>.credential` is an optional ConfigKit secret reference: a scalar
+  environment variable name. Omission or null means no credential. A supplied
+  reference must resolve to a nonempty value, including during validation and
+  inspection. There is no direct key field or CLI key flag. ConfigKit resolves it
+  once and Audita supplies `RunRequest.APIKey`. Profiles must not introduce another
+  `api_key_env` lookup. Backend credentials also support unauthenticated local use.
+- Profiles own model, temperature, output token limit, top-p, per-generation timeout,
+  service tier, reasoning effort, and extra provider parameters using PromptKit's
+  format. They select the application `backend`; their effective endpoint must
+  equal that backend's endpoint. Endpoint-only execution is rejected because it
+  bypasses the requested finite capacity boundary. A named registered backend is
+  mandatory even if both roles use the same profile.
+- `llm.proposal_profile` is required. Omitted `llm.validation_profile` inherits the
+  final proposal profile as a whole, with ConfigKit recording the derivation.
+  Explicit empty or null values fail; no individual model/endpoint/key inheritance.
+- `llm.transient_retries` defaults to **2** additional attempts per generation;
+  explicit zero disables retries; negative is invalid. PromptKit owns which failures
+  are eligible. Do not add an Audita retry loop.
+- `llm.repair_attempts` defaults to **2** additional generations for both roles;
+  explicit zero disables repair when deliberately requested. PromptKit supports
+  0–3. Positive defaults are exercised in the prototype. Audita owns this common
+  setting and supplies a complete request output contract with the role's embedded
+  schema, avoiding separate user-facing repair defaults in each prompt.
+- `llm.operation_timeout` defaults to **15m**, must be positive, and uses a Go
+  duration string. It starts before each logical proposal/assessment call and covers
+  preparation, capacity waiting, transient backoff, generation, and repairs. It is
+  separate from whole-process cancellation. Profile `timeout_seconds` defaults to
+  PromptKit's **600** seconds and is applied afresh to each built-in HTTP attempt.
+  It does not bound the combined retry/repair sequence. A stricter caller/transport timeout can
+  also end execution. The two defaults are selected starting points, not throughput
+  measurements. With the default T=2 and R=2, each logical call dispatches at most
+  **9** attempts before deadline/cancellation; proposal plus assessment may use 18.
+- Omission retains static defaults or invokes the recorded role/backend derivations.
+  Null is not a general synonym for omission: nonoptional scalars reject it. Optional
+  credentials and the future optional section target accept null. Explicit empty
+  module lists will mean an empty correction pipeline, not “restore defaults.”
+  Repeated modules retain their existing ordered meaning.
+- The original prototype disabled ConfigKit imports, selected configuration layers
+  (then called configuration profiles), and secret-directory bootstrap. Selected
+  overlays are now verified under the linked overlay decision; the other two
+  features remain disabled. PromptKit execution profiles retain their meaning.
+
+Configuration validation/inspection must assemble and inspect both selected profiles
+without provider calls. ConfigKit's effective YAML describes only its loaded state;
+a separate selected-profile view shows effective model controls/backend selection.
+Do not imply either snapshot alone represents the full execution configuration.
+The prototype exercises the two APIs but does not implement production subcommands.
+Do not serialize raw profile/library objects as a permanent Audita report schema.
+
+## Migration Inventory
+
+This inventory covers the current configuration and process flags documented in
+[configuration](../config.md) and [CLI](../cli.md), including environment-only fields.
+“Retain” below means retain domain meaning in the new schema, not accept old files.
+All legacy environment overlays listed afterward are removed.
+
+| Current file field / setting | Current CLI flag(s) | Target and rationale |
+| --- | --- | --- |
+| File selection | `--config` | Retain explicit invocation flag; discovery described above. |
+| Transcript input | positional path | Invocation only; retain one transcript path. |
+| Glossary input | `--glossary` | Invocation only; retain required glossary path. |
+| Corrected output | `--output` | Invocation only; retain stdout fallback. |
+| Report output | `--report-json` | Invocation only; retain separate report file. |
+| `pipeline.modules` | `--modules` | Retain sequence/default `[glossary, homophones, glossary, spoken_word, grammar]`; explicit empty clears rather than inherits. |
+| `output.schema` | `--output-schema` | Retain encoder selection/default `bare-segments`. |
+| `llm.proposal.base_url`, `llm.validation.base_url` | `--base-url`, `--validation-base-url` | Replace with backend endpoint; remove per-role endpoint flags. |
+| Both `llm.*.model` | `--model`, `--validation-model` | Replace with profile model; new `--proposal-profile` / `--validation-profile` select whole profiles. |
+| Both `llm.*.api_key_env`, direct runtime keys | `--llm-api-key`, `--validation-llm-api-key` | Replace with backend ConfigKit secret reference; remove direct key flags. |
+| Both `llm.*.timeout` | `--llm-timeout-seconds`, `--validation-llm-timeout-seconds` | Replace with profile `timeout_seconds`, plus `llm.operation_timeout`; no duplicate timeout flags. |
+| Both `llm.*.max_retries` | `--max-retries`, `--validation-max-retries` | Replace with common `llm.transient_retries` and distinct `llm.repair_attempts`; remove misleading old flags. |
+| `concurrency.total_llm` | `--total-llm-concurrency`, `--llm-concurrency` | Replace with per-backend concurrency; remove global/alias flags. |
+| `concurrency.proposal_llm`, `concurrency.validation_llm` | `--proposal-llm-concurrency`, `--validation-llm-concurrency` | Remove role budgets; shared backend handles service capacity. |
+| Validation prompt budget (no old YAML field) | `--validation-max-prompt-tokens` | Add `validation.max_prompt_tokens`, default 2048, positive; retain explicit flag override. Audita owns batching. |
+| `chunking.target_sections` | `--target-sections` | Retain optional positive target; omission/null means automatic, zero invalid. |
+| `chunking.max_section_tokens` | `--max-section-tokens` | Retain 8192, positive. |
+| `chunking.min_section_tokens` | `--min-section-tokens` | Retain 2048, positive, no greater than maximum. |
+| `normalization.max_segment_gap` | `--normalize-max-segment-gap` | Retain 4.0 numeric seconds, non-negative; drop alternate YAML duration syntax for these domain numeric fields. |
+| `normalization.ellipsis_gap` | `--normalize-ellipsis-gap` | Retain 3.5 numeric seconds, non-negative. |
+| `normalization.max_segment_duration` | `--normalize-max-segment-duration` | Retain 60.0 numeric seconds, positive. |
+| `normalization.max_segment_tokens` | `--normalize-max-segment-tokens` | Retain 2048, positive. |
+| `thresholds.glossary` | `--glossary-confidence-threshold` | Retain 0.8 in [0,1]. |
+| `thresholds.homophones` | `--homophones-confidence-threshold` | Retain 0.8 in [0,1]. |
+| `thresholds.spoken_word` | `--spoken-word-confidence-threshold` | Retain 0.8 in [0,1]. |
+| `thresholds.grammar` | `--grammar-confidence-threshold` | Retain 0.8 in [0,1]. |
+| `context.description` | `--transcript-description` | Retain trimming and 500-byte limit; empty clears context. |
+| `diagnostics.work_dir` | `--work-dir` | Retain `/tmp/audita`, nonempty; artifact lifecycle remains Audita-owned. |
+| `diagnostics.retention` | `--work-dir-retention` | Retain `auto` default and selection; existing `never` follow-up is separate. |
+| Help/version/config commands | `--help`, `-h` | Retain command purpose; help should parse before loading configuration. Effective-config output intentionally becomes ConfigKit YAML, not legacy JSON. |
+
+Retained setting flags become explicit ConfigKit overrides only when supplied;
+CLI parser defaults must not overwrite file values. Normalization flag values stay
+numeric seconds. Domain values convert into existing domain structs after validation;
+do not spread ConfigKit or PromptKit types into correction algorithms.
+
+Complete environment removal map (names remain permissible if explicitly named as
+secret references; none have implicit configuration meaning):
+
+| Old environment names | Replacement |
+| --- | --- |
+| `AUDITA_CONFIG` | File selection above. |
+| `AUDITA_MODULES` | `pipeline.modules` / explicit flag. |
+| `AUDITA_LLM_API_KEY`, `OPENROUTER_API_KEY`, `AUDITA_VALIDATION_LLM_API_KEY` | Explicit backend credential references; no implicit fallback. |
+| `AUDITA_MODEL`, `AUDITA_VALIDATION_MODEL` | Profile files and role selection. |
+| `AUDITA_BASE_URL`, `AUDITA_VALIDATION_BASE_URL` | Backend endpoints. |
+| `AUDITA_LLM_TIMEOUT_SECONDS`, `AUDITA_VALIDATION_LLM_TIMEOUT_SECONDS` | Profile timeout and operation deadline. |
+| `AUDITA_MAX_RETRIES`, `AUDITA_VALIDATION_MAX_RETRIES` | Common transient/repair settings. |
+| `AUDITA_TOTAL_LLM_CONCURRENCY`, `AUDITA_LLM_CONCURRENCY`, `AUDITA_PROPOSAL_LLM_CONCURRENCY`, `AUDITA_VALIDATION_LLM_CONCURRENCY` | Backend concurrency; remove alias/role budgets. |
+| `AUDITA_VALIDATION_MAX_PROMPT_TOKENS` | `validation.max_prompt_tokens` / explicit flag. |
+| `AUDITA_TARGET_SECTIONS`, `AUDITA_MAX_SECTION_TOKENS`, `AUDITA_MIN_SECTION_TOKENS` | Corresponding chunking fields / flags. |
+| `AUDITA_NORMALIZE_MAX_SEGMENT_GAP`, `AUDITA_NORMALIZE_ELLIPSIS_GAP`, `AUDITA_NORMALIZE_MAX_SEGMENT_DURATION`, `AUDITA_NORMALIZE_MAX_SEGMENT_TOKENS` | Corresponding normalization fields / flags. |
+| `AUDITA_GLOSSARY_CONFIDENCE_THRESHOLD`, `AUDITA_HOMOPHONES_CONFIDENCE_THRESHOLD`, `AUDITA_SPOKEN_WORD_CONFIDENCE_THRESHOLD`, `AUDITA_GRAMMAR_CONFIDENCE_THRESHOLD` | Corresponding threshold fields / flags. |
+| `AUDITA_WORK_DIR`, `AUDITA_WORK_DIR_RETENTION` | Diagnostics fields / flags. |
+
+## Execution and Failure Mapping
+
+The executable uses grammar proposal plus grammar review, current response schemas,
+current hardening text, transcript-section/description helpers, proposal preview, and
+actual decision-cardinality enforcement from Audita v1.1.0. PromptKit renders the
+adapted templates; the custom HTTP transport supplies OpenAI-compatible responses.
+The prototype pre-expands Audita's `hardening` template function into the embedded
+PromptKit definitions; it passes existing named string variables unchanged. No
+provider, paid call, or injected high-level LLM client participates.
+
+Generated JSON is structurally validated/repaired by PromptKit before Audita decodes
+it. Audita then checks proposal applicability, confidence, and decision identity and
+cardinality. The demonstration previews corrections without mutation. Production
+retains every module's full validator chain, glossary protection, batching, and
+apply-time checks; the representative prototype does not claim to replace them.
+
+| Outcome | Section/batch effect | Overall run |
+| --- | --- | --- |
+| Structural repair succeeds | Continue domain validation. | Normal execution. |
+| Proposal repair exhausted | No proposals from this section; malformed-output warning. | Continue other work; successful completion may include warnings. |
+| Assessment repair exhausted | Reject this batch; malformed-output warning. | Continue other work. |
+| Schema-valid invalid/missing/duplicate decision indices | Reject batch and warn; no semantic retry/repair. | Continue other work. |
+| Inapplicable/low-confidence proposal or negative assessment | Reject affected proposal; preserve domain reasons. | Continue other work. |
+| Provider failure, retry exhaustion, transport or malformed HTTP response | Stop affected operation; retain available observations. | Runtime failure; cancel remaining work, retain existing Audita diagnostic handling. |
+| Deadline, caller cancellation, capacity admission failure | Stop affected operation; no outer retry or fabricated output. | Runtime failure with distinct classified cause. |
+| Invalid settings/profile/schema setup | Fail before processing where detectable. | Configuration error for user input, runtime error for broken embedded assets. |
+
+Zero transient retries produces one ordinary provider failure, not
+`ErrRetriesExhausted`; the latter signals an exhausted **positive** budget.
+Provider failures retain a typed cause. Do not expose raw provider messages through
+user diagnostics: continue Audita's separate secret-redaction boundary. A failed
+logical call may return no result while its observations still record completed
+attempts or an earlier successful generation. Already completed domain results can
+remain in diagnostics; failure must not be presented as a fully corrected output.
+
+## Observation and Report Handoff
+
+The prototype maps library events into Audita-owned records under a short mutex.
+Records contain logical run/role/backend/profile identity, generation and attempt
+indices, dispatch state, queue/generation/retry/wall timing, classified outcomes,
+HTTP status, exhaustion, repair/retry counts, and each usage field's value,
+knownness, and completeness. It omits URLs, credentials, payloads, raw provider
+errors, and prompt text. Identifiers themselves are metadata, not a secret channel.
+
+Use terminal records once for logical totals; retain attempt records for diagnosis
+without adding them to terminal totals. A known zero is distinct from unknown;
+partial totals remain incomplete after unmetered failed attempts. These observations
+are not a billing ledger. Callback code must not wait on its own run's progress.
+
+Production report changes must distinguish logical calls from dispatched provider
+attempts. Replace scheduler utilization fields based on global/proposal/validation
+pools with per-backend admission/active timing. Preserve deterministic ledger and
+validator ordering. Map prompt definition/rendered hashes and selected profile/model
+into request metadata, but do not promise equality with legacy prompt hashes or
+serialize PromptKit's raw event struct as the stable report contract. Update the
+output integration reference, report mapper tests, and downstream migration guide
+when the production representation is implemented.
+
+## Verification and Remaining Handoff
+
+The focused matrix is implemented in the nested module:
+
+| Roadmap scenario | Evidence |
+| --- | --- |
+| Successful assembled path | `go run .`; `TestWorkflowBoundaries/repair_both_roles`; actual Audita domain helpers and both selected profiles. |
+| Presence, overrides, inheritance | `TestConfigurationContract`: omitted role/backend values, explicit zero, null/empty, version, unknown keys, and explicit overrides. |
+| Credentials | Missing/unset and empty references fail; optional null accepted; effective/profile inspection redacted; HTTP authorization verified; no secret/prompt text in observations. `TestProfilesRespectSharedBackendBoundary` rejects competing lookup and capacity bypass. |
+| Both roles repaired | `repair_both_roles`: four attempts, two repairs, one accepted correction. |
+| Repair exhausted | Proposal and assessment exhaustion subtests: bounded counts and selected warning/rejection outcomes. |
+| Semantic boundary | Missing span, missing/duplicate decision indices, negative assessment; no semantic repair loop. |
+| Retry/zero/exhaustion | 429 recovery, zero-budget 503, positive-budget exhaustion after completed proposal; typed error assertions. |
+| Combined budgets | 503 → invalid JSON → repair 503 → valid proposal → assessment: five attempts, two transient retries, one repair. |
+| Shared capacity | `TestSharedCapacityAndRecovery`: caps four and two, both roles, a retry and repair, queue rejection, measured peak equal to cap, later successful call. |
+| Waiting/cancellation/deadline | `TestQueuedCancellationAndDeadline`; `TestDeadlineStopsRetryAfter`: no queued dispatch after cancellation/deadline and a 60-second provider retry delay interrupted by the operation deadline. |
+| Failed-run observations | `TestPartialUsageOnFailedRun`: valid HTTP output with known-zero usage followed by repair failure; nil result, retained events, incomplete known aggregate, correct counts. |
+| Representative assets | `TestRealPromptRenderingParity` compares both rendered message pairs to Audita v1.1.0; `TestSchemaFixturesMatchAudita` compares schema bytes. |
+
+Commands executed from `examples/library-adoption` on Linux amd64 with Go 1.26.8:
+
+```sh
+GOWORK=off go mod download all
+GOWORK=off GOPROXY=off GOSUMDB=off go mod verify
+GOWORK=off GOPROXY=off GOSUMDB=off go run .
+GOWORK=off GOPROXY=off GOSUMDB=off go test -race ./...
+GOWORK=off GOPROXY=off GOSUMDB=off go vet ./...
+```
+
+All passed. The demonstration reports accepted=1, rejected=0, attempts=4,
+transient_retries=0, structural_repairs=2. Module verification reports all modules
+verified. Published tag availability was checked with `git ls-remote --tags` for
+ConfigKit v0.3.1 and PromptKit v0.11.0; successful module acquisition verifies this
+is not merely a local-checkout experiment. The module has no replacements.
+
+The minimum Go 1.25.5 toolchain was also exercised with offline tests and the same
+successful demonstration. The exact installed-toolchain invocation was:
+
+```sh
+GOTOOLCHAIN=local GOWORK=off GOPROXY=off GOSUMDB=off /home/eric/go/pkg/mod/golang.org/toolchain@v0.0.1-go1.25.5.linux-amd64/bin/go test ./...
+GOTOOLCHAIN=local GOWORK=off GOPROXY=off GOSUMDB=off /home/eric/go/pkg/mod/golang.org/toolchain@v0.0.1-go1.25.5.linux-amd64/bin/go run .
+```
+
+An initial attempt to select that cached toolchain through `GOTOOLCHAIN=go1.25.5`
+with checksum verification disabled was rejected by Go's toolchain verifier. Using
+the installed binary with `GOTOOLCHAIN=local` proved offline execution without
+weakening dependency checks. The portable README commands assume an installed
+sufficiently new Go executable on PATH and disable toolchain switching.
+
+From the Audita root, `go test ./...` and
+`go build -o ./bin/audita ./cmd/audita` passed. These check existing production code;
+they do not cover the nested module. Affected relative documentation links,
+inventory coverage, and `git diff --check` were checked separately.
+
+### Findings and Limits
+
+No adoption-blocking upstream API gap was found. Two integration details deserve
+explicit attention: PromptKit's profile GoDoc calls `timeout_seconds` a generation
+deadline, but its built-in HTTP path applies it separately to each attempt; the
+consumer therefore needs the selected total-operation deadline. Also,
+`APIKeyRequired` is an in-memory profile facility, not a profile-YAML field. The
+file-profile design uses the presence of ConfigKit's optional credential reference
+to require local secret resolution; it does not invent a new PromptKit YAML key.
+
+The chosen defaults are proposed operational policy, not measured model-quality or
+cost recommendations. Tests use controlled transport responses and synthetic data;
+they do not establish real provider compatibility, model quality, network behavior,
+throughput, cross-process coordination, complete billing, or release compliance.
+Only the representative LLM-facing subset is implemented in the prototype; the
+inventory specifies the remaining domain fields for the ConfigKit migration.
+Production file discovery, CLI parsing, full module chains, all nine prompts,
+filesystem diagnostics/redaction, and serialized report migration remain their
+assigned production tasks. Those are deliberate scope boundaries, not missing
+claims of completion for the prototype.
+
+The new profile/credential/capacity and deadline choices are explicit above; no
+unresolved upstream requirement prevents starting the ConfigKit implementation
+plan. Production report field names and final runtime wiring are selected in the
+[implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md), following the outcome and observation
+contract here. The coordinated refactor targets both libraries directly; intermediate commits do not require a working legacy runtime.
+
+### Production Implementation Reference
+
+The [implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md) owns ordered production work,
+component changes, tests, and acceptance criteria. The [combined roadmap](library-adoption.md)
+and [configuration roadmap](configkit.md) own the final target. This evidence does
+not require an interim runtime bridge or separately runnable configuration migration.
+
+Retained domain scope includes transcript/glossary schemas, normalization/chunking,
+output encoding, proposals and apply-time checks, deterministic validators, ordered
+modules, LLM-validator batching/cardinality, ledger ordering, and artifact lifecycle.
+The production integration replaces loading, provider execution, prompt rendering,
+and scheduler composition while preserving those responsibilities. Retention and
+timestamp changes remain separate work.
+
+### Distribution Follow-Up
+
+The pinned graph includes ConfigKit, PromptKit, both PromptKit backend catalogs,
+jsonschema/v6, yaml.v3, yaml.v4, x/text, and kr/text (including module/test dependency
+edges). Before distribution, review the actual production dependency graph and
+its license files, including PromptKit/catalog LGPL-3.0-or-later terms. Decide and
+verify the notices, license copies, corresponding library source/modifications,
+and rebuild/relink materials needed for Audita's Go binary/container distribution.
+This unit does not publish binaries or claim that release compliance is complete.
+No upstream relicensing request or library implementation change is required for
+this prototype.
+
+## Selected-Overlay Extension: 2026-09-24
+
+The isolated prototype now pins published ConfigKit v0.4.0. ConfigKit loads
+`overlays/testing.yml` relative to the selected main file; the overlay selects
+complete PromptKit execution profiles from the separate `profiles/` directory.
+The original v0.3.1 observations above are historical and do not prove this
+extension. No production Audita behavior changed.
+
+With Go 1.26.8 on linux/amd64, from `examples/library-adoption/`, these offline
+checks passed with `GOTOOLCHAIN=local GOWORK=off GOPROXY=off GOSUMDB=off`:
+
+```sh
+go test ./...
+go test -race ./...
+go vet ./...
+go run .
+go run . --overlay testing
+go run . --overlay testing --proposal-profile proposal
+```
+
+Each demonstration reported one accepted correction, four attempts, zero
+transient retries, and two structural repairs. The base run selected
+`synthetic-grammar`/`synthetic-review`; testing selected
+`synthetic-grammar-testing`/`synthetic-review-testing`; the explicit proposal
+override restored only `synthetic-grammar`. These are synthetic routing IDs,
+not price, latency, quality, or provider-compatibility measurements.
+
+| Requirement | Consumer evidence |
+| --- | --- |
+| Base, empty/default, testing selection, role overrides, final inheritance, and relative profile resolution | `TestSelectedOverlayConfiguration` loads temporary on-disk layouts and assembles both roles. |
+| Actual base/testing/override HTTP models and repairs | `TestDemonstrationSelectsOutboundModels` captures every request body and matches model IDs to the reported selections and domain outcome. |
+| Missing/invalid selection, version/directive rejection, and ordinary backend-setting validation | `TestOverlayFailuresAndApplicationValidation` checks ConfigKit diagnostic codes and effective backend limits. |
+| Surviving credential resolution, absent/empty failures, null unauthenticated use, and profile-only switching | `TestOverlayCredentialResolutionAndInspection` controls synthetic environment values and observes HTTP Authorization. |
+| Redacted effective output, explanation and metadata origin, profile inspection, and observations | `TestOverlayCredentialResolutionAndInspection` and `TestOverlaySelectionProvenance` inspect safe output and overlay/CLI origins. |
+| Existing retry, repair, domain, capacity, and concurrent-observation behavior | `TestWorkflowBoundaries`, `TestSharedCapacityAndRecovery`, `TestQueuedCancellationAndDeadline`, `TestDeadlineStopsRetryAfter`, and `TestPartialUsageOnFailedRun` pass, including under `-race`. |
+
+Changing a profile does not remove a backend credential reference from the
+effective configuration. ConfigKit validates a surviving reference when only
+profile selection changes. Replacing the reference in an
+overlay resolves only the replacement; explicit null allows unauthenticated use.
+This distinction is part of the selected configuration boundary, not an
+application fallback. Imports and secret-directory bootstrap remain disabled.
+
+The tests use a controlled built-in HTTP transport and synthetic credentials;
+there were no live provider calls. The prototype still covers one representative
+proposal/assessment pair, not all production modules, prompts, CLI paths, or
+diagnostic artifacts. Root `go test ./...` does not run this nested module. No
+upstream API gap appeared in this extension. The [production implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md)
+keeps implementation and release pending.

docs/roadmap/library-boundary.md

@@ -0,0 +1,269 @@
+# Library Adoption: Shared Boundary and Prototype
+
+**Last updated:** 2026-09-24
+
+**Status:** Original boundary-design and offline-prototype unit completed;
+overlay extension and production-planning handoff completed, 2026-09-24.
+The work packages below retain the original scope;
+[decisions, evidence, limitations, and handoff](library-boundary-findings.md)
+record the outcome. The executable prototype was removed after production
+integration; Git history at `590604b` retains it. No publication or upstream
+changes were made in that historical unit.
+
+## Next Unit: Testing Overlays and Production Migration Plan
+
+**Status:** Complete. The [production implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/cb405f6/docs/roadmap/implementation.md)
+owns subsequent work; the [overlay findings](library-boundary-findings.md#selected-overlay-extension-2026-09-24)
+record this unit's checks. The original work packages below remain
+historical scope for the completed first unit.
+
+Extend the existing prototype to prove the [selected-overlay contract](configkit.md#selected-overlays-for-testing)
+with ConfigKit v0.4.0, then turn the verified findings into a detailed production
+ConfigKit migration plan. Do not repeat the original inventory or redesign the
+established library boundary without new evidence.
+
+Required work:
+
+- Upgrade only the nested prototype module from ConfigKit v0.3.1 to v0.4.0.
+- Pass an explicit overlay selection through ConfigKit and add `overlays/testing.yml`
+  plus two PromptKit testing profiles. Select both roles and retain one shared
+  finite backend. Synthetic model names demonstrate routing; they do not benchmark
+  real model cost, speed, or quality.
+- Make the offline executable accept overlay and explicit role-profile selections,
+  and show the selected models alongside domain outcomes and retry/repair counts.
+- Verify base/overlay/CLI precedence, role inheritance, invalid selections,
+  surviving credential references, safe inspection, and overlay provenance.
+  Prove the actual HTTP request models for both roles; loading changed strings
+  alone is insufficient evidence.
+- Record new results separately from the historical v0.3.1 findings. Expand the
+  existing ConfigKit roadmap into a concrete production implementation handoff,
+  including final runtime integration and removal of legacy configuration.
+
+Target end state: reviewers can run the prototype with and without the testing
+selection and see both LLM roles switch models. The offline test suite establishes
+that selection, inspection, and execution agree, and the production ConfigKit work
+has bounded tasks with explicit acceptance criteria.
+
+Acceptance criteria:
+
+- The published v0.4.0 dependency is pinned without local replacements; the original
+  no-overlay path and existing integration tests continue to pass.
+- `testing` switches both roles, explicit profile flags override its selections,
+  and omission/empty/`default` retains base selection according to ConfigKit.
+- Invalid overlays fail clearly; inheritance and secret resolution use surviving
+  effective settings. Inspection records overlay origin without disclosing keys.
+- Both selected testing models reach the controlled HTTP transport; structural
+  repair, shared capacity, observations, and Audita domain checks remain active.
+- Updated examples, executable instructions, findings, and a detailed production
+  migration plan are reviewable. Exact verification commands and limitations are
+  recorded, including the fact that root tests do not cover the nested module.
+- Production code, root dependencies, and upstream libraries remain unchanged.
+  No release or live provider call is part of this unit.
+
+This unit ends with a production migration **plan**, not production implementation.
+The ConfigKit → PromptKit → combined release sequence remains unchanged. Imports
+and secret-directory bootstrap remain disabled; PromptKit execution profiles keep
+their existing names and meaning.
+
+## Objective
+
+Define and prove the boundary between Audita, ConfigKit, and PromptKit before
+replacing Audita's production configuration or LLM runtime. This unit covers the
+shared design and representative prototype supporting the
+[coordinated adoption roadmap](library-adoption.md). The
+[ConfigKit roadmap](configkit.md) continues to own the broader configuration scope.
+
+The deliverable is a concrete target contract supported by an offline executable
+prototype, sufficient to implement the ConfigKit migration without subsequently
+redesigning its LLM-facing settings. A successful prototype is not completion of
+either production migration.
+
+## Established Direction
+
+- This is an intentional breaking migration with a migration guide, not legacy
+  configuration support, aliases, or a permanent compatibility mode.
+- ConfigKit owns shared configuration mechanics; PromptKit owns shared prompt and
+  generation mechanics; Audita retains domain meaning and transcript correction.
+- Structural repair is enabled from the outset for proposal and LLM-validator calls.
+  Positive default repair budgets must be selected, not silently left at zero.
+- Transient retries and structural repair are separate. Explicit zero disables
+  transient retries; Audita must not add a second automatic retry loop.
+- One finite concurrency limit per backend protects local capacity and bounds paid
+  API activity. Calls from both roles share that limit. Fine-grained role budgets
+  and an additional global concurrency control are unnecessary.
+- Keep the sequence: boundary/prototype, ConfigKit migration, PromptKit migration,
+  then combined major release. No intermediate public release is planned.
+
+## Scope and Deliverables
+
+| Deliverable | Required result |
+| --- | --- |
+| Configuration inventory and migration map | Every existing setting and invocation option has a proposed destination, replacement, or removal, with rationale. |
+| Boundary specification | One authoritative source for each setting, explicit precedence/defaults, credential flow, runtime conversion, and failure handling. |
+| Representative executable prototype | New-style configuration drives one proposal and one assessment through released libraries and a controlled provider. |
+| Focused behavioral evidence | Retry, repair, capacity, credential, observation, and domain-boundary scenarios below are demonstrated. |
+| Decision and integration handoff | Concrete choices, measured limitations, affected components, removal plan, and next implementation tasks are recorded. |
+
+Do not switch the production CLI to the prototype, migrate every module/prompt,
+redesign transcript/glossary formats, change correction algorithms, modify retention
+or timestamp behavior, or publish a release in this unit. Do not reproduce the
+libraries' full test suites. Upstream changes require a separately scoped task if
+an actual gap is found.
+
+## Work Package 1: Inventory and Decide Ownership
+
+Read the current [configuration](../config.md), [CLI](../cli.md),
+[LLM runtime](../internal/llm-runtime.md), [prompts](../internal/prompts.md),
+[validators](../internal/validators.md), and [output contracts](../integrations/output-files.md).
+Use graph-first discovery to verify the relevant code and tests, especially
+configuration application, prompt construction, structured responses, scheduling,
+and report assembly.
+
+Build a migration table grouped by application/domain settings, invocation state,
+backend settings, execution profiles, output contracts, and inspection. Include
+current environment settings and aliases without treating their existence as a
+requirement to preserve them. Identify report fields whose meanings change when
+one logical generation contains multiple provider attempts.
+
+Resolve these design decisions explicitly:
+
+| Decision | Proposed direction to prove |
+| --- | --- |
+| File selection and version | Audita selects the main configuration and validates a new schema version; specify discovery and the treatment of old versions. |
+| Application versus invocation state | ConfigKit loads persistent settings; Audita parses invocation-only input/output paths and translates explicitly supplied setting flags into supported overrides. |
+| Backend identity | One registration represents the service capacity shared by proposal and validation calls. Avoid separate identities for the same service merely because roles differ. |
+| Endpoint and capacity ownership | Define them once per backend, with finite limits and an explicit policy for waiting capacity/exhaustion. Do not expose endpoint-only unlimited execution accidentally. |
+| Execution profiles | Prefer complete application-owned profiles using PromptKit conventions. Decide their source and selection without duplicating model/settings fields in application configuration. Catalog selection remains a deliberate option. |
+| Role selection and inheritance | Specify how proposal and validation choose profiles, including omission and any inheritance. Do not inherit old behavior solely for compatibility. |
+| Credentials | Prefer one ConfigKit-owned secret-resolution path passed through PromptKit's supported request credential field. Define backend/profile credential association and avoid competing environment lookups. Validate this against profile requirements. |
+| Retry, repair, and deadlines | Assign one owner and explicit defaults for each control. Use positive structural-repair defaults and document combined limits and total-operation scope. |
+| Optional library features | The original prototype omitted imports, selected configuration layers, and secret-directory bootstrap. Selected overlays are now planned for testing; see the [overlay plan](configkit.md#selected-overlays-for-testing). Imports and secret-directory bootstrap remain out. |
+| Inspection | Define effective configuration and selected-profile inspection without implying that one library's snapshot contains the other's complete state. |
+
+Choose concrete field names, example values, defaults, units, and omission/null/zero
+semantics for the prototype. In particular, decide what validation and inspection
+do when configured credentials are unavailable. Classify choices as proposed,
+selected, or unresolved; do not present candidates as implemented contracts.
+
+## Work Package 2: Specify Execution and Reporting Semantics
+
+Document the flow from selected configuration through backend/profile assembly,
+prompt inputs, generation, structural validation/repair, typed decoding, and Audita
+domain validation. Identify which values cross the boundary and which component
+owns each conversion. Keep library types out of unrelated domain code where a
+small existing domain value is sufficient.
+
+Use the current correction-set and validator-decision schemas as representative
+inputs. Reuse domain checks where practical. Select one real proposal/validator
+pair; account for prompt helpers, template variables, shared fragments, and schema
+references rather than testing only an unrelated hello-world prompt.
+
+Record the following distinctions:
+
+- Repair seeks output conforming to its structural contract. Success does not
+  establish semantic correctness or authorize applying a correction.
+- A schema-valid decision set still requires Audita's cardinality and proposal
+  identity checks. Semantic rejection is not automatically another repair request.
+- Exhausted structural repair yields failed content validation; provider failures,
+  retry exhaustion, cancellation, and capacity rejection have separate outcomes.
+- Specify each outcome's intended effect on an Audita section/batch and overall run,
+  including warnings, rejected work, partial domain results, and terminal failures.
+  Preserve domain intent without preserving obsolete transport behavior by accident.
+- For transient retry budget T and structural repair budget R, one logical call can
+  dispatch up to `(T + 1) * (R + 1)` provider attempts. For example, T=2 and R=2
+  permits nine attempts; these illustrative values do not select production defaults.
+- Choose the total-operation deadline scope for a logical proposal or assessment;
+  distinguish it from per-attempt timeouts and whole-process cancellation. Provider
+  retry timing may exceed local backoff, so deadline behavior must be explicit.
+
+Design the Audita observation record and aggregation rules. Include logical-call
+and attempt identity, role/backend correlation, queue/generation/retry timing,
+repair and retry counts, outcomes, and available usage. Preserve unknown/incomplete
+usage rather than inventing zeros or complete costs. Do not add terminal totals to
+attempt totals. Collect events safely across concurrent calls, without blocking
+callbacks on progress by the same run or persisting raw library events as a stable
+Audita schema. Transcript-bearing artifacts remain separately governed by Audita.
+
+## Work Package 3: Build the Bounded Offline Prototype
+
+Use ConfigKit v0.3.1 and PromptKit v0.11.0 as reviewed starting points; verify their
+published availability and APIs before pinning the prototype. Recheck the combined
+Go minimum (the reviewed PromptKit release requires Go 1.25.5). Do not depend on
+sibling checkouts, an active Go workspace, or committed local module replacements.
+
+A proposed home is `examples/library-adoption/` as an explicitly experimental,
+standalone Go module, isolating prototype dependencies from the production binary.
+Confirm this placement during execution. Its README must give exact offline run
+and test commands after dependency acquisition; root `go test ./...` must not be
+claimed to cover a nested module. Keep complete proposed configuration and prompt
+fixtures with it and clearly label them as prototype-only.
+
+The prototype should:
+
+1. Load representative application settings with ConfigKit, using synthetic secret
+   values supplied by the test environment and an application-owned version rule.
+2. Assemble shared backend registrations and proposal/validation profile selections.
+3. Supply representative transcript/proposal inputs to PromptKit definitions and
+   actual response schemas using supported sources.
+4. Use the built-in HTTP generation path with a controlled local endpoint or
+   transport. An injected LLM client alone cannot prove built-in HTTP retry behavior.
+5. Run proposal generation, decode and check candidates, run assessment, and perform
+   the selected Audita decision checks without modifying real transcript files.
+6. Collect selected safe observations and show a compact result understandable to a
+   reviewer, including repaired output and final domain acceptance/rejection.
+
+No paid calls, real credentials, host environment assumptions, persistent caches,
+or general-purpose test framework are needed. Make concurrency and cancellation
+checks deterministic through controlled synchronization. Keep delays bounded and
+use library-supported controls; do not patch library internals to make tests pass.
+
+## Focused Verification Matrix
+
+| Scenario | Evidence required |
+| --- | --- |
+| Successful assembled path | ConfigKit-loaded selections drive both real prompt roles through PromptKit and Audita checks. |
+| Presence, overrides, and inheritance | Representative omission, explicit values, and role selection follow the proposed contract; unknown fields and unsupported versions fail clearly. |
+| Credentials | Required missing/empty references fail as intended; controlled calls receive the chosen synthetic credential; inspection and observations do not expose it. |
+| Structural repair succeeds | Invalid proposal output and invalid assessment output are repaired in their respective workflows; positive budgets are actually exercised. |
+| Repair exhausted | Repeated invalid output produces the selected Audita outcome and accurate accounting, without an unbounded outer retry. |
+| Domain rejection | Structurally valid but unacceptable correction or decision is rejected by Audita, demonstrating the semantic boundary. |
+| Transient retry and explicit zero | An eligible temporary status recovers; zero makes one attempt. Exhaustion retains a useful classified cause. |
+| Combined repair and retry | A repaired generation encountering a transient failure respects both budgets and reports distinct counts. |
+| Shared capacity | Concurrent calls from both roles, including retry/repair activity, stay within one selected backend cap; test a representative local cap of four and a finite API cap. |
+| Waiting, cancellation, and deadline | Cancellation/deadline stops pending work and retry waits; capacity exhaustion has the intended application outcome and later calls can still proceed. |
+| Failed-run observations | Completed work remains observable after later failure; unknown usage, known zero, and aggregate counts are mapped correctly. |
+
+Tests should protect Audita's integration decisions and important assembled behavior.
+Rely on PromptKit's tests for exhaustive HTTP-status, backoff, and scheduler details;
+rely on ConfigKit's tests for its parser and presence machinery. Use race testing for
+new concurrent observation collection. Keep consumer checks offline and credential-free.
+
+## Completion Criteria and Handoff
+
+This unit is complete only when:
+
+- The settings/ownership table and migration inventory are concrete enough to start
+  the ConfigKit migration, including file/profile sources, CLI scope, credentials,
+  defaults, finite backend limits, positive repair budgets, and deadline policy.
+- The offline prototype and focused matrix demonstrate the selected boundary with
+  both libraries together, including both LLM roles and Audita domain checks.
+- Exact verification commands, results, limitations, and unresolved upstream gaps
+  are recorded. Missing evidence is not reported as successful integration.
+- No adoption-critical decision remains hidden in prototype code. If a gap prevents
+  settling the boundary, record it and revise the plan before declaring readiness.
+- The existing Audita production behavior remains intact in this unit. Temporary
+  code is clearly isolated and has a removal/integration destination during the
+  production migrations; no second supported runtime is introduced.
+- A component-level follow-up identifies code to replace, domain code to retain,
+  tests to update, documentation/migration-guide work, and direct integration of
+  the final configuration and PromptKit runtime.
+- Licensing/distribution follow-up for the selected dependency graph is identified
+  for the later release stage; this prototype does not claim release compliance.
+- Roadmap links and examples resolve, appropriate checks pass, and commits contain
+  only this unit's work. Significant settled architectural choices may be recorded
+  in an ADR under the [documentation policy](../policy/documentation.md), while
+  implementation status stays in the roadmaps.
+
+Retain the overall adoption order unless evidence or the other planned large changes
+justifies revising it. Completing this unit enables detailed production migration;
+it does not authorize an intermediate release or require unrelated runtime fixes.

docs/roadmap/remediation.md

@@ -0,0 +1,161 @@
+# Maintainability Audit: Findings and Recommended Changes
+
+**Reviewed:** 2026-09-25, source commit `011b021`.
+**Scope:** Recent partial-decision reconciliation, validation context, prompt
+alignment, and request-local reference changes, including related existing code.
+**Status:** R27–R30 addressed; see the resolution record below. Previous findings and resolution records remain in Git
+history through `011b021`.
+
+## Architecture Assessment
+
+The recent changes fit Audita's existing boundaries. Proposal generation,
+validation, prompt assets, and outcome reporting already have suitable owners;
+this audit does not recommend a new common package or a broad refactor.
+
+The strongest opportunities are shared prompt instructions and request-scoped
+validation context preparation. Two smaller improvements concern recovery-reason
+classification and partial-result assembly; both predate the latest commits.
+These are bounded maintenance recommendations, not new release gates or changes
+to correction policy, response acceptance, or reporting semantics.
+
+## R27 — P2: Consolidate Shared Prompt Instructions
+
+**Category:** Repeated behavioral-contract text; risk of future drift.
+
+**Evidence:** The four [proposal prompts](../../internal/prompts/assets/prompts/modules.grammar.proposal.yaml)
+repeat literal source-span matching, request-local identity, replacement-text,
+source-based proposal, and confidence instructions. The three active
+[validator prompts](../../internal/prompts/assets/prompts/validators.editorial_review.yaml)
+repeat candidate-reference, supplemental-context, and verdict-confidence guidance.
+Recent contract changes required coordinated edits across these assets. The
+[prompt registry](../../internal/prompts/registry.go) already provides a shared
+hardening fragment supplied through the role-bound adapter.
+
+**Recommended change:** Extract genuinely common passages into embedded fragments
+under `internal/prompts/assets/shared`, following the existing fragment-supply
+approach. Keep the fragments owned by `internal/prompts` and use the existing
+PromptKit rendering path. Keep module-specific examples, replacement-policy
+instructions, protected-name exceptions, and semantic criteria in their individual
+prompt definitions. Do not merge similar-looking rules with different meanings.
+
+**Acceptance:** Common instructions have one maintained source and appear in all
+intended rendered prompts. The existing module and validator distinctions remain
+intact, including glossary replace-all behavior and the spoken-word protected-name
+exception. Focused rendering checks cover fragment inclusion and representative
+module-specific guidance without requiring exact-prose snapshots or introducing
+a separate templating framework.
+
+## R28 — P2: Prepare Validation Context Once Per Request
+
+**Category:** Repeated computation and overlapping segment lookup.
+
+**Evidence:** [BuildLLMValidationRequest](../../internal/framework/validators/llm_validators.go)
+already indexes transcript segments, but
+[validationContext](../../internal/framework/validators/llm_context.go) scans the
+transcript again for each candidate. It also calls
+[NewProtectedVocabulary](../../internal/framework/validators/protected_terms.go)
+for each candidate, rebuilding the same vocabulary and compiling its matching
+expression repeatedly. This is duplicated preparation rather than copied code.
+
+**Recommended change:** Prepare a segment-position lookup and protected vocabulary
+once per validation request, then pass that prepared context into the existing
+context-selection helper. Keep the implementation within `framework/validators`.
+Continue selecting glossary evidence for each candidate individually because its
+proposed replacement can mention different terms. A process-wide cache or new
+context-management package is unnecessary for this cleanup.
+
+**Acceptance:** Candidates in one validation request reuse the prepared lookup and
+vocabulary. Context still consists of the available immediate neighbors within the
+current chunk and relevant glossary entries in glossary order. Candidate-specific
+replacement text still participates in glossary selection. Existing optional-context
+trimming, token estimation, and oversized-input handling retain their behavior.
+Verify these outcomes with focused behavioral tests rather than assertions about
+private helper call counts or unmeasured performance guarantees.
+
+## R29 — P3: Share Execution-Recovery Classification
+
+**Category:** Duplicate reporting rule; predates the latest commits.
+
+**Evidence:** [Proposal generation](../../internal/framework/proposal_generation/generate.go)
+and [LLM validation](../../internal/framework/validators/llm_validators.go)
+independently translate structural-repair and transient-retry counts into
+`output_repaired` and `service_recovered` reasons. The conversion is the same in
+both locations; the surrounding success and domain-outcome decisions differ.
+
+**Recommended change:** Place a small conversion helper in the existing
+`framework/outcomes` package and use it from both callers. Keep the decision to
+record recoveries at the call sites, along with domain-specific reasons such as
+filtered validator decisions. Do not combine proposal and validation orchestration
+or move domain-success classification into this helper.
+
+**Acceptance:** Both callers use the same execution-count conversion. Existing
+reason ordering, successful-operation reporting, and domain-specific additions
+remain intact. A partially unavailable validation batch continues to record its
+failure and execution counts without acquiring recovery records merely because a
+repair or retry occurred. Existing outcome tests should continue to protect these
+distinctions.
+
+## R30 — P3: Centralize Partial Validator Result Assembly
+
+**Category:** Repeated early-return construction; predates the latest commits.
+
+**Evidence:** Several early returns in
+[LLMBackedValidator.Validate](../../internal/framework/validators/llm_validators.go)
+repeat nested slice appends to combine immediate decisions, oversized-input
+decisions, and decisions from completed model batches, together with accumulated
+operation outcomes. This makes it easier for a future failure path to omit facts
+already collected by the validator.
+
+**Recommended change:** Use a small local result-building helper to assemble the
+partial result consistently. Keep error handling, batching, success-path ordering,
+and cardinality enforcement explicit. A general-purpose result-aggregation package
+is not needed.
+
+**Acceptance:** Cancellation and failure paths retain the same previously collected
+decisions and operations. Result ordering and cardinality behavior remain unchanged;
+the helper does not enforce complete cardinality on interrupted work. Retain the
+existing partial-progress and fatal-failure coverage, adding tests only for a
+material gap exposed by the refactor.
+
+## Boundaries to Preserve
+
+- Keep proposal and validator reference mapping separate. Proposals can target the
+  same segment repeatedly; duplicate validator decisions make that candidate
+  unavailable. Similar indexing arithmetic alone does not justify a generic
+  reference-management abstraction.
+- Keep distinct protected-term policy checks separate. They already share the
+  vocabulary implementation and intentionally apply different policies.
+- Preserve exact source-span matching, request-local references, batching, and
+  partial-success behavior. Consolidation is not a reason to strengthen contracts
+  or broaden the accepted correction scope.
+
+## Review Verification and Limits
+
+- This is a source-level maintainability audit, not a finding of demonstrated
+  production failures or a measured performance assessment.
+- No application tests were rerun for the original read-only audit; its roadmap
+  update used documentation link and whitespace checks. Implementation validation
+  is recorded below.
+- Future remediation should run focused behavioral tests and the broader suite
+  appropriate to the actual change, following the testing policy. No live-provider
+  evaluation or new evaluation framework is required by these recommendations.
+- The findings above describe the audited baseline. Implementation and validation
+  are recorded below; no release publication is included.
+
+## Resolution Record
+
+Implemented as focused changes within the existing packages. Proposal and validator
+reference mapping and distinct protected-term policies remain separate.
+
+| Finding | Resolution and evidence |
+| --- | --- |
+| R27 | Common proposal and validation instructions live in two embedded shared fragments, supplied through the existing role-bound PromptKit adapter. Individual YAML definitions retain policy differences and examples. Rendering tests check fragment inclusion once per applicable prompt and representative module-specific guidance. |
+| R28 | Validation request construction builds one segment-position lookup and protected vocabulary, reused by context selection for every candidate. A new test verifies that two replacements of the same source span receive their own relevant glossary entries. Existing neighbor, glossary-order, chunk-boundary, and budget tests pass. |
+| R29 | Both proposal generation and validation use `outcomes.ExecutionRecoveries`. Call sites still control success and domain-specific recovery reasons. Existing tests cover combined recoveries and failed validation with repair counts but no recovery record. |
+| R30 | A local result builder combines collected decisions and operation outcomes for partial returns and completed validation. Sorting and full cardinality checks remain outside it. Existing fatal-batch, partial-progress, runner, and reporting tests pass. |
+
+Validation: `go test ./...` passed, including PromptKit preparation, partial-decision
+reconciliation, subprocess reports, and fatal-progress coverage. Documentation
+links and whitespace checks passed. No live model evaluation or performance
+measurement was performed; the changes make no new model-quality or latency
+promise.

docs/roadmap/resilience.md

@@ -0,0 +1,290 @@
+# Resilient Processing and Focused Run Reports
+
+**Last updated:** 2026-09-24
+
+**Status:** Implemented. This document records the selected behavior and user
+intent. Concrete design decisions and completion evidence are in
+[historical implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/a83a518/docs/roadmap/implementation.md).
+
+## Purpose and User Intent
+
+Audita should reliably produce a useful transcript even when its LLM services
+cannot improve it. Exhausted retries or unusable model responses must not turn
+an otherwise processable transcript into a failed production run.
+
+> Failure of every LLM operation across every stage and chunk still produces a
+> successful output containing the deterministically normalized transcript, with
+> no LLM corrections applied. Partial success preserves as many fully validated,
+> applicable corrections as possible.
+
+Deterministic normalization remains intentional and may change segment layout,
+identifiers, and text before LLM work begins. Output uses the selected transcript
+encoder; byte-for-byte preservation of the original input and mapping corrections
+back to pre-normalization segmentation are not requirements.
+
+The overriding goal is to minimize how much useful processing is lost because of
+one problematic transcript segment. A stage/chunk work unit is an attribution
+boundary, not an all-or-nothing transaction.
+
+Existing reports contain useful information, but there are no compatibility
+obligations for reporting or debug output. Reuse existing machinery only when it
+makes the implementation simpler and clearer; preservation is not itself a goal.
+Internal structures and measurements must serve Audita's processing workflow.
+The reporting package adapts those outcomes to structured and human-facing
+presentations, rather than reshaping processing around a reporting convenience.
+
+## Historical Baseline and Implemented Change
+
+The [pipeline](../internal/pipeline.md), [LLM integration](../integrations/openai-compatible-llm.md),
+[report assembly](../internal/diagnostics-reporting.md), and
+[artifact lifecycle](../operations.md) describe current behavior.
+
+Before this work, exhausted structural repairs could become warnings and skipped
+proposals or rejected validation batches. Other model runtime failures canceled
+peer work, stopped subsequent modules, and prevented final transcript output.
+Successful repairs appeared in execution observations rather than a dedicated
+recovery classification. Detailed artifacts were written during ordinary runs
+and could later be removed by retention. The working transcript was re-chunked
+between modules.
+
+This feature changed those behaviors. It builds on the completed
+[library adoption](library-adoption.md) and [review remediation](remediation.md).
+Fatal cancellation remains necessary; unsuccessful model work must be separated
+from failures that actually require cancellation of the run.
+
+## Fixed Chunks and Ordered Stages
+
+Normalize first, then partition the transcript once. Preserve chunk boundaries,
+segment membership, segment identity, and ordering throughout LLM processing.
+Text evolves as accepted corrections are applied. Recombine chunks in original
+order after the LLM stages finish.
+
+A stage is a configured module instance. Repeated instances of the same module
+have distinct identities. Proposal generation and each validator are operations
+within a stage; validation may use multiple batches within a chunk. A work unit
+is one stage instance applied to one chunk.
+
+Preserve the overall stage sequence: finish a stage across its chunks before
+advancing to the next stage. Concurrent chunk work shares PromptKit's backend
+capacity. Failure in one unit does not prevent other chunks or later stages
+from executing.
+
+Retain operator control over chunk size and the current defaults. Production
+experience indicates that chunk size affects model latency nonlinearly and that
+useful settings are well below backend context limits. Tuning is an empirical
+operator decision. Do not add automatic headroom calculations, cross-profile
+sizing, adaptive repartitioning, or automatic tuning in this feature. A later
+operation that cannot process its input follows the same safe error handling;
+it does not silently alter chunk boundaries. Validation batching within a chunk
+remains supported.
+
+Independent whole-pipeline traversal by chunks and domain-aware response repair
+are explicitly deferred in [future work](future.md). Read-only neighboring
+context, cross-chunk edits, and transcript-wide restructuring are outside this
+unit, without becoming permanent prohibitions on future capabilities.
+
+## Partial Success and Failure Isolation
+
+- Keep bounded transport retries, structural repairs, operation deadlines, and
+  shared backend capacity. Completion comes from safe fallback, not unlimited
+  attempts or an outer retry loop around PromptKit.
+- Apply a correction only after it passes every required validator and the
+  application checks. Failed validation must never be treated as approval.
+- If a validation batch fails, exclude its affected candidates while allowing
+  other candidates to continue through the remaining validators. Preserve fully
+  validated, applicable corrections from successful batches in the same unit.
+- Track candidates by stable proposal identity. Batch membership can change
+  between validators; equal batch numbers do not establish candidate identity.
+- A failed proposal-generation operation leaves its chunk unchanged for that
+  stage because there are no reliable proposals to salvage. Failures later in
+  validation need not discard unrelated successful candidates.
+- Preserve earlier stages' corrections and other chunks' successful work.
+  Record partial success explicitly; a unit can apply corrections and have errors.
+- Continue despite model-operation failures, including exhausted repairs/retries,
+  unusable model decisions, provider/transport failures, local operation deadlines,
+  and admission failures. An operation need not exhaust its attempt budget to fail.
+- A valid empty proposal list or valid assessments rejecting all corrections is
+  successful processing. Correction count alone does not determine failure.
+- Keep safe filtering of defective proposals. Do not retry ordinary editorial
+  rejections. For now, an unusable validator decision set excludes the affected
+  candidates and records an error after existing permitted recovery. New bounded
+  feedback/regeneration for domain-invalid responses belongs to future work.
+- Retain fatal treatment for invalid input/configuration, internal defects,
+  explicit run cancellation, and inability to produce/write the transcript.
+  Caller cancellation is distinct from a local model-operation timeout.
+- Completed runs with model errors return success. Required report-write failures
+  are an additional fatal output condition when reporting is enabled, as below.
+
+PromptKit owns transport, transient retries, structural repair, and admission.
+Audita owns domain acceptance, partial outcomes, fallback, and classification.
+ConfigKit owns configuration resolution. Debug and reporting choices must not
+change correction decisions, retry budgets, or continuation.
+
+## Errors, Recoveries, and Debug
+
+Classify by eventual outcome and impact. These are three reporting categories;
+debug is evidence, not a failure severity. Whether execution continues is a
+separate property.
+
+| Category | Meaning | Representative cases |
+| --- | --- | --- |
+| Errors | Intended work could not complete reliably, or a fatal run-level operation failed. Affected candidates are excluded; other corrections can succeed. | Exhausted repair; unavailable provider; unusable validation decisions; operation deadline. |
+| Recoveries | A problem was handled successfully through retries, repair, or safe filtering. No operator intervention is implied. | Invalid output repaired; transient service failure recovered; defective proposals safely discarded. |
+| Debug | Detailed evidence of processing and ordinary decisions. | Proposals, applied corrections, validator decisions, prompts/responses, individual attempts, execution traces. |
+
+Normal editorial disagreement belongs in debug. A validator correctly rejecting
+a meaning-changing edit is expected behavior. A proposal referring to nonexistent
+original text is defective; safely filtering it is a recovery. An unresolved
+missing/duplicate decision-index response is an error, not a recovery.
+
+Record an operation's eventual outcome once. Several failed attempts followed by
+success produce a consolidated recovery. Attempts ending in failure contribute
+to error details, not additional recoveries. Earlier operations that recovered
+remain recorded if a later operation introduces an error in the same unit.
+
+Separate output-quality recovery, defective-proposal filtering, and service
+recovery with structured reasons. Network instability must not look like a prompt
+defect. Do not introduce an arbitrary warning threshold or a fourth category.
+
+### Counting and Processing Volume
+
+- Count error-bearing work units once per stage/chunk, consolidating underlying
+  operation failures as details. A partially successful unit still counts once.
+- Count recoveries once per recovered operation. Aggregate defective-proposal
+  filtering per proposal response, retaining affected-proposal counts as detail.
+  Attempts and individual discarded proposals are not independent incidents.
+- Keep work-unit coverage separate from operation/recovery counts. Show processing
+  volume with explicitly labeled denominators; do not divide recovered-operation
+  counts by work-unit counts and call that a recovery rate.
+- Report fatal run-level errors separately from the stage/chunk grid when they
+  cannot be attributed to a work unit. Do not invent a chunk to accommodate them.
+- Derive YAML records, Markdown summaries, and CLI counts from actual processing
+  outcomes with deterministic ordering, rather than independently calculating them.
+
+## Production Outputs, Placement, and Lifecycle
+
+Reporting is enabled by default. The ordinary bundle is:
+
+- `errors.yml`: structured error-bearing units, their underlying failures, and
+  applicable run-level failures.
+- `recoveries.yml`: structured recovered operations and pertinent counts.
+- `report.md`: a concise human-facing summary of outcome, corrections, processing
+  volume, errors, recoveries, and their locations in the pipeline.
+
+Write all three documents, including empty event lists for uneventful completed
+runs. Use Audita-owned schemas with safe explanations and stable reason codes;
+retain run, stage-instance, chunk, operation, and validator/batch identities where
+applicable. The implementation plan specifies the versioned schemas; neither old
+report formats nor raw library structures define the new output contract.
+
+Place reports beside the output transcript. Retain transcript-to-stdout support
+with an explicit report-directory selection, defaulting to the invocation working
+directory. Preserve the existing overwrite convention: a new run can replace files
+at its selected destinations. Do not add a mandatory fresh-directory restriction
+or compatibility exports.
+
+Before model work, create initial versions of all three reports carrying a common
+run identifier and explicit `in_progress` status (rendered as prose in Markdown).
+Fail immediately if initialization fails. Also check the transcript destination;
+creating neighboring reports does not prove that an existing transcript file is
+writable. Preflight must not prematurely destroy an existing transcript.
+
+Finalize all required outputs and check every final write. Initial write success
+cannot guarantee later success if space or permissions change. Failure to write
+the transcript or finalize a required report returns nonzero and must not erase
+an already produced transcript. Where possible, finalize available reports with
+an explicit failed status on fatal termination. Abrupt termination may leave
+`in_progress`; elaborate crash-recovery machinery is not required.
+
+Provide an option to disable reporting entirely. It suppresses all three files
+(`errors.yml`, `recoveries.yml`, and `report.md`), including initialization and
+final writes, while retaining the brief CLI summary and independently selectable
+debug output. Disabled reporting must not create or require a writable report
+destination, and must not remove reports from earlier runs.
+Transcript output failures remain fatal regardless of reporting selection.
+
+Ordinary reports survive independently of debug cleanup. Reporting configuration
+and finalization must not depend on retaining a debug run directory.
+
+## Debug and Internal Boundaries
+
+Production issue records are compact. Full responses, transcript excerpts,
+correction lists, and complete validation decisions are opt-in debug evidence.
+Use a single ConfigKit-backed debug boolean with a `--debug` CLI override. Debug
+artifacts may use JSON where useful. Preserve configured-secret redaction in every
+presentation.
+
+Debug-off prevents detailed artifact persistence; writing and later deleting
+those files is insufficient. In-memory domain state needed to process candidates
+correctly is not debug output. Debug and ordinary reporting are independent
+selections and do not alter processing behavior.
+
+There is no obligation to retain `--report-json`, old artifact layouts, parallel
+exports, or old retention settings. Reuse code only where useful to the new
+workflow. Enabled debug output is retained on success or failure; disabled debug
+creates no debug directory. This supersedes the retention question in the
+[runtime follow-up](runtime-behavior.md). The legacy `never` behavior is not the
+meaning of the new reporting-disable option. Optional debug-write failures produce
+a safe stderr diagnostic without changing correction decisions or the run exit
+status; ordinary report and transcript-write failures retain their fatal semantics.
+
+## Deterministic Human-Facing Presentation
+
+Use the CLI style:
+
+> Completed: 18 corrections applied; 3 work units with errors; 7 recoveries.
+
+Do not say that all error-bearing units were skipped: some apply valid corrections.
+Include report locations when reporting is enabled and the files exist. Send the
+summary to stderr so stdout remains exclusively available for transcript output.
+Disabled reporting does not suppress the summary or fatal-error messages.
+
+Keep `report.md` concise and high-level. Use separate error and recovery matrices
+with stages as columns and chunks as rows. Include every chunk in fixed order;
+retain configured stage order and distinct labels for repeated module instances.
+No automatic width detection, thresholds, adaptive omission, or table splitting.
+A wide or long table is acceptable.
+
+The error matrix counts error-bearing work units: each attempted cell is 0 or 1.
+Recovery cells count recovered operations and can exceed 1. An error cell may
+coexist with recoveries and successfully applied corrections in the same unit.
+Distinguish completed work with zero events from unattempted work; do not portray
+canceled/unavailable processing as clean completion. Include totals, processing
+volume, and a short reason breakdown. Detailed inspection belongs in YAML or debug.
+
+## Target Acceptance Criteria
+
+- Total LLM failure still returns success and the encoded normalized transcript,
+  provided required output writes succeed.
+- Mixed-success runs preserve independently validated, applicable corrections
+  within the same stage/chunk as a failed batch. No candidate bypasses a required
+  validator; later validators handle survivors by stable proposal identity.
+- Subsequent stages and unaffected chunks execute after recoverable failures.
+  Fatal errors and explicit cancellation stop and join active work appropriately.
+- Chunk boundaries and identity remain fixed, with operator-selected sizing and
+  unchanged defaults; final recombination preserves ordering.
+- Valid no-change/editorial-rejection outcomes are successful. Repaired/filtered
+  output and unresolved failures receive distinct classifications and counts.
+- YAML, Markdown, and CLI agree on partial outcomes and totals; error matrices
+  count affected work units, recovery matrices count recovered operations.
+- With reporting enabled, all three reports are initialized before model work,
+  finalized and retained, including empty lists. Initial or final required-output
+  write failures produce nonzero exits without destroying completed transcript output.
+- Reporting can be disabled without requiring its destination. Debug can be enabled
+  independently; neither selection changes processing outcomes. Debug-off writes
+  no detailed debug artifacts, and stdout transcript output remains supported.
+- Reports use the fixed deterministic layout without adaptive presentation logic.
+- Meaningful controlled-provider tests cover exhausted recovery, failed validation
+  batches with successful survivors, total failure, fatal cancellation, output
+  failures, and reporting/debug selections. Preserve domain safeguards, capacity
+  limits, and redaction.
+- Canonical CLI, configuration, operations, integration, internal, and architecture
+  documentation describes the implemented final behavior.
+
+## Implementation Reference
+
+The product decisions are settled. The [historical implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/a83a518/docs/roadmap/implementation.md) defines
+the selected settings and precedence, YAML schemas, counts and ordering, interrupted
+work presentation, destination checks, and output finalization. It also owns the
+ordered implementation stages and validation work. This document remains the
+reference for the desired behavior and its acceptance criteria.

docs/roadmap/runtime-behavior.md

@@ -0,0 +1,18 @@
+# Runtime Behavior Decisions
+
+**Last updated:** 2026-09-24
+
+The documentation migration identified existing behaviors for separate product
+consideration. The resilience implementation resolved the diagnostics-retention
+question; the input-timestamp question remains open.
+
+- Resolved: legacy `auto`/`always`/`never` retention settings were removed.
+  Debug is explicitly enabled with `diagnostics.debug` or `--debug`, creates no
+  directory when disabled, and retains its directory on every outcome when enabled.
+- Resolved: [configuration](../config.md) now treats explicit zero as disabling
+  transient retries or structural repair independently; PromptKit performs both.
+- Decide whether missing or null input timestamps should be rejected; current
+  decoding is documented in the [input contract](../integrations/transcript-glossary-files.md).
+
+These decisions are independent of the completed documentation migration and
+release workflow.

docs/roadmap/span-deletion.md

@@ -0,0 +1,203 @@
+# Consistent Span Deletion
+
+**Last updated:** 2026-09-25
+
+**Status:** Implemented. This document owns the desired behavior and user intent.
+The [historical implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/60e2fe7/docs/roadmap/implementation.md)
+records the concrete steps, selected implementation details, and completion
+evidence.
+
+## Purpose and User Intent
+
+Align proposal prompts, the LLM response schema, internal validation, and
+application behavior so legitimate text-span deletions can be considered and
+applied through the ordinary correction workflow. An empty replacement should
+not trigger structural repair merely because it is empty, nor invalidate other
+usable corrections returned alongside it.
+
+This is a consistency fix within Audita's current editing scope. It does not
+authorize whole-segment removal, broaden module responsibilities, or introduce
+new recovery machinery. The accepted [resilience behavior](resilience.md) remains
+in effect, including partial success, fixed chunks, and validation before mutation.
+
+## Historical Baseline (`0b6cd98`)
+
+The source reviewed at `0b6cd98` contains a mismatch:
+
+- The [spoken-word prompt](../../internal/prompts/assets/prompts/modules.spoken_word.proposal.yaml)
+  permits removing fillers and asks for the exact affected span and replacement,
+  without explaining how to represent deletion. Grammar permits punctuation and
+  spacing changes, which can also naturally involve deletion.
+- The shared [correction response schema](../../internal/framework/responseschema/assets/correction_set.v1.json)
+  requires `corrected_text` to have at least one character. Every built-in proposal
+  module uses that schema. An empty replacement invalidates the response before
+  Audita evaluates individual proposals.
+- PromptKit performs bounded structural repair. If repair does not produce a
+  valid response, the proposal operation fails and yields no candidates, including
+  otherwise valid proposals in the same response. The current resilience path
+  records the error and continues other work.
+- [CorrectionProposal](../../internal/framework/proposals/proposal.go) already
+  allows empty replacement text. The
+  [preview/application logic](../../internal/framework/proposals/preview.go)
+  supports literal span deletion when the resulting segment remains nonblank.
+- Domain shape checks currently reject whitespace-only original spans, although
+  the response schema permits them and grammar prompts permit spacing cleanup.
+  This is another restriction to align with the target contract below.
+- The validator named `non_empty_corrected_text` actually checks the resulting
+  segment, not the replacement string. Both this check and application prevent
+  a segment from becoming empty or whitespace-only. Application changes segment
+  text; it does not remove segment records.
+
+A temporary controlled-provider review reproduced whole-response rejection for
+an empty replacement alongside a valid correction, and immediate schema acceptance
+for a space-only replacement. Existing proposal tests confirm internal support for
+partial deletion and rejection of blank resulting segments. These observations
+explain a plausible production failure mechanism; they do not establish the cause
+of any particular historical incident without its payload and running version.
+
+## Target Correction Contract
+
+- `original_text` identifies an existing, nonempty literal span in the current
+  target segment. Whitespace-only spans are eligible; an empty string identifies
+  no source span and remains invalid. Matching and ambiguity checks still apply
+  under the module's replacement policy: `require_unique` for grammar, spoken-word,
+  and homophones; `replace_all` for glossary.
+- `corrected_text` remains a required string. An empty string means delete the
+  identified span; it does not mean remove the segment record or skip validation.
+  Missing fields and non-string values remain structurally invalid.
+- After applying a correction, the segment must contain non-whitespace text.
+  Enforce this using the source segment and replacement result in Audita, not a
+  minimum replacement length in the structural schema.
+- Every accepted deletion passes the same required validators as other edits in
+  its module, including confidence, protected terms, editorial review, and meaning
+  preservation where applicable. Empty replacement is neither automatic approval
+  nor an automatic rejection.
+- Application preserves segment identity, speaker, timing, membership, ordering,
+  and fixed chunk boundaries. Recheck applicability against text produced by
+  earlier accepted corrections, retaining existing conflict handling.
+- Preserve literal replacement behavior. Prompts should select spans and
+  replacements that handle adjacent whitespace and punctuation appropriately;
+  this feature does not add automatic trimming or a second normalization pass.
+  Neither side of a correction is rejected solely for containing only whitespace.
+  For example, reducing a uniquely identified run of two spaces to one space may
+  be appropriate; validity still depends on the resulting text and domain checks.
+
+Preserve the intended safeguards and outcomes, not every detail of the current
+implementation. Validator ordering, internal structure, and payload construction
+may change when needed to achieve this consistent contract. Existing behavior or
+a passing test does not justify retaining a conflicting restriction.
+
+Illustrative cases:
+
+| Segment | Original span | Replacement | Intended handling |
+| --- | --- | --- | --- |
+| `uh hello` | `uh ` | empty string | Structurally valid; may produce `hello` after required validation. |
+| `uh hello` | `uh hello` | `hello` | Remains a valid representation of the same cleanup, subject to validation. |
+| `hello  world` | two spaces | one space | Eligible spacing correction, subject to unique matching and required validation. |
+| `uh` | `uh` | empty string | Structurally valid proposal, then filtered because the resulting segment would be blank. |
+| `hello world` | `hello world` | spaces only | Filtered because the resulting segment would be blank. |
+
+These examples describe eligibility and mechanical results, not unconditional
+editorial approval. An unchanged segment or valid empty correction list remains
+a successful no-change outcome.
+
+## Prompt and Module Alignment
+
+Review relevant prompt definitions for consistency and edit only those needing
+clarification. Where needed, explain empty-string deletion, whitespace-only spans,
+and the resulting-segment limit with a concise, module-appropriate example. The
+model should omit a correction that would leave a blank segment rather than invent
+padding or replacement content. Consistent guidance is the goal; adding deletion
+language to every prompt is not a requirement.
+
+Retain each module's existing scope:
+
+- Spoken-word cleanup can propose appropriate filler or dysfluency deletion while
+  preserving substantive content and intentional emphasis.
+- Grammar can propose permitted punctuation or spacing deletion; allowing an
+  empty replacement does not authorize word removal or spoken-word cleanup.
+- Glossary and homophone modules retain their existing correction purposes.
+  Shared deletion syntax does not grant them new editorial responsibilities.
+
+Review the active LLM-validator prompts and their input construction as well as
+proposal prompts. Validators must understand the before/after effect of deletion
+and assess it under their existing criteria. A correct negative editorial decision
+remains ordinary processing, not a defect that should be regenerated until approved.
+
+The resulting-segment validator's name and documented identity should describe
+that responsibility accurately, rather than imply that the replacement string
+itself must be nonempty. Preserve the existing blank-result failure reason and
+safeguard while making that deliberate identity change; current chain position
+and preview implementation are not independent product requirements.
+
+## Outcomes and Reporting
+
+A legitimate deletion is an ordinary correction. Empty replacement alone must
+produce neither structural repair nor a recovery classification. If another
+problem required successful repair or filtering, retain that actual recovery.
+
+A schema-valid proposal that would blank its stage-input segment is a defective
+proposal handled through existing safe filtering. Preserve the
+`empty_resulting_segment` reason
+and consolidated `proposals_filtered` recovery behavior; other usable proposals
+in the response continue through required validation. Ordinary editorial rejection
+remains debug evidence rather than a model error or filtering recovery.
+
+A stage can propose several edits to one segment in the same response. Each edit
+is evaluated against the stage input, then approved edits are applied sequentially.
+If a proposal was applicable to that input but an earlier accepted edit makes it
+unsafe, skip it as an apply-time conflict. Do not count that conflict as a defective
+proposal, model error, or filtering recovery. For example, in `"uh um"`, deleting
+`"uh "` and deleting `"um"` each leaves text individually; after the first deletion,
+the second must be skipped to prevent a blank result. Retain the skipped decision
+in debug evidence. The existing runner already distinguishes this situation.
+
+Debug evidence must accurately retain the empty replacement and the corresponding
+validation/application outcome. Ordinary reports retain their existing counts and
+purpose; no new report schema, incident category, or deletion-specific counter is
+needed. Unrelated structural defects may still exhaust PromptKit repair and produce
+an error for the operation, as today.
+
+## Acceptance Criteria
+
+- A controlled valid response containing an empty replacement reaches Audita's
+  domain validation without structural repair solely due to that empty string.
+- A permitted span deletion can pass the required validators and reach the final
+  transcript while preserving segment metadata and fixed chunk membership.
+- A mixed response can apply an approved deletion and another approved correction;
+  an empty replacement does not invalidate the whole response.
+- A proposal that would blank its stage-input segment is filtered independently,
+  retains the existing reason, and does not prevent unrelated valid candidates
+  from succeeding. A proposal made unsafe only by an earlier accepted edit is
+  skipped without adding an error or filtering recovery; a blank result is still
+  prevented and the conflict remains visible in debug evidence.
+- A uniquely matching whitespace-only original span can reach validation and
+  application. Empty originals remain invalid, ambiguous matches remain subject
+  to module policy, and whitespace-only resulting segments remain disallowed.
+- Protected-term and meaning-changing deletions remain subject to their normal
+  rejection rules. No candidate bypasses required validation.
+- A deletion needing no recovery counts as an ordinary applied correction; actual
+  repair/filtering and editorial rejection retain their existing classifications.
+  Debug evidence represents the empty replacement faithfully.
+- Schema and domain tests distinguish an empty string from missing/non-string
+  replacement values and distinguish empty replacement from blank resulting text.
+- Prompt guidance and validator naming express the same contract as the code.
+  Tests protect outcomes rather than exact prompt prose or incidental formatting.
+
+Use deterministic, offline controlled responses for behavioral integration tests;
+no live-provider or paid evaluation is a prerequisite. Representative production
+examples can inform prompt quality without expanding this into an evaluation
+framework or guaranteeing that models always follow the instructions.
+
+## Deferred Work
+
+[Whole-segment removal](future.md#whole-segment-removal) is only potential future
+work and first requires deciding whether its benefit warrants implementation.
+Domain-aware regeneration and independent whole-pipeline traversal also remain
+in [future.md](future.md). This feature changes neither retry budgets nor the
+Audita/PromptKit ownership boundary and requires no upstream library feature.
+
+Implementation sequencing and concrete names belong to the
+[historical implementation plan](https://gitea.maximumdirect.net/eric/audita/src/commit/60e2fe7/docs/roadmap/implementation.md).
+The target remains consistent deletion of spans within surviving segments, with
+no new operator controls or report types.

examples/minimal-config.yml

@@ -0,0 +1,12 @@
+version: 2
+backends:
+  local:
+    endpoint: http://127.0.0.1:11434/v1
+    concurrency: 1
+llm:
+  proposal_profile: local-proposal
+  validation_profile: local-validation
+pipeline:
+  modules: [grammar]
+output:
+  schema: bare-segments

examples/overlays/testing.yml

@@ -0,0 +1,3 @@
+llm:
+  proposal_profile: service-proposal-testing
+  validation_profile: service-validation-testing

examples/production-config.yml

@@ -0,0 +1,40 @@
+version: 2
+backends:
+  service:
+    endpoint: https://openrouter.ai/api/v1
+    credential: AUDITA_SERVICE_KEY
+    concurrency: 2
+    queue_capacity: 1024
+llm:
+  proposal_profile: service-proposal
+  validation_profile: service-validation
+  transient_retries: 2
+  repair_attempts: 2
+  operation_timeout: 15m
+pipeline:
+  modules: [glossary, homophones, glossary, spoken_word, grammar]
+output:
+  schema: audita-v1
+chunking:
+  # Requires at least this many normalized segments; omit for automatic chunking.
+  target_sections: 8
+  max_section_tokens: 8192
+  min_section_tokens: 2048
+normalization:
+  max_segment_gap: 4.0
+  ellipsis_gap: 3.5
+  max_segment_duration: 60.0
+  max_segment_tokens: 2048
+thresholds:
+  glossary: 0.8
+  homophones: 0.8
+  spoken_word: 0.8
+  grammar: 0.8
+context:
+  description: "General context for domain vocabulary and speaker names."
+reporting:
+  enabled: true
+  stdout_dir: .
+diagnostics:
+  debug: false
+  work_dir: /tmp/audita

examples/profiles/local-proposal.yaml

@@ -0,0 +1,7 @@
+id: local-proposal
+backend: local
+model: replace-with-local-model
+temperature: 0
+max_tokens: 2048
+top_p: 1
+timeout_seconds: 120

examples/profiles/local-validation.yaml

@@ -0,0 +1,7 @@
+id: local-validation
+backend: local
+model: replace-with-local-model
+temperature: 0
+max_tokens: 1024
+top_p: 1
+timeout_seconds: 120

examples/profiles/service-proposal-testing.yaml

@@ -0,0 +1,7 @@
+id: service-proposal-testing
+backend: service
+model: replace-with-testing-model
+temperature: 0
+max_tokens: 2048
+top_p: 1
+timeout_seconds: 120

examples/profiles/service-proposal.yaml

@@ -0,0 +1,7 @@
+id: service-proposal
+backend: service
+model: replace-with-provider-model
+temperature: 0
+max_tokens: 2048
+top_p: 1
+timeout_seconds: 120

examples/profiles/service-validation-testing.yaml

@@ -0,0 +1,7 @@
+id: service-validation-testing
+backend: service
+model: replace-with-testing-model
+temperature: 0
+max_tokens: 1024
+top_p: 1
+timeout_seconds: 120

examples/profiles/service-validation.yaml

@@ -0,0 +1,7 @@
+id: service-validation
+backend: service
+model: replace-with-provider-model
+temperature: 0
+max_tokens: 1024
+top_p: 1
+timeout_seconds: 120

examples/tiny-glossary.yaml

@@ -0,0 +1,6 @@
+glossary:
+  - name: Audita
+    aliases:
+      - audita
+    category: product
+    summary: The Audita transcript correction CLI.

examples/tiny-transcript.json

@@ -0,0 +1,9 @@
+[
+  {
+    "id": 1,
+    "speaker": "A",
+    "start": 0.0,
+    "end": 1.2,
+    "text": "hello world"
+  }
+]

go.mod

@@ -1,11 +1,20 @@
 module gitea.maximumdirect.net/eric/audita
 
-go 1.24.0
-
-require gopkg.in/yaml.v3 v3.0.1
+go 1.25.5
 
 require (
+	gitea.maximumdirect.net/eric/ConfigKit v0.4.0
+	gitea.maximumdirect.net/eric/promptkit v0.11.0
+	gopkg.in/yaml.v3 v3.0.1
+)
+
+require (
+	gitea.maximumdirect.net/eric/promptkit-backend-openrouter/v2 v2.20260922.1 // indirect
+	gitea.maximumdirect.net/eric/promptkit-backend-rakestrawhome/v2 v2.20260922.0 // indirect
 	github.com/kr/pretty v0.3.1 // indirect
 	github.com/rogpeppe/go-internal v1.14.1 // indirect
+	github.com/santhosh-tekuri/jsonschema/v6 v6.0.2 // indirect
+	go.yaml.in/yaml/v4 v4.0.0-rc.6 // indirect
+	golang.org/x/text v0.14.0 // indirect
 	gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c // indirect
 )

go.sum

@@ -1,4 +1,14 @@
+gitea.maximumdirect.net/eric/ConfigKit v0.4.0 h1:fo+vKff3zTGeWqJIA38NtT84b6OUP//9ZmBH3lFuKgk=
+gitea.maximumdirect.net/eric/ConfigKit v0.4.0/go.mod h1:tYWpsrbLyu05r8+1MapWp5fN6ZrX9I8hU2xEPedYmYk=
+gitea.maximumdirect.net/eric/promptkit v0.11.0 h1:x7JguOUiInPVXK0GEitjyoRFqqvcuSTjYUvqJ+whtik=
+gitea.maximumdirect.net/eric/promptkit v0.11.0/go.mod h1:0zEBVlJJC+ldDdXazAgybYBXNSe74vwFx5dyOuSos/w=
+gitea.maximumdirect.net/eric/promptkit-backend-openrouter/v2 v2.20260922.1 h1:S2/Q/JDE+m8Ca3pDhAjVbfYrdlEhs9blj0STgNopHo0=
+gitea.maximumdirect.net/eric/promptkit-backend-openrouter/v2 v2.20260922.1/go.mod h1:a3VBbh9oaGXD6ruxhtwHj+AWW6DY/SSyby1QtMmYSUo=
+gitea.maximumdirect.net/eric/promptkit-backend-rakestrawhome/v2 v2.20260922.0 h1:Y0EXc3eh12eKW2NrOviVqdk3oBo6UUjopt1YAa1B3Oc=
+gitea.maximumdirect.net/eric/promptkit-backend-rakestrawhome/v2 v2.20260922.0/go.mod h1:mtgD1E0SGm0e/I8Ffhz+AYX7GN7RKUWJXehNFIKHs4Y=
 github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
+github.com/dlclark/regexp2 v1.11.0 h1:G/nrcoOa7ZXlpoa/91N3X7mM3r8eIlMBBJZvsz/mxKI=
+github.com/dlclark/regexp2 v1.11.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
 github.com/kr/pretty v0.2.1/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
 github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
 github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
@@ -10,6 +20,12 @@ github.com/pkg/diff v0.0.0-20210226163009-20ebb0f2a09e/go.mod h1:pJLUxLENpZxwdsK
 github.com/rogpeppe/go-internal v1.9.0/go.mod h1:WtVeX8xhTBvf0smdhujwtBcq4Qrzq/fJaraNFVN+nFs=
 github.com/rogpeppe/go-internal v1.14.1 h1:UQB4HGPB6osV0SQTLymcB4TgvyWu6ZyliaW0tI/otEQ=
 github.com/rogpeppe/go-internal v1.14.1/go.mod h1:MaRKkUm5W0goXpeCfT7UZI6fk/L7L7so1lCWt35ZSgc=
+github.com/santhosh-tekuri/jsonschema/v6 v6.0.2 h1:KRzFb2m7YtdldCEkzs6KqmJw4nqEVZGK7IN2kJkjTuQ=
+github.com/santhosh-tekuri/jsonschema/v6 v6.0.2/go.mod h1:JXeL+ps8p7/KNMjDQk3TCwPpBy0wYklyWTfbkIzdIFU=
+go.yaml.in/yaml/v4 v4.0.0-rc.6 h1:1h7H1ohdUh93/FyE4YaDa1Zh64K6VVbjF4K6WUxMtH4=
+go.yaml.in/yaml/v4 v4.0.0-rc.6/go.mod h1:aZqd9kCMsGL7AuUv/m/PvWLdg5sjJsZ4oHDEnfPPfY0=
+golang.org/x/text v0.14.0 h1:ScX5w1eTa3QqT8oi6+ziP7dTV1S2+ALU0bI+0zXKWiQ=
+golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=

internal/cli/config_input.go

@@ -0,0 +1,222 @@
+package cli
+
+import (
+	"errors"
+	"flag"
+	"fmt"
+	"io"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+
+	ck "gitea.maximumdirect.net/eric/ConfigKit"
+	"gitea.maximumdirect.net/eric/audita/internal/core/config"
+)
+
+var systemConfigPaths = []string{"/usr/local/etc/audita/config.yml", "/etc/audita/config.yml"}
+
+type configInput struct {
+	ConfigPath     string
+	ConfigPathSet  bool
+	Overlay        string
+	View           string
+	TranscriptPath string
+	GlossaryPath   string
+	OutputPath     string
+	Overrides      []ck.Override
+}
+
+type selectedConfig struct {
+	Path   string
+	Source string
+	Result *ck.Result[config.Config]
+}
+
+// parseConfigInput owns the common setting flags and process-only path flags.
+// A transcript before flags and one after flags are both accepted.
+func parseConfigInput(command string, args []string, output io.Writer) (configInput, bool, error) {
+	var input configInput
+	fs := flag.NewFlagSet(command, flag.ContinueOnError)
+	fs.SetOutput(output)
+	configPath := fs.String("config", "", "Main configuration YAML file")
+	overlay := fs.String("overlay", "", "Selected configuration overlay")
+	if command == "print-effective" {
+		fs.String("view", "settings", "Effective view: settings, profiles, or explain")
+	}
+	settingPaths := map[string]ck.Path{
+		"proposal-profile":                 ck.Keys("llm", "proposal_profile"),
+		"validation-profile":               ck.Keys("llm", "validation_profile"),
+		"output-schema":                    ck.Keys("output", "schema"),
+		"validation-max-prompt-tokens":     ck.Keys("validation", "max_prompt_tokens"),
+		"target-sections":                  ck.Keys("chunking", "target_sections"),
+		"max-section-tokens":               ck.Keys("chunking", "max_section_tokens"),
+		"min-section-tokens":               ck.Keys("chunking", "min_section_tokens"),
+		"normalize-max-segment-gap":        ck.Keys("normalization", "max_segment_gap"),
+		"normalize-ellipsis-gap":           ck.Keys("normalization", "ellipsis_gap"),
+		"normalize-max-segment-duration":   ck.Keys("normalization", "max_segment_duration"),
+		"normalize-max-segment-tokens":     ck.Keys("normalization", "max_segment_tokens"),
+		"glossary-confidence-threshold":    ck.Keys("thresholds", "glossary"),
+		"homophones-confidence-threshold":  ck.Keys("thresholds", "homophones"),
+		"spoken-word-confidence-threshold": ck.Keys("thresholds", "spoken_word"),
+		"grammar-confidence-threshold":     ck.Keys("thresholds", "grammar"),
+		"transcript-description":           ck.Keys("context", "description"),
+		"reporting":                        ck.Keys("reporting", "enabled"),
+		"report-dir":                       ck.Keys("reporting", "stdout_dir"),
+		"debug":                            ck.Keys("diagnostics", "debug"),
+		"work-dir":                         ck.Keys("diagnostics", "work_dir"),
+		"modules":                          ck.Keys("pipeline", "modules"),
+	}
+	for name := range settingPaths {
+		switch name {
+		case "reporting", "debug":
+			fs.Bool(name, false, "Setting override")
+		case "validation-max-prompt-tokens", "target-sections", "max-section-tokens", "min-section-tokens", "normalize-max-segment-tokens":
+			fs.Int(name, 0, "Setting override")
+		case "normalize-max-segment-gap", "normalize-ellipsis-gap", "normalize-max-segment-duration", "glossary-confidence-threshold", "homophones-confidence-threshold", "spoken-word-confidence-threshold", "grammar-confidence-threshold":
+			fs.Float64(name, 0, "Setting override")
+		default:
+			fs.String(name, "", "Setting override")
+		}
+	}
+	if command == "process" {
+		fs.String("glossary", "", "Glossary YAML file")
+		fs.String("output", "", "Corrected transcript output file")
+	}
+	parseArgs := args
+	front := ""
+	if command == "process" && len(args) > 0 && !strings.HasPrefix(args[0], "-") {
+		front, parseArgs = args[0], args[1:]
+	}
+	if err := fs.Parse(parseArgs); err != nil {
+		if errors.Is(err, flag.ErrHelp) {
+			return configInput{}, true, nil
+		}
+		return configInput{}, false, err
+	}
+	positions := fs.Args()
+	if front != "" {
+		positions = append([]string{front}, positions...)
+	}
+	if command == "process" {
+		if len(positions) != 1 {
+			return configInput{}, false, fmt.Errorf("expected exactly one transcript JSON path")
+		}
+		input.TranscriptPath = positions[0]
+		input.GlossaryPath = fs.Lookup("glossary").Value.String()
+		input.OutputPath = fs.Lookup("output").Value.String()
+		if strings.TrimSpace(input.GlossaryPath) == "" {
+			return configInput{}, false, fmt.Errorf("--glossary is required")
+		}
+	} else if len(positions) != 0 {
+		return configInput{}, false, fmt.Errorf("unexpected positional arguments")
+	}
+	input.ConfigPath, input.Overlay = *configPath, *overlay
+	if command == "print-effective" {
+		input.View = fs.Lookup("view").Value.String()
+		switch input.View {
+		case "settings", "profiles", "explain":
+		default:
+			return configInput{}, false, fmt.Errorf("unknown effective view %q", input.View)
+		}
+	}
+	fs.Visit(func(f *flag.Flag) {
+		if f.Name == "config" {
+			input.ConfigPathSet = true
+		}
+		path, ok := settingPaths[f.Name]
+		if !ok {
+			return
+		}
+		var value any = f.Value.String()
+		switch f.Name {
+		case "reporting", "debug":
+			value, _ = strconv.ParseBool(f.Value.String())
+		case "modules":
+			modules, err := config.ParseModulesCSV(f.Value.String())
+			if err != nil {
+				value = err
+			} else {
+				values := make([]any, len(modules))
+				for i, module := range modules {
+					values[i] = module
+				}
+				value = values
+			}
+		case "validation-max-prompt-tokens", "target-sections", "max-section-tokens", "min-section-tokens", "normalize-max-segment-tokens":
+			value, _ = strconv.Atoi(f.Value.String())
+		case "normalize-max-segment-gap", "normalize-ellipsis-gap", "normalize-max-segment-duration", "glossary-confidence-threshold", "homophones-confidence-threshold", "spoken-word-confidence-threshold", "grammar-confidence-threshold":
+			value, _ = strconv.ParseFloat(f.Value.String(), 64)
+		}
+		input.Overrides = append(input.Overrides, ck.Override{Path: path, Value: value, Flag: "--" + f.Name})
+	})
+	for _, override := range input.Overrides {
+		if err, ok := override.Value.(error); ok {
+			return configInput{}, false, fmt.Errorf("%s: %w", override.Flag, err)
+		}
+	}
+	return input, false, nil
+}
+
+func selectConfigPath(explicit string, set bool, stat func(string) (os.FileInfo, error), candidates []string) (string, string, error) {
+	if set {
+		if strings.TrimSpace(explicit) == "" {
+			return "", "", fmt.Errorf("--config requires a nonempty path")
+		}
+		if _, err := stat(explicit); err != nil {
+			return "", "", fmt.Errorf("config file %q: %w", explicit, err)
+		}
+		return explicit, "flag", nil
+	}
+	for _, path := range candidates {
+		if _, err := stat(path); err == nil {
+			return path, "default", nil
+		} else if !os.IsNotExist(err) {
+			return "", "", fmt.Errorf("config file %q: %w", path, err)
+		}
+	}
+	return "", "", fmt.Errorf("no configuration file found; select a config with backend and profile settings")
+}
+
+func loadConfigInput(input configInput, command string) (selectedConfig, error) {
+	if command == "validate" && (!input.ConfigPathSet || strings.TrimSpace(input.ConfigPath) == "") {
+		return selectedConfig{}, fmt.Errorf("--config is required")
+	}
+	path, source, err := selectConfigPath(input.ConfigPath, input.ConfigPathSet, os.Stat, systemConfigPaths)
+	if err != nil {
+		return selectedConfig{}, err
+	}
+	loaded, err := config.Load(ck.File(path), input.Overlay, input.Overrides...)
+	if err != nil {
+		return selectedConfig{}, err
+	}
+	return selectedConfig{Path: path, Source: source, Result: loaded}, nil
+}
+
+func configInputError(err error) string {
+	var loadErr *ck.LoadError
+	if !errors.As(err, &loadErr) {
+		return err.Error()
+	}
+	var messages []string
+	for _, diagnostic := range loadErr.Diagnostics() {
+		messages = append(messages, diagnostic.Message)
+	}
+	if len(messages) == 0 {
+		return err.Error()
+	}
+	return strings.Join(messages, "; ")
+}
+
+func selectedReportDir(cfg config.Config, outputPath string) (string, error) {
+	if !cfg.Reporting.Enabled {
+		return "", nil
+	}
+	if outputPath != "" {
+		return filepath.Dir(outputPath), nil
+	}
+	if strings.TrimSpace(cfg.Reporting.StdoutDir) == "" {
+		return "", fmt.Errorf("report directory is required when reporting is enabled and transcript output is stdout")
+	}
+	return cfg.Reporting.StdoutDir, nil
+}

internal/cli/config_input_test.go

@@ -0,0 +1,250 @@
+package cli
+
+import (
+	"errors"
+	"io"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	ck "gitea.maximumdirect.net/eric/ConfigKit"
+	"gitea.maximumdirect.net/eric/audita/internal/core/config"
+)
+
+func TestSharedConfigInput(t *testing.T) {
+	for _, command := range []string{"process", "validate", "print-effective"} {
+		t.Run(command, func(t *testing.T) {
+			args := []string{"--config", "main.yml", "--overlay", "testing", "--proposal-profile", "economy", "--modules", "", "--target-sections", "0"}
+			if command == "process" {
+				args = append([]string{"transcript.json", "--glossary", "glossary.yml"}, args...)
+			}
+			input, help, err := parseConfigInput(command, args, io.Discard)
+			if err != nil || help {
+				t.Fatalf("parse: %+v %v %v", input, help, err)
+			}
+			if !input.ConfigPathSet || input.ConfigPath != "main.yml" || input.Overlay != "testing" || len(input.Overrides) != 3 {
+				t.Fatalf("selection: %+v", input)
+			}
+			for _, override := range input.Overrides {
+				switch override.Path.String() {
+				case ck.Keys("pipeline", "modules").String():
+					if values, ok := override.Value.([]any); !ok || len(values) != 0 {
+						t.Fatalf("empty modules: %#v", override.Value)
+					}
+				case ck.Keys("chunking", "target_sections").String():
+					if override.Value != 0 {
+						t.Fatalf("zero target: %#v", override.Value)
+					}
+				}
+			}
+		})
+	}
+}
+
+func TestConfigInputRejectsRemovedAndProcessOnlyFlags(t *testing.T) {
+	for _, command := range []string{"process", "validate", "print-effective"} {
+		for _, flag := range []string{"--model", "--llm-api-key", "--max-retries", "--llm-concurrency", "--report-json", "--work-dir-retention"} {
+			args := []string{flag, "value"}
+			if command == "process" {
+				args = append([]string{"transcript.json", "--glossary", "glossary.yml"}, args...)
+			}
+			if _, _, err := parseConfigInput(command, args, io.Discard); err == nil {
+				t.Fatalf("%s accepted %s", command, flag)
+			}
+		}
+	}
+	for _, command := range []string{"validate", "print-effective"} {
+		if _, _, err := parseConfigInput(command, []string{"--glossary", "glossary.yml"}, io.Discard); err == nil {
+			t.Fatalf("%s accepted process path", command)
+		}
+		if _, _, err := parseConfigInput(command, []string{"transcript.json"}, io.Discard); err == nil {
+			t.Fatalf("%s accepted positional", command)
+		}
+	}
+	for _, command := range []string{"process", "validate", "print-effective"} {
+		_, help, err := parseConfigInput(command, []string{"--help"}, io.Discard)
+		if err != nil || !help {
+			t.Fatalf("%s help: %v %v", command, help, err)
+		}
+	}
+}
+
+func TestConfigSelection(t *testing.T) {
+	dir := t.TempDir()
+	first, second := filepath.Join(dir, "first.yml"), filepath.Join(dir, "second.yml")
+	if err := os.WriteFile(second, []byte("version: 2\n"), 0600); err != nil {
+		t.Fatal(err)
+	}
+	path, source, err := selectConfigPath("", false, os.Stat, []string{first, second})
+	if err != nil || path != second || source != "default" {
+		t.Fatalf("discovery: %q %q %v", path, source, err)
+	}
+	for _, tc := range []struct {
+		explicit   string
+		set        bool
+		candidates []string
+	}{
+		{"", true, []string{second}}, {first, true, []string{second}}, {"", false, []string{first}},
+	} {
+		if _, _, err := selectConfigPath(tc.explicit, tc.set, os.Stat, tc.candidates); err == nil {
+			t.Fatalf("accepted unavailable selection: %+v", tc)
+		}
+	}
+	statDenied := func(string) (os.FileInfo, error) { return nil, os.ErrPermission }
+	if _, _, err := selectConfigPath("", false, statDenied, []string{first}); !errors.Is(err, os.ErrPermission) {
+		t.Fatalf("inaccessible candidate: %v", err)
+	}
+	path, source, err = selectConfigPath(second, true, os.Stat, []string{first})
+	if err != nil || path != second || source != "flag" {
+		t.Fatalf("explicit path: %q %q %v", path, source, err)
+	}
+}
+
+func TestLegacyEnvironmentDoesNotSelectConfig(t *testing.T) {
+	t.Setenv("AUDITA_CONFIG", filepath.Join(t.TempDir(), "legacy.yml"))
+	_, _, err := selectConfigPath("", false, os.Stat, []string{filepath.Join(t.TempDir(), "missing.yml")})
+	if err == nil || !strings.Contains(err.Error(), "no configuration file") {
+		t.Fatalf("legacy environment changed discovery: %v", err)
+	}
+}
+
+func TestEffectiveViewParsing(t *testing.T) {
+	for _, view := range []string{"settings", "profiles", "explain"} {
+		input, _, err := parseConfigInput("print-effective", []string{"--view", view}, io.Discard)
+		if err != nil || input.View != view {
+			t.Fatalf("%q: %+v %v", view, input, err)
+		}
+	}
+	if _, _, err := parseConfigInput("print-effective", []string{"--view", "raw"}, io.Discard); err == nil {
+		t.Fatal("unknown view accepted")
+	}
+	if _, _, err := parseConfigInput("validate", []string{"--view", "profiles"}, io.Discard); err == nil {
+		t.Fatal("validate accepted inspection-only flag")
+	}
+}
+
+func TestSharedLoadingUsesSelectedFileOverlayAndVisitedOverrides(t *testing.T) {
+	dir := t.TempDir()
+	if err := os.Mkdir(filepath.Join(dir, "overlays"), 0700); err != nil {
+		t.Fatal(err)
+	}
+	main := filepath.Join(dir, "config.yml")
+	base := "version: 2\nbackends:\n  local:\n    endpoint: https://local.invalid/v1\nllm:\n  proposal_profile: base\nchunking:\n  max_section_tokens: 900\n"
+	if err := os.WriteFile(main, []byte(base), 0600); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(filepath.Join(dir, "overlays", "testing.yml"), []byte("llm:\n  proposal_profile: testing\n"), 0600); err != nil {
+		t.Fatal(err)
+	}
+	for _, command := range []string{"process", "validate", "print-effective"} {
+		args := []string{"--config", main, "--overlay", "testing", "--min-section-tokens", "500"}
+		if command == "process" {
+			args = append([]string{"input.json", "--glossary", "glossary.yml"}, args...)
+		}
+		input, _, err := parseConfigInput(command, args, io.Discard)
+		if err != nil {
+			t.Fatal(err)
+		}
+		selected, err := loadConfigInput(input, command)
+		if err != nil {
+			t.Fatal(err)
+		}
+		c := selected.Result.Config
+		if c.LLM.ProposalProfile != "testing" || c.LLM.ValidationProfile != "testing" || c.Chunking.MinSectionTokens != 500 || c.Chunking.MaxSectionTokens != 900 {
+			t.Fatalf("%s resolved %+v", command, c)
+		}
+	}
+	input, _, err := parseConfigInput("validate", []string{"--overlay", "testing"}, io.Discard)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := loadConfigInput(input, "validate"); err == nil {
+		t.Fatal("validate accepted discovery")
+	}
+	input, _, err = parseConfigInput("validate", []string{"--config", main}, io.Discard)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := loadConfigInput(input, "validate"); err == nil {
+		t.Fatal("unmodified invalid lower-precedence section relation accepted")
+	}
+}
+
+func TestReportingAndDebugOverridesRemainIndependent(t *testing.T) {
+	dir := t.TempDir()
+	if err := os.Mkdir(filepath.Join(dir, "overlays"), 0700); err != nil {
+		t.Fatal(err)
+	}
+	main := filepath.Join(dir, "config.yml")
+	base := "version: 2\nbackends:\n  local:\n    endpoint: https://local.invalid/v1\nllm:\n  proposal_profile: base\n"
+	if err := os.WriteFile(main, []byte(base), 0600); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(filepath.Join(dir, "overlays", "testing.yml"), []byte("reporting:\n  enabled: true\n  stdout_dir: reports\ndiagnostics:\n  debug: true\n"), 0600); err != nil {
+		t.Fatal(err)
+	}
+	defaultInput, _, err := parseConfigInput("validate", []string{"--config", main}, io.Discard)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defaults, err := loadConfigInput(defaultInput, "validate")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !defaults.Result.Config.Reporting.Enabled || defaults.Result.Config.Reporting.StdoutDir != "." || defaults.Result.Config.Diagnostics.Debug {
+		t.Fatalf("unexpected defaults: %+v", defaults.Result.Config)
+	}
+	for _, command := range []string{"process", "validate", "print-effective"} {
+		args := []string{"--config", main, "--overlay", "testing", "--reporting=false", "--debug=false"}
+		if command == "process" {
+			args = append([]string{"input.json", "--glossary", "glossary.yml"}, args...)
+		}
+		input, _, err := parseConfigInput(command, args, io.Discard)
+		if err != nil {
+			t.Fatal(err)
+		}
+		if len(input.Overrides) != 2 || input.Overrides[0].Value != false || input.Overrides[1].Value != false {
+			t.Fatalf("boolean overrides must remain typed: %+v", input.Overrides)
+		}
+		selected, err := loadConfigInput(input, command)
+		if err != nil {
+			t.Fatal(err)
+		}
+		if selected.Result.Config.Reporting.Enabled || selected.Result.Config.Diagnostics.Debug || selected.Result.Config.Reporting.StdoutDir != "reports" {
+			t.Fatalf("%s did not honor explicit false: %+v", command, selected.Result.Config)
+		}
+	}
+}
+
+func TestReportDirectorySelectionOnlyUsesActiveDestination(t *testing.T) {
+	cfg := config.Config{Reporting: config.Reporting{Enabled: true, StdoutDir: "reports"}}
+	if got, err := selectedReportDir(cfg, ""); err != nil || got != "reports" {
+		t.Fatalf("stdout destination: %q %v", got, err)
+	}
+	cfg.Reporting.StdoutDir = ""
+	if got, err := selectedReportDir(cfg, filepath.Join("out", "corrected.json")); err != nil || got != "out" {
+		t.Fatalf("file destination ignored stdout setting: %q %v", got, err)
+	}
+	if _, err := selectedReportDir(cfg, ""); err == nil {
+		t.Fatal("enabled stdout reporting accepted blank destination")
+	}
+	cfg.Reporting.Enabled = false
+	if got, err := selectedReportDir(cfg, ""); err != nil || got != "" {
+		t.Fatalf("disabled reporting probed destination: %q %v", got, err)
+	}
+}
+
+func TestBareBooleanFlagsAndReportDirectoryOverride(t *testing.T) {
+	input, _, err := parseConfigInput("process", []string{"input.json", "--glossary", "glossary.yml", "--reporting", "--debug", "--report-dir", "reports"}, io.Discard)
+	if err != nil {
+		t.Fatal(err)
+	}
+	values := map[string]any{}
+	for _, override := range input.Overrides {
+		values[override.Path.String()] = override.Value
+	}
+	if values[ck.Keys("reporting", "enabled").String()] != true || values[ck.Keys("diagnostics", "debug").String()] != true || values[ck.Keys("reporting", "stdout_dir").String()] != "reports" {
+		t.Fatalf("incorrect visited overrides: %+v", values)
+	}
+}

internal/cli/parity_test.go

@@ -1,338 +0,0 @@
-package cli
-
-import (
-	"bytes"
-	"encoding/json"
-	"os"
-	"path/filepath"
-	"reflect"
-	"strings"
-	"testing"
-
-	"gitea.maximumdirect.net/eric/audita/internal/core/reporting"
-)
-
-type parityFixtureCase struct {
-	Name                    string            `json:"name"`
-	TranscriptFile          string            `json:"transcript_file"`
-	GlossaryFile            string            `json:"glossary_file"`
-	ModulesCSV              string            `json:"modules_csv,omitempty"`
-	ProposalResponsesFile   string            `json:"proposal_responses_file,omitempty"`
-	ValidationResponsesFile string            `json:"validation_responses_file,omitempty"`
-	Env                     map[string]string `json:"env,omitempty"`
-	Expect                  parityExpectation `json:"expect"`
-}
-
-type parityExpectation struct {
-	ExitCode                     int      `json:"exit_code"`
-	Status                       string   `json:"status,omitempty"`
-	ErrorPhase                   string   `json:"error_phase,omitempty"`
-	StdoutMode                   string   `json:"stdout_mode,omitempty"` // empty|json
-	StderrContains               string   `json:"stderr_contains,omitempty"`
-	OutputTranscriptFile         string   `json:"output_transcript_file,omitempty"`
-	ModuleInstances              []string `json:"module_instances,omitempty"`
-	ModuleCount                  int      `json:"module_count,omitempty"`
-	TotalAppliedChanges          int      `json:"total_applied_changes,omitempty"`
-	TotalSkippedChanges          int      `json:"total_skipped_changes,omitempty"`
-	FailedModuleInstance         string   `json:"failed_module_instance,omitempty"`
-	ValidatorRejectedReasonCodes []string `json:"validator_rejected_reason_codes,omitempty"`
-	ApplicationSkipReasonCodes   []string `json:"application_skip_reason_codes,omitempty"`
-	RequireErrorLog              bool     `json:"require_error_log,omitempty"`
-	SecretMarkers                []string `json:"secret_markers,omitempty"`
-	ExpectedProposalCalls        []string `json:"expected_proposal_calls,omitempty"`
-	ExpectedValidationCalls      []string `json:"expected_validation_calls,omitempty"`
-	ModuleAppliedCounts          []int    `json:"module_applied_counts,omitempty"`
-	ModuleRejectedCounts         []int    `json:"module_rejected_counts,omitempty"`
-	ModuleSkipCounts             []int    `json:"module_skip_counts,omitempty"`
-	MinResponsePayloadArtifacts  int      `json:"min_response_payload_artifacts,omitempty"`
-}
-
-func TestParityFixtures(t *testing.T) {
-	casePaths, err := filepath.Glob(parityFixturePath("*.case.json"))
-	if err != nil {
-		t.Fatalf("glob parity fixtures: %v", err)
-	}
-	if len(casePaths) == 0 {
-		t.Fatal("expected at least one parity fixture case")
-	}
-
-	for _, casePath := range casePaths {
-		fx := loadParityFixtureCase(t, casePath)
-		t.Run(fx.Name, func(t *testing.T) {
-			runParityFixtureCase(t, filepath.Dir(casePath), fx)
-		})
-	}
-}
-
-func loadParityFixtureCase(t *testing.T, casePath string) parityFixtureCase {
-	t.Helper()
-	var fx parityFixtureCase
-	raw := readFile(t, casePath)
-	if err := json.Unmarshal(raw, &fx); err != nil {
-		t.Fatalf("parse parity case %q: %v", casePath, err)
-	}
-	if strings.TrimSpace(fx.Name) == "" {
-		t.Fatalf("parity case %q missing name", casePath)
-	}
-	return fx
-}
-
-func runParityFixtureCase(t *testing.T, caseDir string, fx parityFixtureCase) {
-	t.Helper()
-	for k, v := range fx.Env {
-		t.Setenv(k, v)
-	}
-
-	proposalClient := &fakeStructuredLLMClient{}
-	validationClient := &fakeStructuredLLMClient{}
-
-	if strings.TrimSpace(fx.ProposalResponsesFile) != "" {
-		raw := readFile(t, filepath.Join(caseDir, fx.ProposalResponsesFile))
-		if err := json.Unmarshal(raw, &proposalClient.proposalResponses); err != nil {
-			t.Fatalf("parse proposal responses: %v", err)
-		}
-		processProposalLLMClient = proposalClient
-	}
-	if strings.TrimSpace(fx.ValidationResponsesFile) != "" {
-		raw := readFile(t, filepath.Join(caseDir, fx.ValidationResponsesFile))
-		if err := json.Unmarshal(raw, &validationClient.validationResponses); err != nil {
-			t.Fatalf("parse validation responses: %v", err)
-		}
-		processValidationLLMClient = validationClient
-	}
-	t.Cleanup(func() {
-		processProposalLLMClient = nil
-		processValidationLLMClient = nil
-	})
-
-	var stdout, stderr bytes.Buffer
-	workDir := t.TempDir()
-	outputPath := filepath.Join(t.TempDir(), "out.json")
-	reportPath := filepath.Join(t.TempDir(), "report.json")
-	args := []string{
-		"process",
-		filepath.Join(caseDir, fx.TranscriptFile),
-		"--glossary",
-		filepath.Join(caseDir, fx.GlossaryFile),
-		"--report-json",
-		reportPath,
-		"--work-dir",
-		workDir,
-		"--work-dir-retention",
-		"always",
-	}
-	// Keep stdout shape deterministic for parity tests.
-	if fx.Expect.StdoutMode != "json" {
-		args = append(args, "--output", outputPath)
-	}
-	if strings.TrimSpace(fx.ModulesCSV) != "" {
-		args = append(args, "--modules", fx.ModulesCSV)
-	}
-
-	exitCode := Run(args, &stdout, &stderr)
-	if exitCode != fx.Expect.ExitCode {
-		t.Fatalf("expected exit code %d, got %d stderr=%q", fx.Expect.ExitCode, exitCode, stderr.String())
-	}
-
-	switch fx.Expect.StdoutMode {
-	case "json":
-		if _, err := json.Marshal(stdout.String()); err != nil {
-			t.Fatalf("unexpected stdout marshal error: %v", err)
-		}
-		if !json.Valid(stdout.Bytes()) {
-			t.Fatalf("expected JSON stdout, got %q", stdout.String())
-		}
-	default:
-		if stdout.Len() != 0 {
-			t.Fatalf("expected empty stdout, got %q", stdout.String())
-		}
-	}
-
-	if fx.Expect.StderrContains != "" && !strings.Contains(stderr.String(), fx.Expect.StderrContains) {
-		t.Fatalf("expected stderr to contain %q, got %q", fx.Expect.StderrContains, stderr.String())
-	}
-
-	report := readProcessReport(t, reportPath)
-	assertParityReport(t, report, fx.Expect)
-
-	runDir := onlyRunDir(t, workDir)
-	runDirReport := readProcessReport(t, filepath.Join(runDir, "report.json"))
-	assertParityReport(t, runDirReport, fx.Expect)
-
-	if fx.Expect.RequireErrorLog {
-		if _, err := os.Stat(filepath.Join(runDir, "error.log")); err != nil {
-			t.Fatalf("expected error.log: %v", err)
-		}
-	}
-
-	if strings.TrimSpace(fx.Expect.OutputTranscriptFile) != "" && fx.Expect.ExitCode == 0 {
-		got := readFile(t, outputPath)
-		want := readFile(t, filepath.Join(caseDir, fx.Expect.OutputTranscriptFile))
-		assertJSONSemanticEqual(t, want, got)
-	}
-
-	if len(fx.Expect.ExpectedProposalCalls) > 0 && !reflect.DeepEqual(proposalClient.calls, fx.Expect.ExpectedProposalCalls) {
-		t.Fatalf("unexpected proposal calls: got %v want %v", proposalClient.calls, fx.Expect.ExpectedProposalCalls)
-	}
-	if len(fx.Expect.ExpectedValidationCalls) > 0 && !reflect.DeepEqual(validationClient.calls, fx.Expect.ExpectedValidationCalls) {
-		t.Fatalf("unexpected validation calls: got %v want %v", validationClient.calls, fx.Expect.ExpectedValidationCalls)
-	}
-
-	if len(fx.Expect.SecretMarkers) > 0 {
-		assertNoSecretMarkers(t, reportPath, fx.Expect.SecretMarkers)
-		assertNoSecretMarkersInTree(t, runDir, fx.Expect.SecretMarkers)
-	}
-	if fx.Expect.MinResponsePayloadArtifacts > 0 {
-		matches, err := filepath.Glob(filepath.Join(runDir, "*", "*response-payload.json"))
-		if err != nil {
-			t.Fatalf("glob response payload artifacts: %v", err)
-		}
-		if len(matches) < fx.Expect.MinResponsePayloadArtifacts {
-			t.Fatalf("expected at least %d response payload artifacts, got %d", fx.Expect.MinResponsePayloadArtifacts, len(matches))
-		}
-	}
-}
-
-func assertParityReport(t *testing.T, report reporting.ProcessReport, exp parityExpectation) {
-	t.Helper()
-	if exp.Status != "" && report.Status != exp.Status {
-		t.Fatalf("expected report status %q, got %q", exp.Status, report.Status)
-	}
-	if exp.ErrorPhase != "" && report.ErrorPhase != exp.ErrorPhase {
-		t.Fatalf("expected report error_phase %q, got %q", exp.ErrorPhase, report.ErrorPhase)
-	}
-	if len(exp.ModuleInstances) > 0 {
-		got := make([]string, 0, len(report.ModuleResults))
-		for _, mr := range report.ModuleResults {
-			got = append(got, mr.ModuleInstance)
-		}
-		if !reflect.DeepEqual(got, exp.ModuleInstances) {
-			t.Fatalf("unexpected module instances: got %v want %v", got, exp.ModuleInstances)
-		}
-	}
-	if exp.ModuleCount > 0 {
-		if report.ModulesSummary == nil || report.ModulesSummary.ModuleCount != exp.ModuleCount {
-			t.Fatalf("expected module_count=%d, got %+v", exp.ModuleCount, report.ModulesSummary)
-		}
-	}
-	if exp.TotalAppliedChanges > 0 {
-		if report.ModulesSummary == nil || report.ModulesSummary.TotalAppliedChanges != exp.TotalAppliedChanges {
-			t.Fatalf("expected total_applied_changes=%d, got %+v", exp.TotalAppliedChanges, report.ModulesSummary)
-		}
-	}
-	if exp.TotalSkippedChanges > 0 {
-		if report.ModulesSummary == nil || report.ModulesSummary.TotalSkippedChanges != exp.TotalSkippedChanges {
-			t.Fatalf("expected total_skipped_changes=%d, got %+v", exp.TotalSkippedChanges, report.ModulesSummary)
-		}
-	}
-	if exp.FailedModuleInstance != "" {
-		if report.ModulesSummary == nil || report.ModulesSummary.FailedModuleInstance != exp.FailedModuleInstance {
-			t.Fatalf("expected failed_module_instance=%q, got %+v", exp.FailedModuleInstance, report.ModulesSummary)
-		}
-	}
-
-	if len(exp.ValidatorRejectedReasonCodes) > 0 {
-		got := collectValidatorRejectedReasonCodes(report.ModuleResults)
-		if !reflect.DeepEqual(got, exp.ValidatorRejectedReasonCodes) {
-			t.Fatalf("unexpected validator rejected reason codes: got %v want %v", got, exp.ValidatorRejectedReasonCodes)
-		}
-	}
-	if len(exp.ApplicationSkipReasonCodes) > 0 {
-		got := collectApplicationSkipReasonCodes(report.ModuleResults)
-		if !reflect.DeepEqual(got, exp.ApplicationSkipReasonCodes) {
-			t.Fatalf("unexpected application skip reason codes: got %v want %v", got, exp.ApplicationSkipReasonCodes)
-		}
-	}
-	if len(exp.ModuleAppliedCounts) > 0 {
-		got := make([]int, 0, len(report.ModuleResults))
-		for _, mr := range report.ModuleResults {
-			got = append(got, len(mr.AppliedChanges))
-		}
-		if !reflect.DeepEqual(got, exp.ModuleAppliedCounts) {
-			t.Fatalf("unexpected per-module applied counts: got %v want %v", got, exp.ModuleAppliedCounts)
-		}
-	}
-	if len(exp.ModuleRejectedCounts) > 0 {
-		got := make([]int, 0, len(report.ModuleResults))
-		for _, mr := range report.ModuleResults {
-			got = append(got, len(mr.ValidatorRejected))
-		}
-		if !reflect.DeepEqual(got, exp.ModuleRejectedCounts) {
-			t.Fatalf("unexpected per-module rejected counts: got %v want %v", got, exp.ModuleRejectedCounts)
-		}
-	}
-	if len(exp.ModuleSkipCounts) > 0 {
-		got := make([]int, 0, len(report.ModuleResults))
-		for _, mr := range report.ModuleResults {
-			got = append(got, len(mr.SkippedChanges))
-		}
-		if !reflect.DeepEqual(got, exp.ModuleSkipCounts) {
-			t.Fatalf("unexpected per-module skip counts: got %v want %v", got, exp.ModuleSkipCounts)
-		}
-	}
-}
-
-func collectValidatorRejectedReasonCodes(results []reporting.ModuleReport) []string {
-	out := make([]string, 0)
-	for _, mr := range results {
-		for _, vr := range mr.ValidatorRejected {
-			out = append(out, vr.ReasonCode)
-		}
-	}
-	return out
-}
-
-func collectApplicationSkipReasonCodes(results []reporting.ModuleReport) []string {
-	out := make([]string, 0)
-	for _, mr := range results {
-		for _, sk := range mr.SkippedChanges {
-			out = append(out, string(sk.SkipReason))
-		}
-	}
-	return out
-}
-
-func assertJSONSemanticEqual(t *testing.T, expected []byte, actual []byte) {
-	t.Helper()
-	var exp any
-	var act any
-	if err := json.Unmarshal(expected, &exp); err != nil {
-		t.Fatalf("unmarshal expected json: %v", err)
-	}
-	if err := json.Unmarshal(actual, &act); err != nil {
-		t.Fatalf("unmarshal actual json: %v", err)
-	}
-	if !reflect.DeepEqual(exp, act) {
-		t.Fatalf("JSON mismatch\nexpected=%s\nactual=%s", string(expected), string(actual))
-	}
-}
-
-func assertNoSecretMarkers(t *testing.T, filePath string, markers []string) {
-	t.Helper()
-	raw := string(readFile(t, filePath))
-	for _, marker := range markers {
-		if marker != "" && strings.Contains(raw, marker) {
-			t.Fatalf("secret marker %q leaked in %s", marker, filePath)
-		}
-	}
-}
-
-func assertNoSecretMarkersInTree(t *testing.T, root string, markers []string) {
-	t.Helper()
-	_ = filepath.WalkDir(root, func(path string, d os.DirEntry, err error) error {
-		if err != nil || d == nil || d.IsDir() {
-			return nil
-		}
-		raw := string(readFile(t, path))
-		for _, marker := range markers {
-			if marker != "" && strings.Contains(raw, marker) {
-				t.Fatalf("secret marker %q leaked in %s", marker, path)
-			}
-		}
-		return nil
-	})
-}
-
-func parityFixturePath(name string) string {
-	return filepath.Join("testdata", "parity", name)
-}

internal/cli/process.go

@@ -0,0 +1,442 @@
+package cli
+
+import (
+	"bytes"
+	"context"
+	"crypto/rand"
+	"encoding/hex"
+	"errors"
+	"fmt"
+	"io"
+	"strings"
+	"sync"
+	"time"
+
+	"gitea.maximumdirect.net/eric/audita/internal/core/chunking"
+	"gitea.maximumdirect.net/eric/audita/internal/core/config"
+	"gitea.maximumdirect.net/eric/audita/internal/core/diagnostics"
+	coreio "gitea.maximumdirect.net/eric/audita/internal/core/io"
+	"gitea.maximumdirect.net/eric/audita/internal/core/normalization"
+	"gitea.maximumdirect.net/eric/audita/internal/core/outputschema"
+	"gitea.maximumdirect.net/eric/audita/internal/core/redaction"
+	"gitea.maximumdirect.net/eric/audita/internal/core/reporting"
+	"gitea.maximumdirect.net/eric/audita/internal/core/schema"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/modules"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/outcomes"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/processreport"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/runner"
+	"gitea.maximumdirect.net/eric/audita/internal/framework/structuredoutput"
+)
+
+func runProcess(ctx context.Context, args []string, stdout, stderr io.Writer) int {
+	input, help, err := parseConfigInput("process", args, stderr)
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", configInputError(err))
+		return 2
+	}
+	if help {
+		writeProcessUsage(stdout)
+		return 0
+	}
+	selected, err := loadConfigInput(input, "process")
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", configInputError(err))
+		return 2
+	}
+	cfg := selected.Result.Config
+	reportDir, err := selectedReportDir(cfg, input.OutputPath)
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		return 2
+	}
+	if cfg.Diagnostics.Debug && strings.TrimSpace(cfg.Diagnostics.WorkDir) == "" {
+		fmt.Fprintln(stderr, "audita process: debug work directory is required")
+		return 2
+	}
+	if err := ctx.Err(); err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", err)
+		return 1
+	}
+	runtime, err := llm.NewRuntime(ctx, cfg, selected.Path, nil)
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		if ctx.Err() != nil || llm.IsAssetError(err) {
+			return 1
+		}
+		return 2
+	}
+	specs, err := contracts.ResolveModuleRunSpecs(cfg.Pipeline.Modules)
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		return 2
+	}
+	runID, err := generateRunID()
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: run ID: %s\n", llm.RedactError(err, cfg))
+		return 1
+	}
+	started := time.Now().UTC()
+	state := &processState{cfg: cfg, runtime: runtime, specs: specs, runID: runID, started: started,
+		transcriptPath: input.TranscriptPath, glossaryPath: input.GlossaryPath, outputPath: input.OutputPath,
+		selected: selected, overlay: input.Overlay, stderr: stderr, stdout: stdout, ctx: ctx}
+	protected := []string{input.GlossaryPath}
+	for _, source := range selected.Result.Sources() {
+		if source.Path != "" {
+			protected = append(protected, source.Path)
+		}
+	}
+	protected = append(protected, runtime.ProfilePaths...)
+	session, err := diagnostics.PrepareOutput(diagnostics.OutputSelection{InputTranscriptPath: input.TranscriptPath, TranscriptPath: input.OutputPath,
+		ReportDir: reportDir, ReportingEnabled: cfg.Reporting.Enabled, ProtectedPaths: protected})
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		return 2
+	}
+	state.session = session
+	initial := state.snapshot(reporting.StatusInProgress, nil, nil)
+	initialBundle, err := state.bundle(initial)
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		return 1
+	}
+	initFailure := []processreport.RunFailure{{Phase: "initialization", Reason: outcomes.ReasonInternalFailure}}
+	failedInit, err := state.bundle(state.snapshot(reporting.StatusFailed, nil, initFailure))
+	if err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		return 1
+	}
+	if err := session.Initialize(initialBundle, failedInit); err != nil {
+		fmt.Fprintf(stderr, "audita process: %s\n", llm.RedactError(err, cfg))
+		return 1
+	}
+	state.startDebug()
+	return state.execute()
+}
+
+type processState struct {
+	cfg            config.Config
+	runtime        *llm.Runtime
+	specs          []contracts.ModuleRunSpec
+	runID          string
+	started        time.Time
+	transcriptPath string
+	glossaryPath   string
+	outputPath     string
+	selected       selectedConfig
+	overlay        string
+	ctx            context.Context
+	stdout         io.Writer
+	stderr         io.Writer
+	session        *diagnostics.OutputSession
+	debugDir       *diagnostics.RunDirectory
+	debugWarnings  []string
+	debugMu        sync.Mutex
+	runOutput      *runner.RunOutput
+}
+
+func generateRunID() (string, error) {
+	var random [12]byte
+	if _, err := rand.Read(random[:]); err != nil {
+		return "", err
+	}
+	return "run-" + hex.EncodeToString(random[:]), nil
+}
+
+func (s *processState) snapshot(status string, completed *time.Time, failures []processreport.RunFailure) reporting.Snapshot {
+	return processreport.Build(processreport.BuildInput{RunID: s.runID, Status: status, StartedAt: s.started,
+		CompletedAt: completed, TranscriptPath: s.transcriptPath, GlossaryPath: s.glossaryPath,
+		OutputPath: s.outputPath, ModuleSpecs: s.specs, RunOutput: s.runOutput, RunFailures: failures})
+}
+
+func (s *processState) bundle(snapshot reporting.Snapshot) (diagnostics.ReportBundle, error) {
+	if !s.cfg.Reporting.Enabled {
+		return diagnostics.ReportBundle{}, nil
+	}
+	secrets := llm.ConfiguredSecrets(&s.cfg)
+	errorsYAML, err := reporting.YAML(snapshot.ErrorsDocument(), secrets...)
+	if err != nil {
+		return diagnostics.ReportBundle{}, err
+	}
+	recoveriesYAML, err := reporting.YAML(snapshot.RecoveriesDocument(), secrets...)
+	if err != nil {
+		return diagnostics.ReportBundle{}, err
+	}
+	return diagnostics.ReportBundle{Errors: errorsYAML, Recoveries: recoveriesYAML,
+		Markdown: reporting.Markdown(snapshot, secrets...)}, nil
+}
+
+func (s *processState) debugWrite(label string, write func(*diagnostics.RunDirectory) error) {
+	if s.debugDir == nil {
+		return
+	}
+	if err := write(s.debugDir); err != nil {
+		s.reportDiagnosticError(fmt.Errorf("%s: %w", label, err))
+	}
+}
+
+func (s *processState) startDebug() {
+	if !s.cfg.Diagnostics.Debug {
+		return
+	}
+	dir, err := diagnostics.NewRunDirectory(s.cfg.Diagnostics.WorkDir, s.runID, llm.ConfiguredSecrets(&s.cfg)...)
+	if err != nil {
+		s.reportDiagnosticError(fmt.Errorf("directory: %w", err))
+		return
+	}
+	s.debugDir = dir
+	s.debugWrite("invocation", func(d *diagnostics.RunDirectory) error {
+		return d.WriteInvocationMetadata(diagnostics.InvocationMetadata{
+			Operation: "process", TranscriptPath: s.transcriptPath, GlossaryPath: s.glossaryPath,
+			OutputPath: s.outputPath, ConfigPath: s.selected.Path, ConfigSource: s.selected.Source,
+			Overlay: s.overlay, ConfigVersion: &s.cfg.Version, TranscriptDescription: s.cfg.Context.Description,
+			Modules: append([]string(nil), s.cfg.Pipeline.Modules...), RunID: s.runID, StartedAt: s.started,
+		})
+	})
+	var effective bytes.Buffer
+	if err := s.selected.Result.WriteEffective(&effective); err == nil {
+		s.debugWrite("effective settings", func(d *diagnostics.RunDirectory) error { return d.WriteEffectiveConfig(effective.Bytes()) })
+	} else {
+		s.reportDiagnosticError(fmt.Errorf("effective settings: %w", err))
+	}
+	if origins, err := llm.ExplanationYAML(s.selected.Result); err == nil {
+		s.debugWrite("configuration origins", func(d *diagnostics.RunDirectory) error {
+			return d.WriteYAMLArtifact(diagnostics.ArtifactConfigOrigins, origins)
+		})
+	} else {
+		s.reportDiagnosticError(fmt.Errorf("configuration origins: %w", err))
+	}
+	if profiles, err := s.runtime.SelectedProfilesYAML(); err == nil {
+		s.debugWrite("selected profiles", func(d *diagnostics.RunDirectory) error {
+			return d.WriteYAMLArtifact(diagnostics.ArtifactSelectedProfiles, profiles)
+		})
+	} else {
+		s.reportDiagnosticError(fmt.Errorf("selected profiles: %w", err))
+	}
+}
+
+func (s *processState) debugPath() string {
+	if s.debugDir == nil {
+		return ""
+	}
+	return s.debugDir.Path()
+}
+
+func (s *processState) execute() int {
+	transcriptBytes, err := coreio.ReadRequiredFile(s.transcriptPath, "transcript")
+	if err != nil {
+		return s.fail("input", outcomes.ReasonInternalFailure, err)
+	}
+	glossaryBytes, err := coreio.ReadRequiredFile(s.glossaryPath, "glossary")
+	if err != nil {
+		return s.fail("input", outcomes.ReasonInternalFailure, err)
+	}
+	source, err := schema.ParseSourceTranscriptJSON(transcriptBytes)
+	if err != nil {
+		return s.fail("input", outcomes.ReasonInputInvalid, err)
+	}
+	glossary, err := schema.ParseGlossaryYAML(glossaryBytes)
+	if err != nil {
+		return s.fail("input", outcomes.ReasonInputInvalid, err)
+	}
+	s.debugWrite("source transcript", func(d *diagnostics.RunDirectory) error { return d.WriteSourceTranscript(source, transcriptBytes) })
+	normalizer := normalization.NewNormalizer(normalization.NormalizationConfig{
+		MaxSegmentGap: s.cfg.Normalization.MaxSegmentGap, EllipsisGap: s.cfg.Normalization.EllipsisGap,
+		MaxSegmentDuration: s.cfg.Normalization.MaxSegmentDuration, MaxSegmentTokens: s.cfg.Normalization.MaxSegmentTokens,
+	})
+	normalized, normSummary := normalizer.Normalize(sourceToCanonicalTranscript(source))
+	s.debugWrite("normalized transcript", func(d *diagnostics.RunDirectory) error { return d.WriteNormalizedTranscript(normalized) })
+	s.debugWrite("normalization summary", func(d *diagnostics.RunDirectory) error { return d.WriteNormalizationSummary(normSummary) })
+	if err := s.ctx.Err(); err != nil {
+		return s.fail("processing", outcomes.ReasonCanceled, err)
+	}
+	observations := llm.NewObservationCollector()
+	proposalClient, err := s.runtime.NewRoleClient("proposal", observations.Hooks())
+	if err != nil {
+		return s.fail("processing", outcomes.ReasonSetupFailure, err)
+	}
+	validationClient, err := s.runtime.NewRoleClient("validation", observations.Hooks())
+	if err != nil {
+		return s.fail("processing", outcomes.ReasonSetupFailure, err)
+	}
+	factory := modules.NewFactory(modules.Dependencies{Config: &s.cfg, Glossary: glossary, DiagnosticsDir: s.debugPath()})
+	result, runErr := runner.New(factory).Run(s.ctx, runner.RunInput{
+		Config: &s.cfg, Transcript: normalized, Glossary: glossary, ModuleSpecs: s.specs,
+		ProposalLLMClient: proposalClient, ProposalDiagnosticsDir: s.debugPath(),
+		ValidationLLMClient: validationClient, ValidationDiagnosticsDir: s.debugPath(),
+		CaptureCandidates:     s.debugDir != nil,
+		ReportDiagnosticError: s.reportDiagnosticError,
+	})
+	result.Observations = observations.Snapshot()
+	result.ObservationsAvailable = true
+	result.LLMSummary = llm.Summarize(result.Observations)
+	s.runOutput = &result
+	s.writeExecutionEvidence()
+	if len(result.Sections) > 0 || runErr == nil {
+		s.writeChunkSummary(result.Sections)
+	}
+	if runErr != nil {
+		return s.fail("processing", fatalReason(s.ctx, runErr), runErr)
+	}
+	encoder, err := outputschema.Resolve(s.cfg.Output.Schema)
+	if err != nil {
+		return s.fail("encoding", outcomes.ReasonSetupFailure, err)
+	}
+	outputBytes, err := encoder.Encoder(result.FinalTranscript)
+	if err != nil {
+		return s.fail("encoding", outcomes.ReasonInternalFailure, err)
+	}
+	return s.finish(outputBytes)
+}
+
+func (s *processState) writeChunkSummary(inventory []contracts.SectionMetadata) {
+	if s.debugDir == nil {
+		return
+	}
+	sections := make([]chunking.Section, len(inventory))
+	for i, metadata := range inventory {
+		sections[i] = chunking.Section{Index: metadata.Index, StartSegmentID: metadata.StartSegmentID,
+			EndSegmentID: metadata.EndSegmentID, EstimatedTokens: metadata.EstimatedTokens,
+			Segments: make([]schema.Segment, metadata.SegmentCount)}
+	}
+	config := chunking.ChunkingConfig{TargetSections: s.cfg.Chunking.TargetSections,
+		MaxSectionTokens: s.cfg.Chunking.MaxSectionTokens, MinSectionTokens: s.cfg.Chunking.MinSectionTokens}
+	summary := chunking.ComputeDetailedSummary(sections, config)
+	s.debugWrite("chunking summary", func(d *diagnostics.RunDirectory) error { return d.WriteChunkingSummary(&summary) })
+}
+
+func (s *processState) writeExecutionEvidence() {
+	if s.debugDir == nil || s.runOutput == nil {
+		return
+	}
+	output := s.runOutput
+	s.debugWrite("model execution", func(d *diagnostics.RunDirectory) error {
+		return d.WriteJSONArtifact(diagnostics.ArtifactLLMExecution, processreport.BuildLLMExecution(output.Observations))
+	})
+	s.debugWrite("utilization", func(d *diagnostics.RunDirectory) error {
+		return d.WriteJSONArtifact(diagnostics.ArtifactUtilizationSummary, processreport.BuildUtilization(output, s.cfg.Backends))
+	})
+	s.debugWrite("correction ledger", func(d *diagnostics.RunDirectory) error {
+		return d.WriteJSONArtifact(diagnostics.ArtifactCorrectionLedger, processreport.BuildCorrectionLedger(processreport.CorrectionLedgerInput{
+			RunDirectoryPath: d.Path(), RunOutput: output}))
+	})
+}
+
+func fatalReason(ctx context.Context, err error) outcomes.ReasonCode {
+	if ctx.Err() != nil || errors.Is(err, context.Canceled) {
+		return outcomes.ReasonCanceled
+	}
+	var failure *structuredoutput.Failure
+	if errors.As(err, &failure) && failure.Reason != "" {
+		return failure.Reason
+	}
+	return outcomes.ReasonInternalFailure
+}
+
+func (s *processState) fail(phase string, reason outcomes.ReasonCode, cause error) int {
+	s.debugWrite("error log", func(d *diagnostics.RunDirectory) error {
+		return d.WriteErrorLog(phase + ": " + llm.RedactError(cause, s.cfg))
+	})
+	failures := []processreport.RunFailure{{Phase: phase, Reason: reason}}
+	completed := time.Now().UTC()
+	base, err := s.bundle(s.snapshot(reporting.StatusFailed, &completed, failures))
+	if err != nil {
+		fmt.Fprintf(s.stderr, "audita process: %s\n", llm.RedactError(err, s.cfg))
+	}
+	writeFailures := s.session.Finalize(nil, base, func(extra []diagnostics.FileFailure) diagnostics.ReportBundle {
+		all := append(append([]processreport.RunFailure(nil), failures...), mapWriteFailures(extra)...)
+		bundle, renderErr := s.bundle(s.snapshot(reporting.StatusFailed, &completed, all))
+		if renderErr != nil {
+			return base
+		}
+		return bundle
+	})
+	fmt.Fprintf(s.stderr, "audita process: %s\n", llm.RedactError(cause, s.cfg))
+	s.printFailures(writeFailures)
+	s.printDebugWarnings()
+	if s.debugDir != nil {
+		fmt.Fprintf(s.stderr, "audita process: diagnostics: %s\n", s.safePath(s.debugDir.Path()))
+	}
+	return 1
+}
+
+func (s *processState) finish(outputBytes []byte) int {
+	completed := time.Now().UTC()
+	snapshot := s.snapshot(reporting.StatusCompleted, &completed, nil)
+	complete, err := s.bundle(snapshot)
+	if err != nil {
+		return s.fail("report_write", outcomes.ReasonInternalFailure, err)
+	}
+	if s.session.TranscriptPath() == "" {
+		if _, err := s.stdout.Write(outputBytes); err != nil {
+			return s.fail("transcript_write", outcomes.ReasonInternalFailure, err)
+		}
+		outputBytes = nil
+	}
+	writeFailures := s.session.Finalize(outputBytes, complete, func(extra []diagnostics.FileFailure) diagnostics.ReportBundle {
+		failed, renderErr := s.bundle(s.snapshot(reporting.StatusFailed, &completed, mapWriteFailures(extra)))
+		if renderErr != nil {
+			return diagnostics.ReportBundle{}
+		}
+		return failed
+	})
+	if len(writeFailures) > 0 {
+		s.printFailures(writeFailures)
+		s.printDebugWarnings()
+		if s.debugDir != nil {
+			fmt.Fprintf(s.stderr, "audita process: diagnostics: %s\n", s.safePath(s.debugDir.Path()))
+		}
+		return 1
+	}
+	fmt.Fprintln(s.stderr, reporting.CompletionLine(snapshot))
+	for _, path := range s.session.ReportPaths() {
+		fmt.Fprintf(s.stderr, "Report: %s\n", s.safePath(path))
+	}
+	if s.debugDir != nil {
+		fmt.Fprintf(s.stderr, "Debug: %s\n", s.safePath(s.debugDir.Path()))
+	}
+	s.printDebugWarnings()
+	return 0
+}
+
+func mapWriteFailures(failures []diagnostics.FileFailure) []processreport.RunFailure {
+	out := make([]processreport.RunFailure, len(failures))
+	for i, failure := range failures {
+		phase := "report_write"
+		if failure.Kind == "transcript write" {
+			phase = "transcript_write"
+		}
+		out[i] = processreport.RunFailure{Phase: phase, Reason: outcomes.ReasonInternalFailure, OutputPath: failure.Path}
+	}
+	return out
+}
+
+func (s *processState) printFailures(failures []diagnostics.FileFailure) {
+	for _, failure := range failures {
+		fmt.Fprintf(s.stderr, "audita process: %s\n", llm.RedactError(failure, s.cfg))
+	}
+}
+
+func (s *processState) printDebugWarnings() {
+	s.debugMu.Lock()
+	defer s.debugMu.Unlock()
+	for _, warning := range s.debugWarnings {
+		fmt.Fprintf(s.stderr, "audita process: %s\n", warning)
+	}
+}
+
+// safePath sanitizes presentation only; filesystem destinations remain literal.
+func (s *processState) safePath(path string) string {
+	return redaction.New(llm.ConfiguredSecrets(&s.cfg)).String(path)
+}
+
+func (s *processState) reportDiagnosticError(err error) {
+	if err == nil {
+		return
+	}
+	message := "debug " + llm.RedactError(err, s.cfg)
+	s.debugMu.Lock()
+	defer s.debugMu.Unlock()
+	s.debugWarnings = append(s.debugWarnings, message)
+}

internal/cli/release_fixtures_test.go

@@ -1,469 +0,0 @@
-package cli
-
-import (
-	"bytes"
-	"encoding/json"
-	"os"
-	"path/filepath"
-	"reflect"
-	"strings"
-	"testing"
-
-	"gitea.maximumdirect.net/eric/audita/internal/core/reporting"
-	"gitea.maximumdirect.net/eric/audita/internal/core/schema"
-	"gitea.maximumdirect.net/eric/audita/internal/framework/proposal_generation"
-	"gitea.maximumdirect.net/eric/audita/internal/framework/validators"
-)
-
-type releaseFixtureExpectations struct {
-	MustApplyTexts         []string `json:"must_apply_texts"`
-	MustNotApplyTexts      []string `json:"must_not_apply_texts"`
-	ProtectedTerms         []string `json:"protected_terms"`
-	ExpectedModuleInstance []string `json:"expected_module_instances"`
-	MinimumCounts          struct {
-		Applied  int `json:"applied"`
-		Rejected int `json:"rejected"`
-		Skipped  int `json:"skipped"`
-	} `json:"minimum_counts"`
-}
-
-func TestReleaseFixtureDefaultPipelineReadiness(t *testing.T) {
-	base := fixturePath(filepath.Join("release", "default-release"))
-
-	var expectations releaseFixtureExpectations
-	if err := json.Unmarshal(readFile(t, base+".expectations.json"), &expectations); err != nil {
-		t.Fatalf("unmarshal release expectations: %v", err)
-	}
-
-	proposalResponses := readProposalResponses(t, base+".proposals.json")
-	validationResponses := readValidationResponses(t, base+".validations.json")
-
-	// First pass: default full pipeline with deterministic fake LLM responses.
-	first := runReleaseFixturePass(t, releaseRunConfig{
-		transcriptPath:         base + ".transcript.json",
-		glossaryPath:           base + ".glossary.yaml",
-		outputSchema:           "bare-segments",
-		proposalResponses:      proposalResponses,
-		validationResponses:    validationResponses,
-		expectedProposalCalls:  []string{"glossary_1:proposal", "homophones:proposal", "glossary_2:proposal", "spoken_word:proposal", "grammar:proposal"},
-		reportSchemaName:       reporting.DefaultProcessReportSchemaName,
-		reportSchemaVersion:    reporting.DefaultProcessReportSchemaVersion,
-		expectedOutputSchema:   "bare-segments",
-		expectModuleInstances:  expectations.ExpectedModuleInstance,
-		expectUtilizationPaths: true,
-	})
-
-	gotTranscript := mustReadTranscript(t, first.outputPath)
-	expectFinalTranscriptContains(t, gotTranscript, expectations.MustApplyTexts)
-	expectFinalTranscriptDoesNotContain(t, gotTranscript, expectations.MustNotApplyTexts)
-	expectFinalTranscriptContains(t, gotTranscript, expectations.ProtectedTerms)
-
-	assertReleaseCounts(t, first.report, expectations)
-	assertPromptAndSchemaMetadataPresent(t, first.runDir)
-	assertReleaseLedgerShape(t, first.report)
-	assertReleaseUtilizationShape(t, first.report)
-	assertStableValidatorKeysPresent(t, first.report)
-	assertStdoutStderrContract(t, first.stdout, first.stderr)
-	assertNoSecretMarkersInTree(t, first.runDir, []string{"release-secret"})
-	assertNoSecretMarkers(t, first.reportPath, []string{"release-secret"})
-
-	// Output schema check: audita-v1 object payload.
-	auditaV1 := runReleaseFixturePass(t, releaseRunConfig{
-		transcriptPath:         base + ".transcript.json",
-		glossaryPath:           base + ".glossary.yaml",
-		outputSchema:           "audita-v1",
-		proposalResponses:      proposalResponses,
-		validationResponses:    validationResponses,
-		expectedProposalCalls:  []string{"glossary_1:proposal", "homophones:proposal", "glossary_2:proposal", "spoken_word:proposal", "grammar:proposal"},
-		reportSchemaName:       reporting.DefaultProcessReportSchemaName,
-		reportSchemaVersion:    reporting.DefaultProcessReportSchemaVersion,
-		expectedOutputSchema:   "audita-v1",
-		expectModuleInstances:  expectations.ExpectedModuleInstance,
-		expectUtilizationPaths: true,
-	})
-	assertAuditaV1OutputShape(t, auditaV1.outputPath)
-
-	// Idempotence-oriented second pass:
-	// run again on first output with deterministic no-op responses.
-	noOpProposals := make([]proposal_generation.StructuredCorrectionSet, 5)
-	for i := range noOpProposals {
-		noOpProposals[i] = proposal_generation.StructuredCorrectionSet{Corrections: nil}
-	}
-	second := runReleaseFixturePass(t, releaseRunConfig{
-		transcriptPath:         first.outputPath,
-		glossaryPath:           base + ".glossary.yaml",
-		outputSchema:           "bare-segments",
-		proposalResponses:      noOpProposals,
-		validationResponses:    nil,
-		expectedProposalCalls:  []string{"glossary_1:proposal", "homophones:proposal", "glossary_2:proposal", "spoken_word:proposal", "grammar:proposal"},
-		reportSchemaName:       reporting.DefaultProcessReportSchemaName,
-		reportSchemaVersion:    reporting.DefaultProcessReportSchemaVersion,
-		expectedOutputSchema:   "bare-segments",
-		expectModuleInstances:  expectations.ExpectedModuleInstance,
-		expectUtilizationPaths: true,
-	})
-	firstSegments := mustReadTranscript(t, first.outputPath)
-	secondSegments := mustReadTranscript(t, second.outputPath)
-	if !reflect.DeepEqual(firstSegments, secondSegments) {
-		t.Fatalf("expected idempotent second pass transcript; first=%+v second=%+v", firstSegments, secondSegments)
-	}
-	if second.report.ModulesSummary == nil {
-		t.Fatalf("expected modules summary on second pass")
-	}
-	if second.report.ModulesSummary.TotalAppliedChanges != 0 {
-		t.Fatalf("expected no-op second pass (0 applied), got %+v", second.report.ModulesSummary)
-	}
-}
-
-type releaseRunConfig struct {
-	transcriptPath         string
-	glossaryPath           string
-	outputSchema           string
-	proposalResponses      []proposal_generation.StructuredCorrectionSet
-	validationResponses    []validators.LLMValidationResponse
-	expectedProposalCalls  []string
-	reportSchemaName       string
-	reportSchemaVersion    string
-	expectedOutputSchema   string
-	expectModuleInstances  []string
-	expectUtilizationPaths bool
-}
-
-type releaseRunResult struct {
-	stdout     string
-	stderr     string
-	outputPath string
-	reportPath string
-	report     reporting.ProcessReport
-	runDir     string
-}
-
-func runReleaseFixturePass(t *testing.T, cfg releaseRunConfig) releaseRunResult {
-	t.Helper()
-
-	processProposalLLMClient = &fakeStructuredLLMClient{proposalResponses: append([]proposal_generation.StructuredCorrectionSet(nil), cfg.proposalResponses...)}
-	processValidationLLMClient = &fakeStructuredLLMClient{validationResponses: append([]validators.LLMValidationResponse(nil), cfg.validationResponses...)}
-	t.Cleanup(func() {
-		processProposalLLMClient = nil
-		processValidationLLMClient = nil
-	})
-
-	workDir := t.TempDir()
-	reportPath := filepath.Join(t.TempDir(), "report.json")
-	outputPath := filepath.Join(t.TempDir(), "out.json")
-	configPath := writeFile(t, "release-config.yml", "version: 1\n")
-
-	args := []string{
-		"process",
-		cfg.transcriptPath,
-		"--glossary",
-		cfg.glossaryPath,
-		"--config",
-		configPath,
-		"--output",
-		outputPath,
-		"--output-schema",
-		cfg.outputSchema,
-		"--report-json",
-		reportPath,
-		"--work-dir",
-		workDir,
-		"--work-dir-retention",
-		"always",
-	}
-
-	var stdout, stderr bytes.Buffer
-	exitCode := Run(args, &stdout, &stderr)
-	if exitCode != 0 {
-		t.Fatalf("expected success, got %d stderr=%q", exitCode, stderr.String())
-	}
-
-	report := readProcessReport(t, reportPath)
-	if report.ReportMetadata.ReportSchemaName != cfg.reportSchemaName || report.ReportMetadata.ReportSchemaVersion != cfg.reportSchemaVersion {
-		t.Fatalf("unexpected report schema metadata: %+v", report.ReportMetadata)
-	}
-	if report.ReportMetadata.OutputSchema != cfg.expectedOutputSchema {
-		t.Fatalf("unexpected output schema metadata: got %q want %q", report.ReportMetadata.OutputSchema, cfg.expectedOutputSchema)
-	}
-	if len(cfg.expectModuleInstances) > 0 {
-		got := make([]string, 0, len(report.ModuleResults))
-		for _, mr := range report.ModuleResults {
-			got = append(got, mr.ModuleInstance)
-		}
-		if !reflect.DeepEqual(got, cfg.expectModuleInstances) {
-			t.Fatalf("unexpected module instances: got %v want %v", got, cfg.expectModuleInstances)
-		}
-	}
-	if report.Diagnostics == nil {
-		t.Fatalf("expected diagnostics metadata")
-	}
-	if cfg.expectUtilizationPaths {
-		if report.Diagnostics.UtilizationSummaryPath == "" || report.Diagnostics.CorrectionLedgerPath == "" {
-			t.Fatalf("expected utilization/ledger artifact paths in report diagnostics: %+v", report.Diagnostics)
-		}
-	}
-
-	runDir := onlyRunDir(t, workDir)
-	if _, err := os.Stat(filepath.Join(runDir, "report.json")); err != nil {
-		t.Fatalf("expected run-dir report: %v", err)
-	}
-
-	if c, ok := processProposalLLMClient.(*fakeStructuredLLMClient); ok {
-		if !reflect.DeepEqual(c.calls, cfg.expectedProposalCalls) {
-			t.Fatalf("unexpected proposal call order: got %v want %v", c.calls, cfg.expectedProposalCalls)
-		}
-	}
-
-	return releaseRunResult{
-		stdout:     stdout.String(),
-		stderr:     stderr.String(),
-		outputPath: outputPath,
-		reportPath: reportPath,
-		report:     report,
-		runDir:     runDir,
-	}
-}
-
-func readProposalResponses(t *testing.T, path string) []proposal_generation.StructuredCorrectionSet {
-	t.Helper()
-	var out []proposal_generation.StructuredCorrectionSet
-	if err := json.Unmarshal(readFile(t, path), &out); err != nil {
-		t.Fatalf("unmarshal proposal responses: %v", err)
-	}
-	return out
-}
-
-func readValidationResponses(t *testing.T, path string) []validators.LLMValidationResponse {
-	t.Helper()
-	var out []validators.LLMValidationResponse
-	if err := json.Unmarshal(readFile(t, path), &out); err != nil {
-		t.Fatalf("unmarshal validation responses: %v", err)
-	}
-	return out
-}
-
-func mustReadTranscript(t *testing.T, path string) []schema.Segment {
-	t.Helper()
-	transcript, err := schema.ParseTranscriptJSON(readFile(t, path))
-	if err != nil {
-		t.Fatalf("parse transcript output: %v", err)
-	}
-	return transcript.Segments
-}
-
-func expectFinalTranscriptContains(t *testing.T, segments []schema.Segment, needles []string) {
-	t.Helper()
-	joined := flattenTranscriptText(segments)
-	for _, needle := range needles {
-		if !strings.Contains(joined, needle) {
-			t.Fatalf("expected transcript to contain %q, got %q", needle, joined)
-		}
-	}
-}
-
-func expectFinalTranscriptDoesNotContain(t *testing.T, segments []schema.Segment, needles []string) {
-	t.Helper()
-	joined := flattenTranscriptText(segments)
-	for _, needle := range needles {
-		if strings.Contains(joined, needle) {
-			t.Fatalf("expected transcript to not contain %q, got %q", needle, joined)
-		}
-	}
-}
-
-func flattenTranscriptText(segments []schema.Segment) string {
-	parts := make([]string, 0, len(segments))
-	for _, s := range segments {
-		parts = append(parts, s.Text)
-	}
-	return strings.Join(parts, "\n")
-}
-
-func assertReleaseCounts(t *testing.T, report reporting.ProcessReport, exp releaseFixtureExpectations) {
-	t.Helper()
-	if report.ModulesSummary == nil {
-		t.Fatalf("expected modules_summary")
-	}
-	if report.ModulesSummary.TotalAppliedChanges < exp.MinimumCounts.Applied {
-		t.Fatalf("expected at least %d applied changes, got %+v", exp.MinimumCounts.Applied, report.ModulesSummary)
-	}
-	validatorRejected := 0
-	skipped := 0
-	for _, mr := range report.ModuleResults {
-		validatorRejected += len(mr.ValidatorRejected)
-		skipped += len(mr.SkippedChanges)
-	}
-	if validatorRejected < exp.MinimumCounts.Rejected {
-		t.Fatalf("expected at least %d validator rejections, got %d", exp.MinimumCounts.Rejected, validatorRejected)
-	}
-	if skipped < exp.MinimumCounts.Skipped {
-		t.Fatalf("expected at least %d application skips, got %d", exp.MinimumCounts.Skipped, skipped)
-	}
-}
-
-func assertReleaseUtilizationShape(t *testing.T, report reporting.ProcessReport) {
-	t.Helper()
-	var payload struct {
-		EffectiveConcurrency struct {
-			TotalLLM int `json:"total_llm"`
-		} `json:"effective_concurrency"`
-		RunTiming struct {
-			SchedulerQueueWaitMS      int64 `json:"scheduler_queue_wait_ms"`
-			LLMExecutionTimeMS        int64 `json:"llm_execution_time_ms"`
-			DeterministicValidationMS int64 `json:"deterministic_validation_time_ms"`
-		} `json:"run_timing"`
-		Modules    []map[string]any `json:"modules"`
-		Validators []map[string]any `json:"validators"`
-	}
-	if err := json.Unmarshal(readFile(t, report.Diagnostics.UtilizationSummaryPath), &payload); err != nil {
-		t.Fatalf("unmarshal utilization diagnostics: %v", err)
-	}
-	if payload.EffectiveConcurrency.TotalLLM <= 0 {
-		t.Fatalf("expected positive total llm concurrency, got %+v", payload.EffectiveConcurrency)
-	}
-	if payload.RunTiming.SchedulerQueueWaitMS < 0 || payload.RunTiming.LLMExecutionTimeMS < 0 || payload.RunTiming.DeterministicValidationMS < 0 {
-		t.Fatalf("expected non-negative run timing values, got %+v", payload.RunTiming)
-	}
-	if len(payload.Modules) == 0 {
-		t.Fatalf("expected module timing summaries")
-	}
-	if len(payload.Validators) == 0 {
-		t.Fatalf("expected validator timing summaries")
-	}
-}
-
-func assertReleaseLedgerShape(t *testing.T, report reporting.ProcessReport) {
-	t.Helper()
-	var entries []struct {
-		ModuleKey            string `json:"module_key"`
-		ModuleInstance       string `json:"module_instance"`
-		ProposalIndex        int    `json:"proposal_index"`
-		Disposition          string `json:"disposition"`
-		DispositionReason    string `json:"disposition_reason_code"`
-		OriginalText         string `json:"original_text"`
-		ProposedCorrected    string `json:"proposed_corrected_text"`
-		ReplacementPolicy    string `json:"replacement_policy"`
-		DeterministicResults []struct {
-			ValidatorKey string `json:"validator_key"`
-		} `json:"deterministic_validator_decisions"`
-		LLMResults []struct {
-			ValidatorKey string `json:"validator_key"`
-		} `json:"llm_validator_decisions"`
-	}
-	if err := json.Unmarshal(readFile(t, report.Diagnostics.CorrectionLedgerPath), &entries); err != nil {
-		t.Fatalf("unmarshal correction ledger: %v", err)
-	}
-	if len(entries) == 0 {
-		t.Fatalf("expected correction ledger entries")
-	}
-	hasApplied := false
-	hasRejected := false
-	hasSkipped := false
-	for _, entry := range entries {
-		if entry.ModuleInstance == "" || entry.ModuleKey == "" {
-			t.Fatalf("expected module identity in ledger entry: %+v", entry)
-		}
-		switch entry.Disposition {
-		case "applied":
-			hasApplied = true
-		case "rejected":
-			hasRejected = true
-		case "skipped":
-			hasSkipped = true
-		}
-	}
-	if !hasApplied || !hasRejected {
-		t.Fatalf("expected applied and rejected entries in correction ledger, got %+v", entries)
-	}
-	if !hasSkipped {
-		// Some deterministic fixture paths do not trigger apply-time skips;
-		// rejections are still captured separately from application skips.
-	}
-}
-
-func assertPromptAndSchemaMetadataPresent(t *testing.T, runDir string) {
-	t.Helper()
-	metadataPaths, err := filepath.Glob(filepath.Join(runDir, "*", "*request-metadata.json"))
-	if err != nil {
-		t.Fatalf("glob request metadata artifacts: %v", err)
-	}
-	if len(metadataPaths) == 0 {
-		t.Fatalf("expected request metadata artifacts with prompt metadata")
-	}
-
-	foundPromptMetadata := false
-	foundSchemaMetadata := false
-	for _, path := range metadataPaths {
-		var payload map[string]any
-		if err := json.Unmarshal(readFile(t, path), &payload); err != nil {
-			t.Fatalf("unmarshal request metadata artifact %q: %v", path, err)
-		}
-		if pm, ok := payload["prompt_metadata"].(map[string]any); ok {
-			if pm["prompt_id"] != nil && pm["prompt_version"] != nil && pm["sha256"] != nil {
-				foundPromptMetadata = true
-			}
-		}
-		if sm, ok := payload["response_schema"].(map[string]any); ok {
-			if sm["id"] != nil && sm["version"] != nil && sm["name"] != nil && sm["sha256"] != nil {
-				foundSchemaMetadata = true
-			}
-		}
-	}
-	if !foundPromptMetadata {
-		t.Fatalf("expected prompt metadata in request metadata artifacts")
-	}
-	if !foundSchemaMetadata {
-		t.Fatalf("expected structured response schema metadata in request metadata artifacts")
-	}
-}
-
-func assertStableValidatorKeysPresent(t *testing.T, report reporting.ProcessReport) {
-	t.Helper()
-	seen := map[string]bool{}
-	for _, module := range report.ModuleResults {
-		for _, decision := range module.ValidatorDecisions {
-			seen[decision.ValidatorName] = true
-		}
-		for _, rejected := range module.ValidatorRejected {
-			seen[rejected.ValidatorName] = true
-		}
-	}
-	expectedAny := []string{
-		"confidence_threshold",
-		"original_text_presence",
-		"no_effect",
-	}
-	for _, key := range expectedAny {
-		if !seen[key] {
-			t.Fatalf("expected stable validator key %q in report decisions/rejections; seen=%v", key, seen)
-		}
-	}
-}
-
-func assertStdoutStderrContract(t *testing.T, stdout, stderr string) {
-	t.Helper()
-	if stdout != "" {
-		t.Fatalf("expected empty stdout with --output, got %q", stdout)
-	}
-	if strings.Contains(stderr, `"module_results"`) || strings.Contains(stderr, `"report_metadata"`) {
-		t.Fatalf("stderr should remain human-readable, not report JSON: %q", stderr)
-	}
-}
-
-func assertAuditaV1OutputShape(t *testing.T, outputPath string) {
-	t.Helper()
-	var payload struct {
-		Schema   string           `json:"schema"`
-		Version  string           `json:"version"`
-		Segments []schema.Segment `json:"segments"`
-	}
-	if err := json.Unmarshal(readFile(t, outputPath), &payload); err != nil {
-		t.Fatalf("unmarshal audita-v1 output: %v", err)
-	}
-	if payload.Schema != "audita-v1" || payload.Version != "v1" {
-		t.Fatalf("unexpected audita-v1 metadata: %+v", payload)
-	}
-	if len(payload.Segments) == 0 {
-		t.Fatalf("expected non-empty audita-v1 segments")
-	}
-}

internal/cli/subprocess_test_hooks.go

@@ -1,113 +0,0 @@
-package cli
-
-import (
-	"context"
-	"errors"
-	"os"
-	"strconv"
-	"strings"
-	"sync"
-	"time"
-
-	"gitea.maximumdirect.net/eric/audita/internal/framework/contracts"
-	"gitea.maximumdirect.net/eric/audita/internal/framework/proposal_generation"
-	"gitea.maximumdirect.net/eric/audita/internal/framework/validators"
-)
-
-const (
-	subprocessTestLLMModeEnv      = "AUDITA_SUBPROCESS_TEST_LLM_MODE"
-	subprocessTestRunTimeoutMSEnv = "AUDITA_SUBPROCESS_TEST_RUN_TIMEOUT_MS"
-)
-
-// ConfigureSubprocessTestHooksFromEnv enables deterministic test-only hooks for
-// subprocess integration tests that run through the Go test binary helper path.
-func ConfigureSubprocessTestHooksFromEnv() {
-	mode := strings.TrimSpace(os.Getenv(subprocessTestLLMModeEnv))
-	timeoutMSRaw := strings.TrimSpace(os.Getenv(subprocessTestRunTimeoutMSEnv))
-	// Only activate in explicit subprocess test mode.
-	if mode == "" && timeoutMSRaw == "" {
-		return
-	}
-
-	if mode != "" {
-		client := &subprocessTestLLMClient{mode: mode}
-		processProposalLLMClient = client
-		processValidationLLMClient = client
-	}
-
-	if timeoutMSRaw == "" {
-		return
-	}
-	timeoutMS, err := strconv.Atoi(timeoutMSRaw)
-	if err != nil || timeoutMS <= 0 {
-		return
-	}
-	processRunnerContext = func() (context.Context, context.CancelFunc) {
-		return context.WithTimeout(context.Background(), time.Duration(timeoutMS)*time.Millisecond)
-	}
-}
-
-type subprocessTestLLMClient struct {
-	mode      string
-	mu        sync.Mutex
-	proposals int
-}
-
-func (c *subprocessTestLLMClient) CompleteStructured(ctx context.Context, req contracts.StructuredCompletionRequest, out any) (contracts.StructuredCompletionResponse, error) {
-	_ = req
-
-	switch c.mode {
-	case "backend_error":
-		return contracts.StructuredCompletionResponse{}, errors.New("synthetic backend failure")
-	case "block_until_cancel":
-		<-ctx.Done()
-		return contracts.StructuredCompletionResponse{}, ctx.Err()
-	case "malformed_structured":
-		switch target := out.(type) {
-		case *proposal_generation.StructuredCorrectionSet:
-			*target = proposal_generation.StructuredCorrectionSet{
-				Corrections: []proposal_generation.StructuredCorrectionProposal{
-					{TargetSegmentID: 0, OriginalText: "x", CorrectedText: "y", Confidence: 0.99},
-				},
-			}
-		case *validators.LLMValidationResponse:
-			*target = validators.LLMValidationResponse{
-				Validations: []validators.LLMValidationDecision{
-					{CorrectionIndex: 999, Approved: true, Confidence: 0.9, Reason: "bad index"},
-				},
-			}
-		}
-	case "mid_pipeline_fail":
-		switch target := out.(type) {
-		case *proposal_generation.StructuredCorrectionSet:
-			c.mu.Lock()
-			c.proposals++
-			proposalCall := c.proposals
-			c.mu.Unlock()
-
-			if proposalCall >= 3 {
-				*target = proposal_generation.StructuredCorrectionSet{
-					Corrections: []proposal_generation.StructuredCorrectionProposal{
-						{TargetSegmentID: 0, OriginalText: "x", CorrectedText: "y", Confidence: 0.99},
-					},
-				}
-			} else {
-				*target = proposal_generation.StructuredCorrectionSet{
-					Corrections: []proposal_generation.StructuredCorrectionProposal{
-						{TargetSegmentID: 1, OriginalText: "Segment", CorrectedText: "Segment", Confidence: 0.99},
-					},
-				}
-			}
-		case *validators.LLMValidationResponse:
-			*target = validators.LLMValidationResponse{Validations: nil}
-		}
-	default:
-		switch target := out.(type) {
-		case *proposal_generation.StructuredCorrectionSet:
-			*target = proposal_generation.StructuredCorrectionSet{Corrections: nil}
-		case *validators.LLMValidationResponse:
-			*target = validators.LLMValidationResponse{Validations: nil}
-		}
-	}
-	return contracts.StructuredCompletionResponse{}, nil
-}

internal/cli/testdata/parity/application-skip-ambiguous.case.json

@@ -1,22 +0,0 @@
-{
-  "name": "ambiguous_match_rejected_before_application",
-  "transcript_file": "application-skip-ambiguous.transcript.json",
-  "glossary_file": "default-handoff.glossary.yaml",
-  "modules_csv": "homophones",
-  "proposal_responses_file": "application-skip-ambiguous.proposals.json",
-  "validation_responses_file": "application-skip-ambiguous.validations.json",
-  "expect": {
-    "exit_code": 0,
-    "status": "success",
-    "output_transcript_file": "application-skip-ambiguous.expected-transcript.json",
-    "module_instances": ["homophones"],
-    "module_count": 1,
-    "total_applied_changes": 0,
-    "total_skipped_changes": 1,
-    "module_applied_counts": [0],
-    "module_rejected_counts": [1],
-    "module_skip_counts": [0],
-    "validator_rejected_reason_codes": ["ambiguous_original_text"],
-    "expected_proposal_calls": ["homophones:proposal"]
-  }
-}

internal/cli/testdata/parity/application-skip-ambiguous.expected-transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0,"end":1,"text":"the site near another site"}
-]

internal/cli/testdata/parity/application-skip-ambiguous.proposals.json

@@ -1,3 +0,0 @@
-[
-  {"corrections": [{"id": 1, "original_text": "site", "corrected_text": "sight", "confidence": 0.99}]}
-]

internal/cli/testdata/parity/application-skip-ambiguous.transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0.0,"end":1.0,"text":"the site near another site"}
-]

internal/cli/testdata/parity/application-skip-ambiguous.validations.json

@@ -1,4 +0,0 @@
-[
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]}
-]

internal/cli/testdata/parity/default-full-pipeline.case.json

@@ -1,28 +0,0 @@
-{
-  "name": "default_full_pipeline_shape_and_reports",
-  "transcript_file": "default-full-pipeline.transcript.json",
-  "glossary_file": "default-full-pipeline.glossary.yaml",
-  "proposal_responses_file": "default-full-pipeline.proposals.json",
-  "validation_responses_file": "default-full-pipeline.validations.json",
-  "env": {
-    "AUDITA_LLM_API_KEY": "parity-secret",
-    "AUDITA_VALIDATION_LLM_API_KEY": "parity-secret"
-  },
-  "expect": {
-    "exit_code": 0,
-    "status": "success",
-    "output_transcript_file": "default-full-pipeline.expected-transcript.json",
-    "module_instances": ["glossary_1", "homophones", "glossary_2", "spoken_word", "grammar"],
-    "module_count": 5,
-    "total_applied_changes": 3,
-    "total_skipped_changes": 3,
-    "secret_markers": ["parity-secret"],
-    "expected_proposal_calls": [
-      "glossary_1:proposal",
-      "homophones:proposal",
-      "glossary_2:proposal",
-      "spoken_word:proposal",
-      "grammar:proposal"
-    ]
-  }
-}

internal/cli/testdata/parity/default-full-pipeline.expected-transcript.json

@@ -1,9 +0,0 @@
-[
-  {
-    "id": 1,
-    "speaker": "Alice",
-    "start": 0,
-    "end": 1,
-    "text": "Hello, there were Jesters hmm"
-  }
-]

internal/cli/testdata/parity/default-full-pipeline.glossary.yaml

@@ -1,7 +0,0 @@
-glossary:
-  - name: Jesters
-    aliases:
-      - jester
-    plural: jesters
-    category: faction
-    summary: A protected in-world faction term.

internal/cli/testdata/parity/default-full-pipeline.proposals.json

@@ -1,28 +0,0 @@
-[
-  {
-    "corrections": [
-      {"id": 1, "original_text": "gestures", "corrected_text": "Jesters", "confidence": 0.99}
-    ]
-  },
-  {
-    "corrections": [
-      {"id": 1, "original_text": "Jesters", "corrected_text": "jesters", "confidence": 0.99},
-      {"id": 1, "original_text": "Jesters", "corrected_text": "JESTERX", "confidence": 0.99}
-    ]
-  },
-  {
-    "corrections": [
-      {"id": 1, "original_text": "jesters", "corrected_text": "JESTERS", "confidence": 0.99}
-    ]
-  },
-  {
-    "corrections": [
-      {"id": 1, "original_text": "uh", "corrected_text": "hmm", "confidence": 0.99}
-    ]
-  },
-  {
-    "corrections": [
-      {"id": 1, "original_text": "hello ,", "corrected_text": "Hello,", "confidence": 0.99}
-    ]
-  }
-]

internal/cli/testdata/parity/default-full-pipeline.transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0.0,"end":1.0,"text":"hello , there were gestures uh"}
-]

internal/cli/testdata/parity/default-full-pipeline.validations.json

@@ -1,12 +0,0 @@
-[
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": false, "confidence": 0.99, "reason": "reject cleanup"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "parity-secret"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]}
-]

internal/cli/testdata/parity/default-handoff.case.json

@@ -1,27 +0,0 @@
-{
-  "name": "default_pipeline_handoff_and_module_order",
-  "transcript_file": "default-handoff.transcript.json",
-  "glossary_file": "default-handoff.glossary.yaml",
-  "proposal_responses_file": "default-handoff.proposals.json",
-  "validation_responses_file": "default-handoff.validations.json",
-  "expect": {
-    "exit_code": 0,
-    "status": "success",
-    "output_transcript_file": "default-handoff.expected-transcript.json",
-    "module_instances": ["glossary_1", "homophones", "glossary_2", "spoken_word", "grammar"],
-    "module_count": 5,
-    "total_applied_changes": 5,
-    "total_skipped_changes": 0,
-    "module_applied_counts": [1, 1, 1, 1, 1],
-    "module_rejected_counts": [0, 0, 0, 0, 0],
-    "module_skip_counts": [0, 0, 0, 0, 0],
-    "expected_proposal_calls": [
-      "glossary_1:proposal",
-      "homophones:proposal",
-      "glossary_2:proposal",
-      "spoken_word:proposal",
-      "grammar:proposal"
-    ],
-    "min_response_payload_artifacts": 15
-  }
-}

internal/cli/testdata/parity/default-handoff.expected-transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0,"end":1,"text":"Hello, there were Jesters at the Sight um"}
-]

internal/cli/testdata/parity/default-handoff.glossary.yaml

@@ -1,6 +0,0 @@
-glossary:
-  - name: Jesters
-    aliases: [jester]
-    plural: jesters
-    category: faction
-    summary: A protected in-world faction term.

internal/cli/testdata/parity/default-handoff.proposals.json

@@ -1,7 +0,0 @@
-[
-  {"corrections": [{"id": 1, "original_text": "gestures", "corrected_text": "Jesters", "confidence": 0.99}]},
-  {"corrections": [{"id": 1, "original_text": "site", "corrected_text": "sight", "confidence": 0.99}]},
-  {"corrections": [{"id": 1, "original_text": "sight", "corrected_text": "Sight", "confidence": 0.99}]},
-  {"corrections": [{"id": 1, "original_text": "um um", "corrected_text": "um", "confidence": 0.99}]},
-  {"corrections": [{"id": 1, "original_text": "hello ,", "corrected_text": "Hello,", "confidence": 0.99}]}
-]

internal/cli/testdata/parity/default-handoff.transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0.0,"end":1.0,"text":"hello , there were gestures at the site um um"}
-]

internal/cli/testdata/parity/default-handoff.validations.json

@@ -1,12 +0,0 @@
-[
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]},
-  {"validations": [{"correction_index": 0, "approved": true, "confidence": 0.99, "reason": "ok"}]}
-]

internal/cli/testdata/parity/deterministic-validator-low-confidence.case.json

@@ -1,17 +0,0 @@
-{
-  "name": "deterministic_validator_low_confidence",
-  "transcript_file": "deterministic-validator-low-confidence.transcript.json",
-  "glossary_file": "default-full-pipeline.glossary.yaml",
-  "modules_csv": "grammar",
-  "proposal_responses_file": "deterministic-validator-low-confidence.proposals.json",
-  "expect": {
-    "exit_code": 0,
-    "status": "success",
-    "output_transcript_file": "deterministic-validator-low-confidence.expected-transcript.json",
-    "module_instances": ["grammar"],
-    "module_count": 1,
-    "total_skipped_changes": 1,
-    "validator_rejected_reason_codes": ["low_confidence"],
-    "expected_proposal_calls": ["grammar:proposal"]
-  }
-}

internal/cli/testdata/parity/deterministic-validator-low-confidence.expected-transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0,"end":1,"text":"hello , world"}
-]

internal/cli/testdata/parity/deterministic-validator-low-confidence.proposals.json

@@ -1,7 +0,0 @@
-[
-  {
-    "corrections": [
-      {"id": 1, "original_text": "hello ,", "corrected_text": "Hello,", "confidence": 0.1}
-    ]
-  }
-]

internal/cli/testdata/parity/deterministic-validator-low-confidence.transcript.json

@@ -1,3 +0,0 @@
-[
-  {"id":1,"speaker":"Alice","start":0.0,"end":1.0,"text":"hello , world"}
-]

internal/cli/testdata/parity/glossary-schema-error.case.json

@@ -1,12 +0,0 @@
-{
-  "name": "glossary_schema_handling",
-  "transcript_file": "default-full-pipeline.transcript.json",
-  "glossary_file": "glossary-schema-error.glossary.yaml",
-  "expect": {
-    "exit_code": 1,
-    "status": "failed",
-    "error_phase": "glossary_schema",
-    "stderr_contains": "glossary_schema",
-    "require_error_log": true
-  }
-}

internal/cli/testdata/parity/glossary-schema-error.glossary.yaml

@@ -1,2 +0,0 @@
-glossary:
-  - name: MissingCategoryAndSummary