Upgrade PromptKit and prepare version 0.7.1

.woodpecker/release.yml

@@ -0,0 +1,33 @@
+when:
+  - event: tag
+
+steps:
+  - name: validate-release
+    image: golang:1.25.5
+    commands:
+      - |
+        set -eu
+
+        version="$CI_COMMIT_TAG"
+        release_note="docs/releases/$version.md"
+
+        if ! printf '%s\n' "$version" | grep -E -x 'v(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)' >/dev/null; then
+          printf '%s\n' "invalid release tag: $version" >&2
+          exit 1
+        fi
+        if [ ! -s "$release_note" ]; then
+          printf '%s\n' "missing release note: $release_note" >&2
+          exit 1
+        fi
+        if ! grep -F -x "# Notarius $version" "$release_note" >/dev/null; then
+          printf '%s\n' "release note heading does not match $version" >&2
+          exit 1
+        fi
+        for heading in '## Summary' '## Compatibility' '## Upgrade' '## Changes'; do
+          if ! grep -F -x "$heading" "$release_note" >/dev/null; then
+            printf '%s\n' "release note is missing heading: $heading" >&2
+            exit 1
+          fi
+        done
+
+        ./scripts/check-release-source.sh "$version"

AGENTS.md

@@ -1,2 +1,7 @@
 Please review `docs/development.md` for initial orientation in this repository
 and follow its task-specific reading guide.
+
+For every planning and implementation task, follow the
+[`Implementation Discipline`](docs/policy/architecture.md#implementation-discipline):
+keep machinery proportionate, do not invent restrictions, and do not elevate
+incidental implementation details into repository policy.

README.md

@@ -28,6 +28,20 @@ For the complete ordered D&D workflow, use
 [its synthetic transcript](examples/dnd-complete-transcript.json). It
 demonstrates all implemented D&D lanes and the supporting campaign references.
 
+## Install A Source Release
+
+Install a pinned source release with Go:
+
+~~~
+GOWORK=off go install \
+  gitea.maximumdirect.net/eric/notarius/cmd/notarius@<tag>
+~~~
+
+Replace `<tag>` with a stable release tag such as `vMAJOR.MINOR.PATCH`. The
+installed command's diagnostic version is described in the [CLI
+reference](docs/cli.md); maintainers preparing a release should follow [Source
+Releases](docs/release.md).
+
 ## Documentation
 
 - [CLI reference](docs/cli.md) — commands, flags, output streams, and exits.
@@ -39,6 +53,8 @@ demonstrates all implemented D&D lanes and the supporting campaign references.
   artifact formats.
 - [Subprocess consumer guide](docs/consumers/subprocess.md) — invoke Notarius
   from an orchestrator and consume a published result.
+- [Complete D&D consumer guide](docs/consumers/dnd-pipeline.md) — run the full
+  D&D pipeline as a subprocess and discover its structured artifacts.
 - [Internal overview](docs/internal/overview.md) — implemented component map
   for maintainers.
 - [Developer guide](docs/development.md) — contributor orientation and

assets/dnd/combat-turns/prompts/prompt.yaml

@@ -42,4 +42,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_combat_turns_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/combat-turns/schemas/dnd_combat_turns_llm.v1.json

@@ -16,7 +16,8 @@
             "type": "string"
           },
           "turn_kind": {
-            "type": "string"
+            "type": "string",
+            "enum": ["turn", "reaction", "legendary_action", "lair_action", "other"]
           },
           "source_refs": {
             "type": "array",

assets/dnd/enemy-events/prompts/prompt.yaml

@@ -50,4 +50,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_enemy_events_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/enemy-events/schemas/dnd_enemy_events_llm.v1.json

@@ -13,7 +13,10 @@
         "required": ["name", "kind", "source_refs"],
         "properties": {
           "name": {"type": "string"},
-          "kind": {"type": "string"},
+          "kind": {
+            "type": "string",
+            "enum": ["engaged", "killed", "fled", "captured", "incapacitated"]
+          },
           "source_refs": {
             "type": "array",
             "items": {

assets/dnd/item-occurrences/prompts/instructions.md

@@ -18,12 +18,20 @@ when the transcript explicitly describes it being physically destroyed or
 expended as a non-payment component. Use `transferred` only when possession
 moves between two distinct named party members.
 
-Return both `from` and `to` for every occurrence, using `null` when a holder does not
-apply. For `discovered`, set both holders to `null`. For `acquired`, set `from`
-to `null` and provide `to`; for `lost` and `consumed`, provide `from` and set
-`to` to `null`; and for `transferred`, provide both holders. Use `party` only
-for collective or unresolved party possession, never for either side of a
-transfer. Do not emit a transfer for a gift, sale, or payment outside the party.
+Return both `from` and `to` for every occurrence. Use JSON `null`, not an empty
+string, whenever a holder does not apply. Follow this holder matrix exactly:
+
+| `kind` | required `from` | required `to` |
+| --- | --- | --- |
+| `discovered` | `null` | `null` |
+| `acquired` | `null` | `party` or the named party member gaining possession |
+| `lost` | `party` or the named party member losing possession | `null` |
+| `consumed` | `party` or the named party member consuming the item | `null` |
+| `transferred` | one named party member | a different named party member |
+
+Use `party` only for collective or unresolved party possession, never for
+either side of a transfer. Do not emit a transfer for a gift, sale, or payment
+outside the party.
 
 Ordinary non-depleting use is not an occurrence. Do not infer acquisition from a
 discovery, or discovery from an acquisition: emit both only when each is

assets/dnd/item-occurrences/prompts/prompt.yaml

@@ -42,4 +42,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_item_occurrences_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/item-occurrences/schemas/dnd_item_occurrences_llm.v1.json

@@ -13,7 +13,10 @@
         "required": ["name", "kind", "quantity", "from", "to", "source_refs"],
         "properties": {
           "name": {"type": "string"},
-          "kind": {"type": "string"},
+          "kind": {
+            "type": "string",
+            "enum": ["discovered", "acquired", "lost", "consumed", "transferred"]
+          },
           "quantity": {"type": ["integer", "null"]},
           "from": {"type": ["string", "null"]},
           "to": {"type": ["string", "null"]},

assets/dnd/item-registry/extract/prompts/prompt.yaml

@@ -37,4 +37,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_item_registry_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/item-registry/normalize/prompts/instructions.md

@@ -5,5 +5,5 @@ evidence, similar objects, or a shared owner as sufficient.
 Keep currency denominations and materially different item types separate. Keep
 uncertain aliases separate. Do not infer an item property or uniqueness.
 
-When selecting a canonical display name, choose one supplied candidate name
-that is the clearest established designation.
+Set `canonical_candidate_number` to the supplied candidate number whose label
+is the clearest established designation.

assets/dnd/item-registry/normalize/prompts/prompt.yaml

@@ -27,4 +27,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: semantic_reconciliation_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/location-occurrences/prompts/prompt.yaml

@@ -42,4 +42,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_location_occurrences_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/location-registry/extract/prompts/prompt.yaml

@@ -37,4 +37,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_location_registry_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/location-registry/normalize/prompts/instructions.md

@@ -5,4 +5,5 @@ nearby evidence, nested places, or generic labels as sufficient.
 Keep parent and child places separate, as well as similarly named places and
 uncertain aliases.
 
-When selecting a canonical display name, prefer the clearest established name.
+Set `canonical_candidate_number` to the supplied candidate number whose label
+is the clearest established name.

assets/dnd/location-registry/normalize/prompts/prompt.yaml

@@ -27,4 +27,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: semantic_reconciliation_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/npc-occurrences/prompts/prompt.yaml

@@ -42,4 +42,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_npc_occurrences_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/npc-occurrences/schemas/dnd_npc_occurrences_llm.v1.json

@@ -16,7 +16,8 @@
             "type": "string"
           },
           "kind": {
-            "type": "string"
+            "type": "string",
+            "enum": ["mentioned", "noncombat_presence", "dialogue", "combat_ally", "combat_opponent", "other"]
           },
           "source_refs": {
             "type": "array",

assets/dnd/npc-registry/extract/prompts/prompt.yaml

@@ -37,4 +37,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_npc_registry_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/npc-registry/normalize/prompts/instructions.md

@@ -3,7 +3,8 @@ contextual labels and cited transcript windows. Preserve distinct individuals
 even when their names are similar or their contextual descriptions are
 identical.
 
-When selecting a canonical display name, prefer a complete, stable proper name
+Set `canonical_candidate_number` to the supplied candidate number whose label
+is the preferred canonical display name. Prefer a complete, stable proper name
 over an abbreviation. Prefer an unadorned proper name over that name plus a
 contextual class, role, title, or relationship descriptor unless the transcript
 establishes the descriptor as part of the person's name. A longer display name

assets/dnd/npc-registry/normalize/prompts/prompt.yaml

@@ -27,4 +27,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: semantic_reconciliation_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/scene-descriptions/prompts/instructions.md

@@ -5,9 +5,7 @@ into multiple scenes or use facts that are not supported by it.
 Return one kind, one concise title, and one concise summary. Choose exactly one
 kind:
 
-- combat: active combat materially organizes the scene, including
-  initiative-like exchanges or sustained hostile action. Planning a fight or
-  discussing a completed fight is not combat by itself.
+- combat: a scene classified as combat under the shared combat policy.
 - narrative: current-session in-world play that is not principally active
   combat, a prior-session recap, or sustained out-of-character session
   discussion. This includes exploration, travel, dialogue, investigation,
@@ -20,15 +18,13 @@ kind:
   play.
 
 Narrative is the default for actual current-session gameplay that does not meet
-another definition. When the accepted chunk is mixed:
+another definition. When the accepted chunk has no substantive active combat:
 
-1. use combat when active combat is a substantive central activity, even with
-   brief setup, rules clarification, or immediate aftermath;
-2. otherwise use recap when recounting a previous session is the chunk's
+1. use recap when recounting a previous session is the chunk's
    primary table purpose;
-3. otherwise use meta when sustained out-of-character session discussion is
+2. otherwise use meta when sustained out-of-character session discussion is
    primary and in-world progression is no more than incidental; and
-4. use narrative for all remaining current-session in-world play.
+3. use narrative for all remaining current-session in-world play.
 
 Brief table talk, dice resolution, rules clarification, jokes, or
 administrative comments do not make a gameplay scene meta. A short recollection

assets/dnd/scene-descriptions/prompts/prompt.yaml

@@ -27,6 +27,8 @@ messages:
     content_file: ./sharedassets/common-dnd-transcript-chunk.md
     cache_control:
       type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-scene-combat-policy.md
   - role: user
     content_file: ./instructions.md
     cache_control:
@@ -35,4 +37,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_scene_descriptions_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/scene-descriptions/schemas/dnd_scene_descriptions_llm.v1.json

@@ -6,7 +6,8 @@
   "required": ["kind", "title", "summary"],
   "properties": {
     "kind": {
-      "type": "string"
+      "type": "string",
+      "enum": ["combat", "narrative", "recap", "meta"]
     },
     "title": {
       "type": "string"

assets/dnd/scene-descriptions/validate/combat-semantics/prompts/instructions.md

@@ -0,0 +1,8 @@
+Classify only whether substantive active combat occurs in the supplied
+transcript chunk under the shared combat policy. Do not judge the scene title,
+summary, non-combat subtype, scene boundary, or any other aspect of a scene
+description.
+
+Return `combat` when the chunk contains substantive active combat and
+`non_combat` otherwise. Give a concise, transcript-grounded explanation for
+the classification.

assets/dnd/scene-descriptions/validate/combat-semantics/prompts/prompt.yaml

@@ -0,0 +1,23 @@
+id: dnd.scene_descriptions.validate_combat
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-scene-combat-policy.md
+  - role: user
+    content_file: ./instructions.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_scene_combat_semantics_llm.v1.json
+  repair_attempts: 1

assets/dnd/scene-descriptions/validate/combat-semantics/schemas/dnd_scene_combat_semantics_llm.v1.json

@@ -0,0 +1,18 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.scene_descriptions.combat_semantics.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["classification", "explanation"],
+  "properties": {
+    "classification": {
+      "type": "string",
+      "enum": ["combat", "non_combat"]
+    },
+    "explanation": {
+      "type": "string",
+      "minLength": 1,
+      "maxLength": 512
+    }
+  }
+}

assets/dnd/scenes/prompts/prompt.yaml

@@ -31,4 +31,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_scenes_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/dnd/shared/prompts/common-dnd-extraction-evidence.md

@@ -1,6 +1,8 @@
 Transcript units are the only evidence for extracted events and factual claims.
 Every reported factual claim must be supported by cited transcript units. Use
 integer `start_unit_id` and `end_unit_id` values from the transcript.
+Within each range, `start_unit_id` must identify the earlier transcript unit and
+`end_unit_id` the same or a later unit according to transcript order.
 
 When supporting evidence is non-contiguous, use multiple narrow ranges rather
 than a broad range that bridges unrelated conversation.

assets/dnd/shared/prompts/common-dnd-scene-combat-policy.md

@@ -0,0 +1,13 @@
+Use `combat` only when substantive active combat materially organizes the
+scene. Active combat includes initiative or turn exchanges, attacks, combat
+spells, damage, saves, movement, or similarly sustained hostile action.
+
+Do not use `combat` for planning or preparing for a possible fight; threats,
+hostile dialogue, or a tense confrontation; immediate aftermath, looting,
+healing, or discussion of a completed fight; a recap or in-world recollection
+of earlier combat; or out-of-character rules discussion without active
+encounter play.
+
+When a chunk contains substantive active combat alongside brief setup, rules
+clarification, interruption, phase transition, or immediate aftermath, classify
+it as `combat`.

assets/dnd/spells/prompts/prompt.yaml

@@ -47,4 +47,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: dnd_spells_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/generic/normalize/deduplication/prompts/candidates.md

@@ -1,3 +1,3 @@
-Candidate material:
+Candidate material, including the exact valid candidate-number range:
 
 {{ input "candidates" }}

assets/generic/normalize/deduplication/prompts/prompt.yaml

@@ -24,4 +24,4 @@ output:
   format: json
   validation_mode: json_schema
   schema_path: semantic_reconciliation_llm.v1.json
-  repair_attempts: 0
+  repair_attempts: 1

assets/generic/normalize/deduplication/prompts/protocol.md

@@ -1,7 +1,9 @@
-Use only the positive integer `candidate_id` values supplied in the candidate material.
+Use only the positive integer `candidate_number` values supplied in the candidate material. Valid candidate numbers are exactly the inclusive `first` through `last` values declared in `candidate_number_range`; use the explicit number attached to each candidate.
 
-Return a duplicate group only when the evidence supports that every selected candidate describes the same underlying entity. Each group must contain at least two distinct candidate IDs, and its `canonical_candidate_id` must be one of those IDs. A candidate may appear in at most one group.
+Transcript unit `id` values and evidence `start_unit_id` and `end_unit_id` values identify source positions. They are not candidate numbers and must never be used in `candidate_numbers` or `canonical_candidate_number`.
 
-Omit uncertain matches and candidates that should remain distinct. Do not invent candidates or infer an ID from list position. An empty `duplicate_groups` array is valid.
+Return a duplicate group only when the evidence supports that every selected candidate describes the same underlying entity. Each group must contain at least two distinct candidate numbers, and its `canonical_candidate_number` must be one of those numbers. A candidate may appear in at most one group.
 
-The response must conform exactly to the selected JSON schema. Return IDs only: do not copy candidate names, evidence, transcript text, source identifiers, or source ranges into the response.
+Omit uncertain matches and candidates that should remain distinct. Do not invent candidates or infer a number that is not explicitly supplied. An empty `duplicate_groups` array is valid.
+
+The response must conform exactly to the selected JSON schema. Return candidate numbers only: do not copy candidate names, evidence, transcript text, source identifiers, or source ranges into the response.

assets/generic/normalize/deduplication/schemas/semantic_reconciliation_llm.v1.json

@@ -11,9 +11,9 @@
       "items": {
         "type": "object",
         "additionalProperties": false,
-        "required": ["candidate_ids", "canonical_candidate_id"],
+        "required": ["candidate_numbers", "canonical_candidate_number"],
         "properties": {
-          "candidate_ids": {
+          "candidate_numbers": {
             "type": "array",
             "minItems": 2,
             "items": {
@@ -21,7 +21,7 @@
               "minimum": 1
             }
           },
-          "canonical_candidate_id": {
+          "canonical_candidate_number": {
             "type": "integer",
             "minimum": 1
           }

docs/adr/0013-use-request-local-candidate-handles-for-semantic-reconciliation.md

@@ -83,9 +83,7 @@ entity identity, checkpoint contracts, or cross-request correlation. Changes
 to shared protocol and policy assets must participate in the normal prompt,
 schema, and checkpoint fingerprint mechanisms.
 
-Acceptance of this decision does not imply that the shared mechanism or its
-consumer migrations are implemented. The
-[feature roadmap](../roadmap/semantic-reconciliation.md) owns target behavior
-and status, and the
-[implementation plan](../roadmap/implementation.md) owns delivery sequence
-until the work is complete.
+The shared mechanism and its initial D&D registry consumers are now
+implemented. Current behavior is documented in
+[Module Internals](../internal/modules.md#semantic-reconciliation) and
+[D&D Module Internals](../internal/dnd.md#semantic-registry-reconciliation).

docs/adr/0014-feedback-aware-validation-retries.md

@@ -0,0 +1,59 @@
+# ADR-0014: Use feedback-aware validation retries
+
+**Status:** Accepted
+**Date:** 2026-08-26
+
+## Context
+
+Validation can identify a candidate defect after a producer has returned an
+otherwise well-formed result. Retrying without the validator's deterministic,
+bounded feedback wastes the useful diagnosis, while treating validator
+execution failures as defects would ask a producer to repair conditions it
+cannot control. The mechanism must preserve typed producer ownership,
+checkpoint safety, and the repository's sensitive-data boundaries.
+
+## Decision
+
+The implementation will keep three independent budgets: the producer binding's
+outer `retries` budget, PromptKit's structured-output repair budget, and each
+validator's execution-retry budget. Validators will run sequentially in their
+configured order and aggregate both rejections and execution failures before a
+candidate disposition is selected.
+
+A correction-capable producer will provide the exact single LLM response that
+controlled its candidate using the `single_response_v1` protocol. A correction
+attempt will reconstruct the ordinary request and append exactly two fresh
+messages: that latest response as `assistant`, followed by one deterministic
+aggregate correction request as `user`. Earlier turns will not accumulate.
+
+Validator failures will not recurse into correction. Pipeline policy owns
+terminal disposition, with field-by-field producer overrides over pipeline
+defaults: structural failure and semantic rejection default to `fail_run`, and
+validator execution failure defaults to `warn_continue`. Validators can report
+facts and bounded corrective guidance, but never decide disposition.
+
+Rejected and structurally invalid candidates will not advance. A candidate
+allowed through after a validator execution failure will retain explicit
+incomplete-validation provenance and will not be checkpointed. Exact response
+and correction text remain attempt-local: they are excluded from ordinary
+errors, warnings, manifests, receipts, caches, checkpoints, and default debug
+summaries.
+
+## Alternatives considered
+
+- Retry every producer after any validation outcome. This conflates producer
+  defects with validator operational failures and wastes retry budget.
+- Let validators decide whether to continue. This would distribute pipeline
+  disposition policy across validators and undermine consistent defaults.
+- Reuse the full prior conversation. Accumulated turns introduce unbounded
+  prompt growth and make correction behavior depend on incidental history.
+- Persist raw responses to simplify diagnosis. Raw model output and correction
+  guidance may be sensitive and do not belong in durable pipeline records.
+
+## Consequences
+
+The framework gains transport-neutral correction and candidate contracts,
+producer capability checks, policy resolution, aggregated validation outcomes,
+and conservative checkpoint handling. Prompt construction remains inside the
+LLM adapter, while modules remain responsible for accurately exposing the
+single response that directly controlled a candidate.

docs/adr/0015-separate-process-warnings-from-quality-diagnostics.md

@@ -0,0 +1,67 @@
+# ADR-0015: Separate process warnings from quality diagnostics
+
+**Status:** Accepted
+**Date:** 2026-08-27
+
+## Context
+
+Notarius currently represents process degradation, incomplete validation,
+extraction-quality doubt, and routine normalization with one flat warning
+record. That makes ordinary successful runs noisy, loses the framework context
+needed to explain a finding, and gives `warning_count` no stable operational
+meaning. It also permits output encoders to add a warning after the durable
+warning file has already been written.
+
+The application needs one bounded diagnostic model that preserves exact
+occurrence counts while retaining only safe, representative samples. Fresh and
+resumed logical runs must present the same groups. The model must not alter
+validation decisions, retry budgets, rejected-output behavior, or process exit
+policy.
+
+## Decision
+
+Warnings are reserved for a completed run that advanced under an allowed
+process-level degradation or incomplete-work policy. Extraction-quality signals
+are advisories, and routine accepted transformations are observations. A
+non-degraded successful run therefore has zero actionable warnings.
+
+Modules and validators own a diagnostic's disposition, category, reason code,
+scope, and safe message. The framework adds pipeline origin, including stage,
+step, lane, module, validator, and chunk context where applicable. It then
+aggregates deterministically by disposition, category, reason code, and full
+origin. Chunk context remains on representative samples so equivalent findings
+across chunks aggregate together.
+
+Diagnostics carry exact occurrence counts, at most three distinct samples, and
+numeric omitted-sample metadata. Producers and validators are bounded to 64
+local groups. Final actionable warning groups are bounded without truncation;
+the non-warning collection may truncate represented groups while preserving an
+exact total occurrence count and explicit truncation metadata.
+
+The public contracts will be versioned: grouped actionable warnings use
+`notarius.warnings.v2`, grouped advisories and observations use
+`notarius.diagnostics.v1`, and the run receipt uses
+`notarius.run-result.v2`. Successful output encoders return logical files or
+an error; they do not add post-encoding warnings.
+
+## Alternatives considered
+
+- Keep one warning list and filter only CLI output. This would leave durable
+  consumers with the same semantically mixed, unbounded contract.
+- Map reason codes to severity in a central framework registry. This would
+  split module-owned meaning between synchronized policy tables and make new
+  diagnostic meaning implicit.
+- Preserve local omission warning records. They inflate visible group counts
+  and lose exact occurrence semantics.
+- Keep output-encoder warnings. A one-pass encoder cannot include those
+  records consistently in files it has already serialized; a two-phase encoder
+  protocol is deferred until a demonstrated need exists.
+
+## Consequences
+
+The framework gains validated diagnostic primitives, local collection,
+origin-aware aggregation, and versioned durable presentation. Existing warning
+transport remains temporarily while producers migrate. Current architecture,
+operator, integration, and internal documentation will describe the behavior
+only as each implementation step lands; this accepted decision does not claim
+that the migration is complete.

docs/adr/0016-feedback-aware-module-requested-retries.md

@@ -0,0 +1,76 @@
+# ADR-0016: Use feedback-aware module-requested retries
+
+**Status:** Accepted
+**Date:** 2026-08-28
+
+## Context
+
+An LLM-backed module can discover that a structurally valid model response is
+unsafe while translating it into a typed candidate, before the ordinary
+validator chain runs. Semantic registry reconciliation is the first such case:
+the shared core can identify invalid duplicate-group proposals, and typed
+application can reject a domain-incompatible group while preserving a safe
+partial result. Repeating the original request without the rejected response or
+corrective explanation gives the model no information with which to improve.
+
+The existing feedback-aware validation mechanism already establishes the safe
+correction protocol, but module-owned retry directives also carry internal
+reason codes, operator messages, and fallback diagnostics. Those values are not
+model instructions. Some module retry conditions, including exhausted
+structured-output repair, also have no valid exact response to append.
+
+## Decision
+
+An LLM-backed normalizer may attach optional, bounded model-facing correction
+guidance to a module-requested retry. Guidance is a separate contract field; the
+framework never derives it from a reason code, operator message, diagnostic, or
+error. A feedback-capable directive must include the exact model candidate that
+controlled the safe fallback and must use `single_response_v1`.
+
+The artifact-neutral producer-attempt state machine constructs the correction
+from that exact latest response and the supplied guidance. The producer rebuilds
+its complete ordinary request and appends the response as one assistant message
+and the guidance as one user message. Earlier attempts do not accumulate, and
+the attempt consumes the same configured stage retry budget as operational,
+structural, validation, and feedback-free module retries.
+
+A module retry without guidance remains valid and starts a fresh attempt. This
+is the required behavior when no valid model candidate exists, including after
+structured-output repair is exhausted. If feedback is supplied without a valid
+supported candidate, the framework fails the module contract rather than
+retrying blindly or inventing guidance.
+
+After retry exhaustion, the normalizer's safe fallback continues through its
+validator chain. Safe semantic groups may remain applied, unsafe groups remain
+separate, and bounded fallback diagnostics may explain the process degradation.
+Exact responses and correction text remain attempt-local and are excluded from
+ordinary errors, warnings, manifests, receipts, caches, and checkpoints.
+
+This decision extends, rather than supersedes,
+[ADR-0014](0014-feedback-aware-validation-retries.md): both correction sources
+use the same transport-neutral payload, replacement-request semantics, outer
+retry budget, and sensitive-data boundary.
+
+## Alternatives considered
+
+- Continue blind module retries. This preserves a smaller contract but wastes
+  the module's deterministic diagnosis and commonly repeats the same defect.
+- Convert module safety checks into validators. Typed reconciliation must apply
+  only safe proposal groups and retain a fallback before validation; moving
+  artifact-owned translation and application policy into validators would blur
+  stage ownership.
+- Copy the retry reason or operator message into the model request. Those values
+  are written for provenance and humans, can contain opaque internal labels,
+  and do not reliably describe a correct replacement.
+- Require feedback for every module retry. Structural failures may have no
+  valid exact candidate, so this would either prevent useful fresh retries or
+  fabricate prior-response material.
+
+## Consequences
+
+The normalize retry contract and generic producer-attempt directive gain an
+optional correction-guidance field and candidate-pairing validation. Modules
+that use it must provide semantically meaningful bounded prose and an exact
+candidate. Registry reconciliation maintains separate operator and model
+renderers, and policy fingerprints change so checkpoints created under blind
+retry behavior are not reused.

docs/adr/0017-defer-cache-access-policy-to-deployment.md

@@ -0,0 +1,59 @@
+# ADR-0017: Defer cache access policy to the deployment
+
+**Status:** Accepted
+**Date:** 2026-09-11
+
+## Context
+
+Canonical chunk plans and invocation checkpoints are reconstructible cache
+state. They may include source-derived content, so access and retention matter,
+but Notarius cannot determine whether a deployment needs a private cache or a
+cache shared by a trusted group. The current application-enforced `0700` and
+`0600` modes prevent ordinary Unix provisioning from making that choice.
+
+[ADR-0005](0005-cache-canonical-chunk-plans-by-source.md) governs canonical
+chunk-plan reuse, while [ADR-0006](0006-separate-output-cache-and-debug-state.md)
+established the output, cache, and debug separation refined by
+[ADR-0007](0007-separate-checkpoint-recording-from-reuse.md). None of those
+decisions requires Notarius to own exact cache access modes.
+
+## Decision
+
+Deployments own cache access policy through provisioned directories, process
+umask, setgid inheritance, and default ACLs. Notarius requests conventional
+maximum creation modes for cache entries and leaves the kernel to select their
+effective permissions.
+
+Notarius continues to own confined paths, symlink rejection, validation,
+atomic publication, cleanup, and recovery. Those correctness boundaries are
+independent from deciding which accounts may access cache state.
+
+Source-derived cache contents can be sensitive. Operators own their access and
+retention policy, while API credentials remain outside cache payloads under the
+existing secret boundary. Existing cache inode modes are not migrated
+automatically.
+
+This decision does not supersede ADR-0005, ADR-0006, or ADR-0007.
+
+## Alternatives considered
+
+- Continue enforcing private `0700` directory and `0600` file modes. This
+  makes one deployment choice for every operator and prevents trusted-group
+  provisioning.
+- Add configurable Notarius permission modes or a private/group cache setting.
+  This preserves application ownership of a deployment concern and expands the
+  public configuration surface without a durable policy reason.
+- Recursively converge existing cache ownership, modes, or ACLs. This would
+  make upgrades mutate operator-managed state and cannot safely infer an
+  intended policy.
+
+## Consequences
+
+Operators can use a restrictive umask for private cache state or provision a
+setgid/default-ACL cache root for trusted-group access. They must deliberately
+recreate or administratively migrate pre-existing cache trees when their
+existing modes do not match the desired deployment policy.
+
+Notarius does not add user or group management, ACL mutation, permission
+configuration, or recursive cache migration. Output and debug permission
+contracts remain separate from this decision.

docs/adr/0018-adopt-configkit.md

@@ -0,0 +1,53 @@
+# ADR-0018: Adopt ConfigKit for configuration loading
+
+**Status:** Accepted
+**Date:** 2026-09-13
+
+## Context
+
+Notarius's former configuration boundary owned strict YAML parsing, input
+presence bookkeeping, defaults, normalization, environment overlays, and
+configuration inspection. Those general-purpose responsibilities had grown
+alongside application-specific pipeline and module semantics.
+
+ConfigKit v0.3.0 provides strict typed YAML loading, imports, configuration
+profiles, explicit CLI overrides, recorded defaults and normalization, safe
+inspection rendering, and released source-origin APIs. The application needs
+those source records to resolve filesystem paths according to the file that
+supplied a surviving declaration.
+
+## Decision
+
+Notarius will adopt the root public `gitea.maximumdirect.net/eric/ConfigKit`
+package at v0.3.0 for its version 5 configuration contract. ConfigKit owns
+source composition, strict decoding, input presence, metadata, and safe
+effective/explanation rendering. Notarius retains its typed document schema,
+field semantics, runtime conversion, catalog-dependent pipeline resolution,
+PromptKit profile handling, and physical run preparation.
+
+Version 5 is a deliberate breaking file-contract change. It uses canonical
+object module bindings and reference sources, and enables ConfigKit imports
+and configuration profiles. The selected configuration file's canonical
+directory is the base for its declared application paths; ConfigKit's released
+`Resolution.InputInfo` and `Resolution.Source` APIs provide their origins.
+
+`NOTARIUS_CONFIG` remains file discovery, while field-level operational
+environment overlays are retired. Explicit configuration-valued CLI flags are
+represented as ConfigKit overrides; invocation actions remain outside that
+boundary.
+
+## Alternatives considered
+
+- Continue extending the custom YAML loader. This preserves duplicated
+  general-purpose parsing, presence, composition, and inspection machinery.
+- Retain version 4 alongside version 5. This would leave a supported dual
+  configuration boundary and delay removal of the machinery being replaced.
+- Wrap ConfigKit in a compatibility pre-parser. This weakens ConfigKit's
+  strict contract and recreates the application-owned parser complexity.
+
+## Consequences
+
+The released application has one ConfigKit-backed loader and one version 5
+contract. Operator documentation and maintained examples move with the cutover.
+ConfigKit metadata is retained only at the configuration publication boundary;
+its types do not become module or runtime pipeline interfaces.

docs/cli.md

@@ -10,14 +10,26 @@ defined in [Operations](operations.md).
 
 ~~~
 notarius help
+notarius --version
 notarius run <pipeline-id> --input path/to/source.json [--json] [flags]
-notarius config validate [--config path/to/config.yml] [--pipeline pipeline-id] [--only lane-a,lane-b]
-notarius pipelines list [--config path/to/config.yml] [--json]
+notarius config validate [--config path/to/config.yml] [--config-profile name] [--pipeline pipeline-id] [--only lane-a,lane-b]
+notarius pipelines list [--config path/to/config.yml] [--config-profile name] [--json]
 ~~~
 
 Running Notarius without arguments, or with **help**, **--help**, or **-h**,
 writes the command summary to standard output and exits with status 0.
 
+`notarius --version` is valid only as the sole root argument. It writes exactly
+`notarius <version>` followed by a newline to standard output and exits with
+status 0. A tagged `go install` build can report its main-module stable tag,
+and controlled builds can inject a stable tag at link time through
+`gitea.maximumdirect.net/eric/notarius/internal/buildinfo.Override`; an ordinary
+unversioned checkout reports `development`. Invalid injected version content is
+a runtime error with exit status 1, while extra `--version` arguments are a
+syntax error with exit status 2. This diagnostic does not replace the
+[run-result](integrations/run-result.md) or artifact contracts for downstream
+compatibility decisions.
+
 ## run
 
 ~~~
@@ -30,6 +42,7 @@ pipeline ID and **--input** are required.
 | Flag | Meaning |
 | --- | --- |
 | **--config path** | Use this configuration file. When omitted, configuration discovery applies; see [Configuration](config.md). |
+| **--config-profile name** | Select a ConfigKit configuration overlay. Empty and **default** select the composed base configuration. |
 | **--input path** | Source input file to process. Required. |
 | **--output-dir path** | Override the configured output root for this run. |
 | **--json** | Write the successful run-result receipt as JSON to standard output. |
@@ -43,8 +56,8 @@ pipeline ID and **--input** are required.
 | **--session-id id** | Override the generated prompt session identifier with a non-empty value for LLM-backed module calls. |
 | **--reasoning-effort value** | Replace the selected PromptKit profile's reasoning effort for every LLM-backed call in this run. The value must be non-empty and the flag may be specified only once. |
 | **--clear-reasoning-effort** | Clear reasoning effort inherited from the selected PromptKit profile for every LLM-backed call in this run. |
-| **--reference selector=path** | Add or replace a file reference binding. Repeatable. |
-| **--without-reference selector** | Remove a configured optional reference binding. Repeatable. |
+| **--reference selector=path** | Add or replace external file reference bindings at pipeline, lane, chunk, or binding scope. Repeatable. |
+| **--without-reference selector** | Remove matching configured external reference bindings. Repeatable. |
 
 **--chunk_cache** accepts only **auto**, **bypass**, or **refresh**.
 **--debug-dir**, **--output-dir**, **--session-id**, and
@@ -69,33 +82,62 @@ guidance.
 
 ### Reference selectors
 
-Use **--reference** only for a reference slot declared by the selected
-configured target. The accepted selector forms are:
+Use **--reference** only for reference slots declared by the selected
+configured targets. Qualification narrows the scope of an override:
 
 | Form | Target |
 | --- | --- |
-| slot=path | The unique selected target that declares slot. |
+| slot=path | Every selected target that declares slot. |
 | chunk.slot=path | The chunker. |
-| merge.slot=path | The unique selected merger that declares slot. |
-| lane.slot=path | The unique extractor, merger, or normalizer in lane that declares slot. |
+| lane.slot=path | Every extractor, merger, or normalizer in lane that declares slot. |
 | lane.extract.slot=path | The extractor in lane. |
 | lane.merge.slot=path | The merger in lane. |
 | lane.normalize.slot=path | The normalizer in lane. |
 
-**--without-reference** uses the same selector forms without =path. Slot
-names, requiredness, and configured bindings are part of the
+Pipeline- and lane-scoped selectors are expected to match multiple targets and
+fail if they match none. A stage-specific selector fails when its lane is not
+selected or its target does not declare the slot. There is no stage-wide
+`merge.slot` shorthand; name the lane when targeting a merger.
+
+CLI bindings override configured external paths. For overlapping CLI
+selectors, a binding-specific or chunk selector overrides a lane selector, and
+a lane selector overrides a pipeline selector. The last occurrence wins at
+equal scope. Binding and unbinding the same concrete target at equal scope is
+an error; a narrower bind or unbind may create an intentional exception to a
+broader action.
+
+**--without-reference** uses the same selector forms without `=path` and
+removes external bindings only. Neither flag replaces or removes a generated
+artifact handoff; an external/generated collision is a resolution error.
+Required slots are checked after all effective changes. CLI reference paths
+are resolved relative to the process working directory, so subprocess and
+service callers should use absolute paths. Slot names, accepted media types,
+size limits, requiredness, and configured generated bindings are part of the
 [configuration contract](config.md).
 
+For example, one shared campaign reference can reach every compatible target,
+with an optional lane-specific exception:
+
+~~~
+notarius run dnd-session \
+  --input /data/transcript.json \
+  --reference party=/data/references/party.txt \
+  --reference npc-registry.party=/data/references/npc-party-context.txt
+~~~
+
 ### Run output
 
 Without **--json**, standard output contains the completed pipeline ID, counts
 of normalized and rejected outputs, and the output directory. A debug-enabled
 run also prints its debug-bundle path to standard output. A successful run with
-warnings reports the warning count to standard error. The published JSON bundle
+actionable process warnings reports their group and occurrence counts to
+standard error. When the selected output module publishes `warnings.json`, the
+summary also reports that durable file's path. Advisory and observation findings
+do not produce a warning line. The published JSON bundle
 is defined by the [JSON output contract](integrations/json-output.md).
 
 With **--json**, successful standard output is exactly one
-`notarius.run-result.v1` JSON document followed by a newline, with no
+`notarius.run-result.v2` JSON document followed by a newline, with no
 human-oriented status or debug-path line. Its fields and compatibility policy
 are defined by the [run-result contract](integrations/run-result.md). A caller
 must check for exit status 0 before decoding this output; a failed write can
@@ -113,10 +155,10 @@ OPENROUTER_API_KEY=your-api-key \
 ## config validate
 
 ~~~
-notarius config validate [--config path/to/config.yml] [--pipeline pipeline-id] [--only lane-a,lane-b]
+notarius config validate [--config path/to/config.yml] [--config-profile name] [--pipeline pipeline-id] [--only lane-a,lane-b]
 ~~~
 
-This command loads and validates a configuration. With **--pipeline**, it also
+This command loads and validates the selected composed configuration. With **--pipeline**, it also
 resolves that pipeline against the production module catalog. **--only** selects
 lanes during that resolution and requires **--pipeline**.
 
@@ -139,10 +181,32 @@ OPENROUTER_API_KEY=validation-placeholder \
 The placeholder in the second command is sufficient only for offline
 validation; it cannot run a provider-backed pipeline.
 
+## config effective
+
+~~~
+notarius config effective [--config path/to/config.yml] [--config-profile name]
+~~~
+
+This command writes the selected composed configuration as deterministic,
+redacted YAML. Standard output contains only that YAML: it does not validate a
+pipeline, prepare modules, materialize references, or contact a provider.
+
+## config explain
+
+~~~
+notarius config explain [--config path/to/config.yml] [--config-profile name] <json-pointer>
+~~~
+
+This command writes ConfigKit's redacted, human-readable explanation for one
+canonical JSON Pointer. Supply `""` explicitly to explain the root. A malformed
+pointer is a syntax error; a well-formed unknown path is a command failure.
+The explanation can describe a value displaced by a selected profile. Standard
+output contains only the explanation.
+
 ## pipelines list
 
 ~~~
-notarius pipelines list [--config path/to/config.yml] [--json]
+notarius pipelines list [--config path/to/config.yml] [--config-profile name] [--json]
 ~~~
 
 This command lists configured pipeline IDs in sorted order. By default, it
@@ -159,7 +223,11 @@ go run ./cmd/notarius pipelines list \
 Successful commands write their primary result to standard output. Warnings and
 errors are written to standard error.
 
-For **run --json**, warnings remain on standard error and standard output is a
+Configuration load errors include ConfigKit's field diagnostics and source
+coordinates, followed by a source-ID legend identifying the acquired files or
+source labels. This includes imported files and the selected configuration profile.
+
+For **run --json**, actionable process warnings remain on standard error and standard output is a
 machine-readable success result only. Syntax and runtime diagnostics remain on
 standard error. Parse the result only after the process exits with status 0.
 
@@ -170,6 +238,6 @@ standard error. Parse the result only after the process exits with status 0.
 | 2 | The command or flag syntax was invalid, including unknown commands, missing required arguments, invalid flag values, or invalid flag combinations. |
 
 The root help spellings are the supported help path. Invoking **--help** on
-**run**, **config validate**, or **pipelines list** is handled by the flag
+**run**, a **config** subcommand, or **pipelines list** is handled by the flag
 parser as a usage error: it writes an error to standard error and exits with
 status 2.

docs/config.md

@@ -1,7 +1,7 @@
 # Configuration
 
 This is the canonical reference for Notarius configuration. Configuration files
-are YAML and must declare version 4. They select pipelines and their modules;
+are YAML and must declare version 5. They select pipelines and their modules;
 the [CLI reference](cli.md) owns invocation syntax, and
 [Operations](operations.md) owns run-state procedures.
 
@@ -19,12 +19,15 @@ The command fails if none of these paths provides a configuration file.
 For configuration values, precedence is:
 
 1. built-in defaults;
-2. the selected YAML file;
-3. supported operational environment variables; and
-4. the CLI run overrides that apply to a command.
+2. composed YAML sources, including imports;
+3. the selected ConfigKit profile; and
+4. explicitly supplied configuration-valued run flags.
 
-Environment variables do not provide a second configuration schema. They only
-override the fields listed below.
+`--config-profile` selects a ConfigKit overlay from `profiles/<name>.yml`
+beneath the selected configuration root. Empty and `default` select no overlay.
+Configuration profiles are distinct from PromptKit LLM profiles selected by
+`llm_profile` or `--llm-profile`. The retired operational environment variables
+have no effect; `NOTARIUS_CONFIG` remains a source-discovery input.
 
 ## Maintained Examples
 
@@ -33,7 +36,8 @@ override the fields listed below.
 - [Complete D&D configuration](../examples/dnd-complete.config.yml) uses
   ordered steps, all implemented D&D lanes, generated references, state
   settings, bounded LLM concurrency, and the maintained
-  [operator profile](../examples/profiles/dnd-extraction.yml).
+  [LLM profile](../examples/llm-profiles/dnd-extraction.yml) and imports its
+  pipeline from [a fragment](../examples/pipelines/dnd-session.yml).
 
 Use these complete files as starting points rather than combining the
 illustrative fragments in this reference.
@@ -46,7 +50,7 @@ other than **version** is optional.
 
 | Field | Type | Default | Rules |
 | --- | --- | --- | --- |
-| **version** | integer | none | Required; must be 4. |
+| **version** | integer or quoted string | none | Required; must be 5. Maintained files use `version: 5`. |
 | **promptkit** | object | none | Profile source and optional local-backend configuration. |
 | **pipelines** | map | empty | Maps pipeline IDs to pipeline definitions. |
 | **concurrency** | object | see below | Global LLM and extraction limits. |
@@ -69,6 +73,22 @@ Built-in defaults are:
 
 An empty cache directory in YAML deliberately selects the corresponding
 per-user root. An explicit empty output or debug directory is invalid.
+Selecting a chunk-plan or checkpoint directory does not configure filesystem
+access policy. Cache permissions come from deployment provisioning, process
+umask, setgid inheritance, and default ACLs; see [Operations](operations.md#cache-provisioning).
+
+Every nonempty relative configured destination or reference path resolves from
+the canonical directory of the file that supplies its surviving declaration.
+This includes imported fragments and selected configuration profiles. Built-in
+relative destinations use the selected main configuration file's directory;
+explicit `--output-dir` and `--debug-dir` values use the invocation working
+directory. Absolute values remain unchanged.
+
+The root file can use ConfigKit imports to compose disjoint definitions. Imports
+and selected profiles remain under the configuration root; they do not merge or
+append a pipeline's `steps` sequence. A profile replaces sequences as defined
+by ConfigKit, so changing an individual ordered binding requires supplying the
+whole replacement sequence.
 
 ## PromptKit Profiles
 
@@ -76,7 +96,7 @@ The optional **promptkit** object selects one source of profile definitions and
 may register one conventional local OpenAI-compatible backend:
 
 ~~~yaml
-version: 4
+version: 5
 
 promptkit:
   profile_dir: ./profiles
@@ -94,17 +114,18 @@ promptkit:
 | **local_backend.endpoint** | string | Required when **local_backend** is present; absolute HTTP or HTTPS URL with a host. |
 | **local_backend.concurrency_limit** | integer | Optional non-negative limit; defaults to 0. |
 
-Set at most one of **profile_dir** and **profile_file**. Relative values use
-the process working directory, not the configuration file's directory. The
-complete example's `./examples/profiles/dnd-extraction.yml` value is therefore
-valid when Notarius is launched from the repository root; use an absolute path
-for services and containers.
+Set at most one of **profile_dir** and **profile_file**. Every nonempty relative
+configured filesystem path is resolved from the canonical directory of the
+file that supplies it. The complete example's
+`./llm-profiles/dnd-extraction.yml` value therefore resolves beside that example.
+Use absolute paths for services and containers when their deployment layout is
+not configuration-relative.
 
 An operator source is optional. For a requested ID, PromptKit checks the
 configured operator source first, then Notarius's embedded fallback profiles,
 then its own built-in catalog. A matching profile is complete: it replaces a
 lower-precedence definition rather than merging with it. The maintained
-[`dnd-extraction` operator profile](../examples/profiles/dnd-extraction.yml)
+[`dnd-extraction` LLM profile](../examples/llm-profiles/dnd-extraction.yml)
 is a secret-free deployment artifact; production, development, and local
 deployments can each provide a complete definition with that same workload ID.
 Use workload-oriented IDs for new profiles instead of model names.
@@ -132,7 +153,11 @@ model: example-model
 Keep credentials out of the local-backend object. A PromptKit profile may name
 its credential environment variable through `api_key_env`; set that variable
 only in the run environment. PromptKit owns the
-[pinned profile-file format](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/formats.md).
+[pinned profile-file format](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.9.0/docs/formats.md),
+including `base_profile` inheritance. Notarius passes profiles through without
+merging them. Filesystem profiles cannot express PromptKit's in-memory
+`APIKeyRequired` setting; an unset `api_key_env` is optional and may reach the
+provider without authorization.
 The [PromptKit upstream boundary](integrations/pkg-promptkit.md) identifies the
 supported package API, and [Operations](operations.md#operational-limits)
 describes the effective concurrency layers.
@@ -143,30 +168,15 @@ requiring credential values. It rejects absent, malformed, or incompatible
 profiles before a run prepares modules. Credential availability is checked only
 when a generation is prepared.
 
-## Migrating Version 3 Configuration
+## Version 5 Migration
 
-Version 3 files are not decoded or rewritten. Change **version: 3** to
-**version: 4** and rename the top-level **scriptorium:** section to
-**promptkit:**. Version 4 decoding is strict, so a remaining **scriptorium**
-field is rejected as unknown.
-
-## Operational Environment Variables
-
-These variables are applied after YAML values:
-
-| Variable | Overrides | Rules |
-| --- | --- | --- |
-| **NOTARIUS_TOTAL_LLM_CONCURRENCY** | **concurrency.total_llm** | Integer. |
-| **NOTARIUS_STAGE_WORKERS_EXTRACT** | **concurrency.stage_workers.extract** | Integer. |
-| **NOTARIUS_OUTPUT_DIR** | **output.directory** | Non-empty path. |
-| **NOTARIUS_CACHE_CHUNK_PLANS_MODE** | **cache.chunk_plans.mode** | **auto**, **bypass**, or **refresh**. |
-| **NOTARIUS_CACHE_CHUNK_PLANS_DIR** | **cache.chunk_plans.directory** | Non-empty path. |
-| **NOTARIUS_CACHE_CHECKPOINTS_DIR** | **cache.checkpoints.directory** | Non-empty path. |
-| **NOTARIUS_DEBUG_DIR** | **debug.directory** | Non-empty path. |
-
-Integer values are trimmed then parsed as base-10 integers. Directory and
-output values reject NUL characters. **NOTARIUS_CONFIG** participates only in
-configuration discovery.
+Version 3 and version 4 files are rejected and are not rewritten at load time.
+Use `version: 5`, rename the former `scriptorium` section to `promptkit`, and
+convert every module binding to an object containing `module`. Convert every
+external reference to `{path: ...}` and move any lane-level references to the
+corresponding extraction binding. Put durable operational values in YAML or
+use the documented explicit run flags; the former field-level environment
+variables are not read.
 
 ## Concurrency, Output, Cache, And Debug
 
@@ -191,11 +201,14 @@ debug:
 **concurrency.total_llm** must be greater than zero. The only supported
 **concurrency.stage_workers** key is **extract**; its value must be from 1
 through **total_llm**. When omitted, it is recalculated from the effective
-**total_llm** after YAML and environment precedence.
+**total_llm** after imports, profile selection, and explicit ConfigKit
+overrides.
 
 **cache.chunk_plans.mode** accepts **auto**, **bypass**, or **refresh**.
 **cache.checkpoints.enabled** is a boolean. The CLI can override the output
 directory and chunk-plan mode for one run; see [CLI reference](cli.md#run).
+Cache directory fields select only locations; Notarius has no cache permission
+field or operational environment override.
 
 ## Pipelines
 
@@ -205,19 +218,27 @@ Each **pipelines** entry has a unique, non-empty ID and the following shape:
 pipelines:
   dnd-session:
     llm_profile: dnd-extraction
-    input: seriatim
-    chunk: generic
-    output: json
+    input:
+      module: seriatim
+    chunk:
+      module: generic
+    output:
+      module: json
     artifacts:
       spells:
-        extract: dnd/spells
-        merge: appendorder
-        normalize: dnd/spells
+        extract:
+          module: dnd/spells
+        merge:
+          module: appendorder
+        normalize:
+          module: dnd/spells
 ~~~
 
 | Field | Type | Default | Rules |
 | --- | --- | --- | --- |
 | **llm_profile** | string | none | Optional non-empty default PromptKit profile ID for selected LLM-backed bindings and validators. An explicitly present blank value is invalid. |
+| **structured_output_repair_attempts** | integer | prompt-owned (1 in maintained production prompts) | Optional structural-repair limit from 0 through 3 for selected LLM-backed bindings and validators. Omission leaves the prompt's declared policy in control; explicit 0 disables structural repair at that scope. |
+| **validation_policy** | object | see below | Optional terminal policy defaults for producer validation. Its fields inherit independently into chunk, extract, merge, and normalize bindings. |
 | **input** | module binding | none | Required. |
 | **chunk** | module binding | **generic** | Optional. |
 | **output** | module binding | **json** | Optional. |
@@ -237,6 +258,42 @@ run-level **--llm-profile** value first, then the binding's **llm_profile**,
 then the pipeline's **llm_profile**, and finally the PromptKit default.
 Deterministic bindings do not receive these defaults or run overrides.
 
+Structural output repair is resolved after module, validator, and `--only` lane
+selection. An object's **structured_output_repair_attempts** value takes
+precedence over the pipeline value; otherwise, an LLM-backed binding or
+validator inherits the pipeline value. If both are omitted, PromptKit uses the
+prompt's declared repair policy. The value must be an integer from 0 through 3;
+explicit `null` and non-integer values are invalid. An explicit value on a
+deterministic binding or validator is invalid, while a pipeline value simply
+does not apply to deterministic selections.
+
+`validation_policy` controls terminal disposition for one complete producer
+attempt and validator chain. It may appear on a pipeline or a **chunk**,
+**extract**, **merge**, or **normalize** module binding; input, output, and
+validator bindings reject it. Every field is optional and resolves in binding,
+pipeline, then application-default order:
+
+| Field | Values | Default |
+| --- | --- | --- |
+| **producer_structural_failure** | **fail_run**, **reject_output** | **fail_run** |
+| **semantic_rejection** | **fail_run**, **reject_output** | **fail_run** |
+| **validator_failure** | **warn_continue**, **fail_run** | **warn_continue** |
+
+The policy object and its fields must be non-null, and unknown fields are
+rejected. A deterministic producer may not explicitly set
+**producer_structural_failure** on its binding, although a pipeline-level
+default remains valid for pipelines that include LLM-backed producers.
+
+After the producer binding's retry budget is exhausted, an invalid structured
+response uses **producer_structural_failure**. One or more semantic validator
+rejections use **semantic_rejection**; rejection takes precedence over an
+exhausted validator failure or skip. With no rejection, an exhausted validator
+failure or skip uses **validator_failure**. `reject_output` records the
+terminal rejection without advancing that candidate. `warn_continue` is valid
+only for validator execution failure: it advances a structurally valid,
+otherwise unrejected result with incomplete-validation provenance and without
+making it reusable checkpoint state.
+
 A lane has these fields:
 
 | Field | Type | Default | Rules |
@@ -244,19 +301,20 @@ A lane has these fields:
 | **extract** | module binding | none | Required. |
 | **merge** | module binding | **appendorder** | Optional. |
 | **normalize** | module binding | **noop** | Optional. |
-| **references** | map | none | Supported compatibility alias for **extract.references**. |
+| **references** | map | none | Not supported at lane scope; place references on the eligible binding. |
 | **validators** | list | none | Non-empty lane-level lists are rejected. Set validator overrides on a binding instead. |
 
-The lane-level **references** alias remains accepted. When the alias and
-**extract.references** bind the same slot, **extract.references** wins. Use
-the binding-local form in new configurations.
+Lane-level **references** are not supported. Put a lane extraction reference
+under **extract.references**.
 
 ## Module Bindings And Validators
 
-Use a module key directly when no other binding fields are needed:
+Every module and validator binding uses an object, even when it has no fields
+beyond its module key:
 
 ~~~yaml
-input: seriatim
+input:
+  module: seriatim
 ~~~
 
 Use an object for fields:
@@ -267,24 +325,40 @@ extract:
   llm_profile: dnd-extraction
   retries: 2
   references:
-    spell_catalog: ./dnd-spell-catalog.json
+    spell_catalog:
+      path: ./dnd-spell-catalog.json
 ~~~
 
 | Binding field | Type | Default | Rules |
 | --- | --- | --- | --- |
 | **module** | string | none | Required for an object binding. Must be a registered compatible key. |
 | **llm_profile** | string | none | Optional non-empty PromptKit profile ID for an LLM-backed binding. It overrides the pipeline default unless the run supplies **--llm-profile**. |
-| **retries** | integer | 0 | Non-negative additional attempts for chunk, extract, merge, and normalize bindings. |
+| **structured_output_repair_attempts** | integer | pipeline or prompt-owned (1 in maintained production prompts) | Optional structural-repair limit from 0 through 3 for an LLM-backed binding. It overrides the pipeline value; explicit 0 disables structural repair. |
+| **validation_policy** | object | pipeline or application defaults | Optional field-by-field terminal-policy override for a chunk, extract, merge, or normalize binding. |
+| **retries** | integer | 0 | Non-negative additional complete producer attempts for chunk, extract, merge, and normalize bindings. This single budget covers operational errors, invalid structured output, module-requested normalization retry, and semantic correction. |
 | **options** | object | none | Must satisfy the selected module. |
 | **references** | map | none | Valid only on chunk, extract, merge, and normalize bindings. |
 | **validators** | list | production chain | Valid only on chunk, extract, merge, and normalize bindings. |
 
 Omitting **validators** uses the registered chain. **validators: []** selects
 an empty chain; a non-empty list replaces the chain in the listed order.
-Validator bindings accept only **module**, **llm_profile**, and **options**.
-They reject **references**, **retries**, and nested **validators**. Deterministic
-validators reject an explicit **llm_profile**. Deterministic module bindings
-also reject an explicit **llm_profile**.
+Validator bindings accept only **module**, **llm_profile**,
+**structured_output_repair_attempts**, **retries**, and **options**. Their
+**retries** value is a non-negative additional validator-execution budget and
+is valid only when the selected validator is LLM-backed. A validator retry
+rechecks the same immutable candidate; it never regenerates the producer.
+They reject
+**validation_policy**, **references**, and nested **validators**. Deterministic
+validators reject explicit **llm_profile** and
+**structured_output_repair_attempts**.
+Deterministic module bindings also reject those explicit fields.
+
+An LLM-backed chunk, extract, merge, or normalize producer with both a
+non-empty validator chain and positive **retries** must declare the supported
+single-response correction capability. Preparation rejects a configuration
+that could require semantic correction from a producer that cannot provide an
+exact prior response. A deterministic producer, or an LLM attempt that did
+not make a model call, cannot consume a semantic retry after rejection.
 
 The **json** output module accepts optional **include_chunk_map** and
 **evidence_context** settings:
@@ -324,18 +398,20 @@ When enabled, it publishes the selected source-unit excerpt defined by the
 
 ## References And Ordered Handoffs
 
-Reference maps bind named slots that the selected target declares. A scalar is
-an external path. Pipeline-level maps accept only external paths; step-local
-and binding-local maps may also select a normalized artifact from an earlier
-step:
+Reference maps bind named slots that the selected target declares. An external
+reference is an object with `path`. Pipeline-level maps accept only external
+paths; step-local and binding-local maps may also select a normalized artifact
+from an earlier step:
 
 ~~~yaml
 steps:
   - id: describe-session
     artifacts:
       npc-registry:
-        extract: dnd/npc-registry
-        normalize: dnd/npc-registry
+        extract:
+          module: dnd/npc-registry
+        normalize:
+          module: dnd/npc-registry
   - id: extract-events
     references:
       npc_registry:
@@ -344,8 +420,10 @@ steps:
           lane: npc-registry
     artifacts:
       spells:
-        extract: dnd/spells
-        normalize: dnd/spells
+        extract:
+          module: dnd/spells
+        normalize:
+          module: dnd/spells
 ~~~
 
 An artifact selector contains only **step** and **lane**. The producer must be
@@ -354,10 +432,16 @@ slot. A generated binding supplies one accepted normalized artifact; it does
 not name a file. A configured generated dependency remains required even when
 that consumer slot is otherwise optional.
 
-Pipeline references are defaults. A matching step-local or binding-local
-external path overrides a pipeline default. Required slots must be bound after
-these configuration values and any CLI reference overrides are applied.
-Reference paths in YAML are resolved relative to the configuration file.
+Pipeline references are configuration defaults. A matching step-local or
+binding-local external path overrides a pipeline default. CLI reference
+bindings are then operational overrides of configured external paths; their
+pipeline, lane, and binding scopes and precedence are defined by the
+[CLI reference](cli.md#reference-selectors). A CLI file reference cannot
+replace a configured generated artifact handoff. Required slots must be bound
+after configuration and CLI reference actions are applied. A YAML reference
+path resolves from the file that supplies its surviving declaration, including
+an imported fragment or selected profile. CLI reference paths resolve relative
+to the process working directory.
 
 ### D&D Reference Slots
 
@@ -447,11 +531,33 @@ Available validator keys are:
 | Item occurrences | **extract/dnd/item-occurrences/shape**, **extract/dnd/item-occurrences/registry**, **extract/dnd/item-occurrences/source_refs**, **extract/dnd/item-occurrences/source_relatedness**, **normalize/dnd/item-occurrences/invariants** |
 | Item registry | **extract/dnd/item-registry/shape**, **extract/dnd/item-registry/source_refs**, **extract/dnd/item-registry/source_relatedness**, **normalize/dnd/item-registry/identity** |
 | NPC occurrences | **extract/dnd/npc-occurrences/shape**, **extract/dnd/npc-occurrences/registry**, **extract/dnd/npc-occurrences/source_refs**, **extract/dnd/npc-occurrences/source_relatedness**, **normalize/dnd/npc-occurrences/invariants** |
-| Scene descriptions | **extract/dnd/scene-descriptions/shape**, **extract/dnd/scene-descriptions/source_refs**, **extract/dnd/scene-descriptions/source_relatedness**, **normalize/dnd/scene-descriptions/invariants** |
+| Scene descriptions | **extract/dnd/scene-descriptions/shape**, **extract/dnd/scene-descriptions/source_refs**, **extract/dnd/scene-descriptions/source_relatedness**, **extract/dnd/scene-descriptions/combat_semantics** (LLM-backed, opt-in), **normalize/dnd/scene-descriptions/invariants** |
 | Enemy events | **extract/dnd/enemy-events/shape**, **extract/dnd/enemy-events/engagements**, **extract/dnd/enemy-events/source_refs**, **extract/dnd/enemy-events/source_relatedness**, **normalize/dnd/enemy-events/invariants** |
 | Location registry | **extract/dnd/location-registry/shape**, **extract/dnd/location-registry/source_refs**, **extract/dnd/location-registry/source_relatedness**, **normalize/dnd/location-registry/identity** |
 | Location occurrences | **extract/dnd/location-occurrences/shape**, **extract/dnd/location-occurrences/registry**, **extract/dnd/location-occurrences/source_refs**, **extract/dnd/location-occurrences/source_relatedness**, **normalize/dnd/location-occurrences/invariants** |
 
+`extract/dnd/scene-descriptions/combat_semantics` is not in a production default chain. To opt in, replace the scene extractor validator chain with the current ordered chain plus the semantic validator last, and set a positive producer retry budget if a rejection should request a corrected scene:
+
+~~~yaml
+extract:
+  module: dnd/scene-descriptions
+  retries: 1
+  validators:
+    - module: generic/valid_json
+    - module: extract/dnd/scene-descriptions/shape
+    - module: extract/dnd/scene-descriptions/source_refs
+    - module: generic/valid_json_schema
+    - module: extract/dnd/scene-descriptions/source_relatedness
+    - module: extract/dnd/scene-descriptions/combat_semantics
+      retries: 1
+~~~
+
+An override replaces, rather than extends, the default chain. See [Module Bindings And Validators](#module-bindings-and-validators) for binding, profile, repair, retry, and failure-policy rules.
+
+The validator retry shown above permits one additional execution against the
+same scene candidate when the LLM-backed validator itself fails; it is separate
+from both the extractor's producer retry and PromptKit structural repair.
+
 When no override is configured, production D&D bindings use the following
 ordered chains. Each row lists extract then normalize; spell chains are the
 same at both stages.
@@ -484,7 +590,16 @@ go run ./cmd/notarius config validate \
 ~~~
 
 Configuration validation rejects invalid YAML, unsupported fields, invalid
-defaults or environment overrides, incompatible module keys, unknown options,
+defaults or explicit CLI overrides, incompatible module keys, unknown options,
 invalid reference bindings, missing required reference slots, invalid validator
 overrides, and incompatible generated artifact handoffs. Use
 [pipelines list](cli.md#pipelines-list) to inspect configured IDs.
+
+## Inspection
+
+Use [config effective](cli.md#config-effective) to write the selected composed
+configuration as deterministic, redacted YAML. Use
+[config explain](cli.md#config-explain) with a canonical JSON Pointer to inspect
+the effective or historical provenance for one path, including a value displaced
+by a selected profile. Neither command resolves a pipeline, prepares modules,
+materializes references, or requires provider credentials.

docs/consumers/dnd-pipeline.md

@@ -0,0 +1,221 @@
+# Consuming The Complete D&D Pipeline
+
+Use this workflow when an orchestrator runs the maintained complete D&D
+pipeline and consumes its structured JSON artifacts. The generic
+[subprocess consumer guide](subprocess.md) owns process-level responsibilities;
+this guide connects that workflow to the complete D&D configuration, its
+Seriatim input, and its artifact inventory.
+
+The [CLI reference](../cli.md), [configuration reference](../config.md),
+[run-result receipt](../integrations/run-result.md), and
+[published JSON output contract](../integrations/json-output.md) remain the
+canonical definitions of those public interfaces.
+
+## Prepare And Validate The Deployment
+
+Start from the maintained
+[complete D&D configuration](../../examples/dnd-complete.config.yml). It uses
+the `dnd-session` pipeline and demonstrates every implemented D&D lane, ordered
+artifact handoffs, campaign references, chunk-map publication, and evidence
+context.
+
+A deployment must provide its own PromptKit profile and campaign reference
+files. Use absolute paths for service and subprocess deployments. In
+particular, observe these different resolution rules:
+
+- reference and PromptKit profile paths in YAML are resolved relative to the
+  file that supplies their surviving declaration;
+- reference paths passed with `--reference` are resolved relative to the
+  Notarius process working directory.
+
+Do not copy the repository example's relative profile path into a deployment
+without preserving its configuration-relative layout. The complete path and
+profile rules are defined in [Configuration](../config.md).
+
+Preflight the deployed configuration before processing sessions and whenever
+it changes:
+
+```sh
+notarius config validate \
+  --config /absolute/path/to/notarius.yml \
+  --pipeline dnd-session
+```
+
+Provide credentials through the environment or the documented configuration
+mechanism. Do not put credentials in command arguments, generated
+configuration, or logs.
+
+## Supply The Transcript
+
+The complete pipeline consumes a Seriatim JSON document. The
+[Seriatim input contract](../integrations/seriatim.md) defines its required
+metadata, segments, and validation rules. Preserve segment IDs: D&D artifact
+citations use those segment IDs as source-unit ranges.
+
+When the caller maintains several transcript tiers, use the final trimmed JSON
+transcript so extraction operates on the same session content presented to
+later consumers. For example, Narratio identifies this implemented artifact as
+`narratio.transcript.final_trimmed` and normally stores it at
+`transcripts/final.trimmed.json`.
+
+Notarius generates a stable prompt session from the resolved input module and
+the exact input bytes. An ordinary orchestrator should not pass `--session-id`.
+Use that override only when intentionally changing the routing relationship
+between invocations; it is not a credential or output identity.
+
+## Run Notarius
+
+Invoke the pipeline with explicit absolute paths and request its
+machine-readable receipt:
+
+```sh
+notarius run dnd-session \
+  --config /absolute/path/to/notarius.yml \
+  --input /absolute/path/to/transcripts/final.trimmed.json \
+  --output-dir /absolute/path/to/notarius-output \
+  --reference party=/absolute/path/to/references/party.txt \
+  --reference players=/absolute/path/to/references/players.txt \
+  --reference glossary=/absolute/path/to/references/glossary.txt \
+  --reference spell_catalog=/absolute/path/to/references/spells.json \
+  --json
+```
+
+Each unqualified reference is pipeline-scoped: Notarius supplies it to every
+selected D&D target that declares the slot. A deployment may omit an optional
+reference it does not maintain, and may use the lane- or binding-qualified
+forms from the [CLI reference](../cli.md#reference-selectors) for an exceptional
+override. The registry, scene-description, combat-turn, and NPC-occurrence
+references declared between ordered steps in the complete configuration are
+generated artifacts. Do not pass those handoffs on the CLI.
+
+The caller should:
+
+- capture stdout and stderr separately;
+- propagate cancellation and impose an operator-appropriate timeout;
+- wait for process completion before interpreting stdout; and
+- retain stderr for diagnosis without copying secrets or transcript content
+  into other logs.
+
+Only exit status 0 permits decoding stdout as a receipt. Ignore stdout after a
+nonzero exit because a failed receipt write can leave partial bytes. The
+[CLI reference](../cli.md#output-streams-and-exit-statuses) defines the complete
+stream and exit-status contract.
+
+## Discover The Published Bundle
+
+Decode the successful stdout document as a supported run-result schema. For
+the current contract, `schema_version` is `notarius.run-result.v2`. Tolerate
+unknown fields allowed by that version, but reject an unsupported schema
+version.
+
+Use the receipt's absolute `output_directory` as the exact run-specific bundle
+root. Do not scan the output root for its newest directory, guess a run ID, or
+construct a bundle path. Resolve `index_file` beneath `output_directory` and
+reject an absolute logical path or any result that escapes the bundle root.
+
+The complete configuration uses the application validation defaults. A caller
+that requires fully validated D&D artifacts must also require receipt
+`validation_status: approved`; a successful `incomplete` result reflects the
+configured validator-failure continuation policy and carries its bounded
+validator provenance in `validation_summaries`.
+
+Read `index.json` and locate each requested lane in `output_files` by its exact
+`lane_id`. Do not guess a lane filename. Before decoding a payload:
+
+1. resolve its descriptor's relative `file` beneath the bundle root with the
+   same confinement check;
+2. verify the descriptor's media type and schema identity against the linked
+   artifact contract; and
+3. decode the payload according to that contract.
+
+The [published JSON output contract](../integrations/json-output.md) defines
+the index and bundle layout. Treat all paths obtained from a decoded external
+document as untrusted until confined to their documented root.
+
+## Complete Artifact Inventory
+
+When every configured lane is accepted, the complete example publishes these
+lane artifacts:
+
+| Lane ID | Purpose | Canonical contract |
+| --- | --- | --- |
+| `item-registry` | Canonical registry of encountered items and currency. | [Item registry](../integrations/dnd-item-registry-artifacts.md) |
+| `npc-registry` | Canonical registry of named NPCs. | [NPC registry](../integrations/dnd-npc-registry-artifacts.md) |
+| `location-registry` | Canonical registry of named locations. | [Location registry](../integrations/dnd-location-registry-artifacts.md) |
+| `scene-descriptions` | Classification, title, and summary for each scene. | [Scene descriptions](../integrations/dnd-scene-description-artifacts.md) |
+| `item-occurrences` | Source-grounded item discovery, acquisition, use, transfer, and loss events. | [Item occurrences](../integrations/dnd-item-occurrence-artifacts.md) |
+| `spells` | Source-grounded spell casts and casters. | [Spell casts](../integrations/dnd-spell-artifacts.md) |
+| `combat-turns` | Source-grounded combat turn participation. | [Combat turns](../integrations/dnd-combat-turn-artifacts.md) |
+| `npc-occurrences` | Source-grounded NPC interaction occurrences. | [NPC occurrences](../integrations/dnd-npc-occurrence-artifacts.md) |
+| `location-occurrences` | Source-grounded location occurrences. | [Location occurrences](../integrations/dnd-location-occurrence-artifacts.md) |
+| `enemy-events` | Source-grounded enemy combat events. | [Enemy events](../integrations/dnd-enemy-event-artifacts.md) |
+
+The JSON encoder always publishes these bundle-management files:
+
+| File | Purpose |
+| --- | --- |
+| `index.json` | Discovery document for lane and pipeline-wide artifacts. |
+| `manifest.json` | Run provenance and result summaries. |
+| `rejected.json` | Rejected pipeline outputs. |
+| `warnings.json` | Actionable process-degradation warnings. |
+| `diagnostics.json` | Advisory and observation findings for accepted artifacts. |
+
+The complete configuration also requests two pipeline-wide artifacts:
+
+- [`chunk-map.json`](../integrations/chunk-map.md), the accepted chunk plan and
+  chunk metadata; and
+- [`evidence-context.json`](../integrations/evidence-context.md), a reading
+  excerpt containing the union of selected cited source units and the
+  configured surrounding window.
+
+Discover both from their top-level `index.json` descriptors rather than
+treating them as lanes. Evidence context is convenient reading material, not
+authoritative provenance; citations in the normalized lane payloads remain the
+evidence contract.
+
+Every optional or lane file is published only when its corresponding artifact
+is available. A successful process does not guarantee that all configured
+lanes were accepted.
+
+## Decide What Counts As Consumer Success
+
+Exit status 0 means Notarius completed the pipeline and published its result
+bundle. The receipt or bundle may still report warnings, rejected outputs, or
+missing lane descriptors. A downstream consumer must define its own required
+artifact set explicitly.
+
+A caller that claims to consume the complete D&D workflow should normally
+require all ten lane IDs in the table and verify each descriptor's expected
+contract. If any required lane is missing, rejected, or incompatible, fail the
+caller's extraction step while retaining the Notarius bundle for diagnosis. A
+consumer that needs only a subset may define and document a narrower policy.
+
+Keep the successful receipt with the complete published bundle. Retain
+`manifest.json`, `rejected.json`, `warnings.json`, and captured process logs as
+required by the caller's provenance, diagnosis, and retention policies. Avoid
+selectively copying payload files without also preserving enough index and
+manifest information to identify their originating run and contracts.
+
+The transcript, lane artifacts, evidence context, manifest, debug data, and
+logs can all contain private campaign information. Apply the same access,
+publication, and retention controls used for the source transcript.
+
+## Consumer Checklist
+
+- Validate the deployed Notarius configuration and `dnd-session` pipeline.
+- Pass the final trimmed Seriatim JSON transcript with stable segment IDs.
+- Use absolute configuration, input, output-root, profile, and reference paths
+  in service deployments.
+- Capture stdout and stderr separately and enforce cancellation and timeout.
+- Parse stdout only after exit status 0.
+- Accept only supported receipt, index, and artifact schema versions while
+  tolerating permitted unknown fields.
+- Use the receipt's `output_directory`; never guess the run directory.
+- Confine `index_file` and every descriptor path to the published bundle root.
+- Discover lanes by `lane_id` and verify descriptor compatibility before
+  decoding payloads.
+- Enforce an explicit required-lane policy and inspect rejections and warnings.
+- Preserve the receipt and sufficient bundle provenance for every retained
+  artifact.
+- Protect all transcript-derived files and diagnostic streams as sensitive
+  campaign data.

docs/consumers/subprocess.md

@@ -6,14 +6,21 @@ statuses, while the [run-result receipt](../integrations/run-result.md) and
 [Published JSON Output contract](../integrations/json-output.md) own the
 durable result formats.
 
+For the maintained complete D&D workflow, including its transcript input,
+configured lane inventory, and downstream acceptance checklist, see
+[Consuming The Complete D&D Pipeline](dnd-pipeline.md).
+
 ## Run And Check The Process
 
 Optionally preflight a selected configuration and pipeline before work starts:
 
 ```sh
-notarius config validate --config /path/to/notarius.yml --pipeline pipeline-id
+notarius config validate --config /path/to/notarius.yml --config-profile production --pipeline pipeline-id
 ```
 
+Omit `--config-profile` when the composed base configuration is the intended
+selection; it is independent from any `--llm-profile` run override.
+
 Invoke the run with explicit paths and machine-readable output. Capture
 standard output and standard error separately; do not combine them before
 processing the result.
@@ -21,19 +28,43 @@ processing the result.
 ```sh
 notarius run pipeline-id \
   --config /path/to/notarius.yml \
+  --config-profile production \
   --input /path/to/source.json \
   --output-dir /path/to/output-root \
   --json
 ```
 
 Use absolute paths for supplied input, configuration, output-root, and
-reference files. Notarius generates a stable prompt session for the resolved
-input module and exact input bytes. Pass **--session-id** only when intentionally
-grouping different invocations under a different session. Supply credentials
-through Notarius's documented configuration and environment mechanisms, never
-as command-line arguments or generated secret-bearing configuration. In
-particular, a session identifier is provider-visible and is not a credential
-mechanism.
+reference files. Pass each external reference as its own argument-vector pair;
+do not construct and invoke a shell command. An unqualified reference selector
+supplies that file to every compatible selected target. Lane and stage
+qualification are available for exceptional overrides, while generated
+same-run references remain part of configured pipeline composition. The
+[CLI reference](../cli.md#reference-selectors) owns the exact selector and
+precedence contract.
+
+The maintained D&D subprocess workflow uses this facility for campaign context:
+
+```sh
+notarius run dnd-session \
+  --config /absolute/path/to/notarius.yml \
+  --input /absolute/path/to/transcripts/final.trimmed.json \
+  --output-dir /absolute/path/to/notarius-output \
+  --reference party=/absolute/path/to/references/party.txt \
+  --reference players=/absolute/path/to/references/players.txt \
+  --reference glossary=/absolute/path/to/references/glossary.txt \
+  --reference spell_catalog=/absolute/path/to/references/spells.json \
+  --json
+```
+
+Only pass the external references available to and desired by the deployment.
+Notarius generates a stable prompt session for the resolved input module and
+exact input bytes; reference changes do not change it. Pass **--session-id**
+only when intentionally grouping different invocations under a different
+session. Supply credentials through Notarius's documented configuration and
+environment mechanisms, never as command-line arguments or generated
+secret-bearing configuration. In particular, a session identifier is
+provider-visible and is not a credential mechanism.
 
 Wait for the process before interpreting standard output. Only an exit status
 of 0 permits decoding the receipt. On a nonzero exit, retain standard error for
@@ -65,10 +96,17 @@ lanes. The caller decides which lane IDs are required for its own work and
 which are optional; it should make that decision explicitly rather than infer
 failure from the receipt counts alone.
 
+When complete validation is required, also require receipt
+`validation_status: approved` and inspect `validation_summaries`. A successful
+run with `validation_status: incomplete` contains a structurally valid result
+that advanced after validator execution could not complete under the configured
+`warn_continue` policy. It is not reusable checkpoint state and should not be
+silently treated as fully reviewed by the caller.
+
 ## Preserve Provenance And Handle Data Carefully
 
 Keep the receipt with the published `manifest.json`, and retain
-`rejected.json` and `warnings.json` when review or later provenance requires
+`rejected.json`, `warnings.json`, and `diagnostics.json` when review or later provenance requires
 them. Treat the input, output bundle, cache, debug bundle, and captured process
 logs as potentially sensitive data. Apply the caller's access controls and
 retention policy, and avoid copying secrets into arguments, logs, or

docs/development.md

@@ -18,13 +18,14 @@ implemented component map.
 | Any documentation addition or revision | [Documentation Policy](policy/documentation.md) | It defines canonical homes, audiences, current-behavior rules, and maintenance requirements. |
 | Adding, changing, reviewing, or deleting tests | [Testing Policy](policy/testing.md) | It defines risk-based sufficiency, durable test boundaries, test-double guidance, and criteria for retaining tests. |
 | CLI composition or command behavior | [CLI Internals](internal/cli.md) and [CLI Reference](cli.md) | The internal guide owns composition and command flow; the reference owns public syntax. |
-| Building a subprocess caller or changing its result protocol | [Subprocess Consumer Guide](consumers/subprocess.md), [Run Result Receipt](integrations/run-result.md), and [CLI Internals](internal/cli.md) | These separate caller workflow, durable receipt contract, and CLI implementation behavior. |
+| Building a subprocess caller or changing its result protocol | [Subprocess Consumer Guide](consumers/subprocess.md), [Complete D&D Consumer Guide](consumers/dnd-pipeline.md), [Run Result Receipt](integrations/run-result.md), and [CLI Internals](internal/cli.md) | These separate generic caller workflow, the complete D&D workflow, the durable receipt contract, and CLI implementation behavior. |
 | Configuration loading, resolution, or user-visible configuration behavior | [Configuration Internals](internal/configuration.md) and [Configuration](config.md) | The internal guide owns loading and resolution mechanics; the reference owns the configuration contract. |
 | Pipeline resolution or execution | [Pipeline Internals](internal/pipeline.md) | It documents profiles, references, validation, retries, checkpoints, and runner behavior. |
 | Production modules or validators | [Module Internals](internal/modules.md), [D&D Module Internals](internal/dnd.md), and [D&D integration contracts](integrations/) | The generic guide owns extension mechanics, the D&D guide owns shared family conventions, and the contracts own durable output shapes. |
 | LLM clients, prompts, schemas, profiles, or scheduling | [LLM Runtime](internal/llm.md) | It documents the transport boundary and PromptKit integration. |
 | Output, cache, resume, or debug artifacts | [Run State Internals](internal/state.md), [Operations](operations.md), and [Configuration](config.md) | These separate implementation details, operator behavior, and configuration contracts. |
 | External input formats, artifact schemas, or durable output files | [Integration Contracts](integrations/) | Integration documents define external and durable data contracts. |
+| Release preparation, tagging, publication, or verification | [Source Releases](release.md) and [Documentation Policy](policy/documentation.md) | The release procedure owns maintainer guards and immutable-tag recovery; the policy assigns release-note ownership. |
 | Proposed or unimplemented behavior | [Roadmap](roadmap/) | Future work belongs only in roadmap documentation until implemented. |
 
 For an existing subsystem, also inspect its focused tests and the package-local

docs/integrations/json-output.md

@@ -9,7 +9,7 @@ Output configuration, including chunk-map and evidence-context publication, belo
 ## Bundle Layout
 
 All paths below are logical, relative, slash-separated bundle paths. The
-encoder always emits the first four JSON files below and adds lane or
+encoder always emits the first five JSON files below and adds lane or
 pipeline-wide artifact files when their corresponding artifacts are available:
 
 A subprocess caller first obtains the physical bundle root from the
@@ -21,7 +21,8 @@ root for the logical discovery described here.
 | `index.json` | Entry point that names the other published files and lane payloads. |
 | `manifest.json` | Run provenance and result summaries. |
 | `rejected.json` | Rejected pipeline outputs. |
-| `warnings.json` | Accepted-output and run warnings. |
+| `warnings.json` | Actionable process-degradation warnings. |
+| `diagnostics.json` | Accepted-artifact quality advisories and normalization observations. |
 | `lanes/<safe-lane-id>.json` | One normalized artifact payload for each lane. |
 | `chunk-map.json` | Optional accepted chunk map, when its export is enabled and available. |
 | `evidence-context.json` | Optional selected source-unit excerpt, when evidence publication is enabled. |
@@ -39,7 +40,8 @@ normalized lanes has this valid minimal index:
   "manifest_file": "manifest.json",
   "output_files": [],
   "rejected_file": "rejected.json",
-  "warnings_file": "warnings.json"
+  "warnings_file": "warnings.json",
+  "diagnostics_file": "diagnostics.json"
 }
 ```
 
@@ -49,6 +51,7 @@ normalized lanes has this valid minimal index:
 | `output_files` | Yes | Lane descriptors sorted by `lane_id`. |
 | `rejected_file` | Yes | Always `rejected.json`. |
 | `warnings_file` | Yes | Always `warnings.json`. |
+| `diagnostics_file` | Yes | Always `diagnostics.json`. |
 | `chunk_map` | No | Descriptor for the pipeline-wide `chunk-map.json`; never a lane descriptor. |
 | `evidence_context` | No | Descriptor for the pipeline-wide `evidence-context.json`; never a lane descriptor. |
 
@@ -92,7 +95,7 @@ group into the following externally observable summaries:
 | Run identity and result | `run_id`, `pipeline_id`, `pipeline_digest`, `schema_version`, `validation_status`, `started_at`, `completed_at` |
 | Resolved components | `input_module`, `chunker`, `extractors`, `merger`, `normalizer`, `output_encoder`, `artifact_lanes`, `validator_chains`, `module_metadata` |
 | Source and references | `source_digests`, `references` |
-| Published result summaries | `normalized_outputs`, `rejected_outputs` |
+| Published result summaries | `normalized_outputs`, `rejected_outputs`, `validation_summaries` |
 | Execution summaries | `chunk_plan`, `checkpoint_decisions`, `llm_profiles`, `metadata` |
 
 `references` records provenance such as the target, slot, origin, digest,
@@ -102,6 +105,16 @@ summarize results without embedding lane payload bytes. A chunk-plan summary is
 provenance for the plan used by this run; cache records, debug artifacts, and
 other operational state are not published as bundle files.
 
+Each `validation_summaries` entry is a bounded outcome for one producer result.
+It has required `status`, `producer_attempt_count`, and `terminal_action`;
+the stage and affected step, lane, module, or chunk identity are present when
+applicable. `status` is `complete`, `rejected`, or `incomplete`.
+`rejecting_validators`, `reason_codes`, and `incomplete_validators` preserve
+configured validator order and omit later duplicates. Entries contain no raw
+candidate response, correction guidance, validator diagnostic message, or
+artifact payload. The same shape may appear as `validation` on an affected
+rejection entry.
+
 When present, `metadata.session_id` is the effective non-secret routing
 correlation identifier used for the run. It can be visible to providers and is
 not a substitute for a cache or checkpoint identity. Its generation and
@@ -122,16 +135,57 @@ These values describe observed execution; they are not a backend-registration
 interface. Entries that differ by backend or effective reasoning remain
 distinct even when their profile, provider, and model are otherwise equal.
 
-## Rejections And Warnings
+## Rejections, Warnings, And Diagnostics
 
 `rejected.json` is always an object with a `rejected` array. Each entry has
 required `stage` and `message`; `step_id`, `lane_id`, `module_key`, `chunk_id`,
 `chunk_index`, `validator_name`, `reason_code`, `attempt_count`, and
-`diagnostic_artifact_path` are present only when applicable.
+`diagnostic_artifact_path` are present only when applicable. An entry may also
+contain the bounded `validation` summary described above; the existing singular
+validator and reason fields remain the first configured rejection for
+compatibility.
 
-`warnings.json` is always an object with a `warnings` array. Each warning has
-`reason_code` and `message`; `scope` is optional. Both arrays are empty when
-there is nothing to report.
+`warnings.json` is always the `notarius.warnings.v2` envelope:
+
+```json
+{
+  "schema_version": "notarius.warnings.v2",
+  "group_count": 0,
+  "occurrence_count": 0,
+  "groups": []
+}
+```
+
+It contains only process warnings. `group_count` is exact, and
+`occurrence_count` is the exact sum of its group occurrence counts.
+
+`diagnostics.json` is always the `notarius.diagnostics.v1` envelope:
+
+```json
+{
+  "schema_version": "notarius.diagnostics.v1",
+  "group_count": 0,
+  "occurrence_count": 0,
+  "truncated": false,
+  "unrepresented_occurrence_count": 0,
+  "groups": []
+}
+```
+
+It contains only advisory and observation groups. `group_count` counts groups
+represented in `groups`; `occurrence_count` includes both represented and
+unrepresented occurrences. When `truncated` is true,
+`unrepresented_occurrence_count` is the exact number omitted from group
+representation.
+
+Each group has `disposition`, `category`, `reason_code`, framework-owned
+`origin`, exact `occurrence_count`, bounded `samples`, and
+`omitted_sample_count`. Samples carry safe `scope` and `message`, plus a chunk
+ID and zero-based chunk index when applicable. A group retains at most three
+distinct samples. The framework fails rather than truncating actionable
+warnings beyond 128 groups; it represents at most 256 advisory/observation
+groups and records further occurrences through the diagnostic truncation
+fields above.
 
 ## Compatibility
 

docs/integrations/pkg-configkit.md

@@ -0,0 +1,13 @@
+# ConfigKit Integration
+
+Notarius uses the public root package
+[`gitea.maximumdirect.net/eric/ConfigKit`](https://gitea.maximumdirect.net/eric/ConfigKit)
+at v0.3.0 for version 5 configuration acquisition, imports, profiles, explicit
+overrides, diagnostics, and safe inspection rendering.
+
+Notarius owns the typed document, domain defaults and validation, runtime
+conversion, catalog-dependent pipeline resolution, source-aware configured
+paths, and command discovery. It uses only ConfigKit's public path, result,
+diagnostic, and resolution-source APIs. Upstream documentation owns general
+ConfigKit semantics; [Configuration](../config.md) owns Notarius's file and
+operator contract.

docs/integrations/pkg-promptkit.md

@@ -1,11 +1,11 @@
 # PromptKit Integration
 
 Notarius pins
-[`gitea.maximumdirect.net/eric/promptkit` v0.5.0](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0)
+[`gitea.maximumdirect.net/eric/promptkit` v0.10.0](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.10.0)
 as its in-process prompt engine. The upstream
-[Go package consumer guide](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/consumers/pkg-promptkit.md)
+[Go package consumer guide](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.10.0/docs/consumers/pkg-promptkit.md)
 owns the public engine API, and the upstream
-[format reference](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/formats.md)
+[format reference](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.10.0/docs/formats.md)
 owns prompt, profile, and schema file contracts.
 
 ## Supported Boundary
@@ -15,7 +15,8 @@ Notarius relies on the root `promptkit` package to:
 - construct an `Engine` with filesystem-backed prompt, schema, and optional
   operator and application-fallback profile sources;
 - prepare one frozen execution from a `RunRequest` with named inline artifacts,
-  variables, a direct session ID, prompt identity, and profile selection, then
+  variables, a direct session ID, prompt identity, profile selection, and
+  optional appended rendered messages, then
   record credential-redacted details and run that exact execution;
 - return rendered debug material, validated structured output, selected
   profile, backend, effective model metadata, and token usage;
@@ -26,7 +27,7 @@ Notarius relies on the root `promptkit` package to:
   admission exhaustion through `ErrCapacityExceeded`.
 
 The pinned
-[`BackendLocal`, `LocalBackend`, and `WithBackend` API](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/backends.go)
+[`BackendLocal`, `LocalBackend`, and `WithBackend` API](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.10.0/backends.go)
 owns the registration and backend-capacity contract.
 
 For one completion, the adapter calls `PrepareExecution`, takes a
@@ -52,7 +53,7 @@ Notarius sends one stable effective session through PromptKit's direct session
 field, which is authoritative for provider session behavior. It also retains
 the same value as the `session_id` prompt variable for maintained prompt
 compatibility. The generated identifier is 76 ASCII characters, within
-PromptKit v0.5.0's 256-code-point session limit. Session IDs are non-secret
+PromptKit's 256-code-point session limit. Session IDs are non-secret
 correlation identifiers and may be exposed to providers and provider
 observability. The CLI contract owns generation and override behavior.
 
@@ -84,7 +85,55 @@ configuration and deployment workflow are defined in
 [Configuration](../config.md#promptkit-profiles) and
 [Operations](../operations.md#promptkit-profile-deployment).
 
-Notarius supports this boundary against PromptKit v0.5.0. Its fallback source,
+PromptKit owns `base_profile` resolution under its
+[pinned format rules](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.10.0/docs/formats.md).
+Notarius records the selected leaf identity and resolved target without parsing
+or merging inheritance. An unset filesystem `api_key_env` is optional and may
+reach the provider without authorization, which can result in a 401 or 403.
+
+PromptKit accepts only the `developer`, `system`, `user`, and
+`assistant` text-chat roles after normalizing case and surrounding whitespace.
+Maintained Notarius prompt definitions use only `system` and `user`.
+
+For application-owned semantic correction, Notarius uses PromptKit's
+`RunRequest.AppendedMessages` after the ordinary rendered prompt. It supplies
+exactly two messages in order: the latest validated producer response with
+role `assistant`, then deterministic validation guidance with role `user`.
+It never exposes a general caller-selected role API, accumulates earlier
+correction turns, or changes the ordinary prompt prefix. Ordinary requests
+leave appended messages unset.
+
+PromptKit preserves supplied content but does not own Notarius's correction
+bounds. Notarius rejects invalid UTF-8, blank, or oversized assistant material
+(at most 1 MiB), guidance (at most 64 KiB), and combined content (at most
+1,114,112 bytes) before preparing the request. The transport-neutral
+application contract owns defensive copying and these limits. Default request
+and terminal summaries retain only safe counts, digests, identities, and usage;
+complete appended messages remain limited to the explicitly requested detailed
+debug trace.
+
+PromptKit obtains its maintained OpenRouter and Rakestrawhome backend and
+profile catalogs from independently versioned transitive modules. Notarius
+does not import or register either catalog; PromptKit retains catalog source,
+identity, precedence, credential, and capacity ownership.
+
+External catalog profile names and settings can change between dependency
+releases, including removal without aliases. Review directly selected IDs and
+inherited `base_profile` values when upgrading. An application-owned alias that
+inherits an external profile does not freeze that profile's settings.
+PromptKit's opt-in `promptkit-openrouter-standard` and
+`promptkit-openrouter-auto` names provide name continuity within its
+selected-and-tested dependency scope, not a fixed model or provider availability
+guarantee. See the upstream
+[catalog compatibility policy](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.10.0/docs/release.md#catalog-profile-compatibility).
+
+Use a complete operator-owned profile with explicit backend or endpoint, model,
+and settings when those choices must remain operator-controlled. The embedded
+`dnd-extraction` fallback already follows this model and does not inherit an
+external catalog profile. An upgrade updates Notarius's compiled-in catalog
+checkpoint identity so prior profile-source checkpoints are not reused.
+
+Notarius supports this boundary against PromptKit v0.10.0. Its fallback source,
 prepared-execution, inspection, and typed capacity APIs are used as public
 upstream contracts; other PromptKit APIs or file-format behavior are not
 implicitly supported. A dependency upgrade requires reviewing the adapter,
@@ -97,6 +146,10 @@ pinned upstream documentation.
 module assets, maps its transport-neutral completion contract, prepares and
 executes requests, validates output, records provenance, captures debug
 material, redacts errors, and preserves timeout ownership.
+
+PromptKit provider error details do not cross the ordinary completion boundary.
+Notarius exposes a provider-neutral generation category and optional status;
+redacted provider details are retained only in requested debug material.
 [D&D Module Internals](../internal/dnd.md) owns the embedded
 `dnd-extraction` fallback profile and the maintained D&D prompt defaults.
 [Configuration](../config.md#promptkit-profiles) defines how a Notarius
@@ -105,4 +158,7 @@ the conventional local backend.
 
 PromptKit API or format changes outside this boundary are not implicitly
 supported. Updating the pinned version requires reviewing the adapter and
-profile/configuration contracts against the upstream documentation.
+profile/configuration contracts against the upstream documentation. Maintained
+production prompts use PromptKit's bounded structural-repair contract; their
+current declaration is one additional repair attempt. Notarius retains the
+transport-neutral boundary and does not expose PromptKit types to modules.

docs/integrations/run-result.md

@@ -9,36 +9,58 @@ Command syntax, streams, and exit statuses are defined in the
 
 ## Schema
 
-The current schema version is `notarius.run-result.v1`.
+The current schema version is `notarius.run-result.v2`.
 
 | Field | Required | Meaning |
 | --- | --- | --- |
-| `schema_version` | Yes | Exactly `notarius.run-result.v1`. |
+| `schema_version` | Yes | Exactly `notarius.run-result.v2`. |
 | `run_id` | Yes | The finalized Notarius run identifier. |
 | `pipeline_id` | Yes | The effective pipeline identifier. |
 | `output_directory` | Yes | Absolute path to the published, run-specific output bundle. |
 | `index_file` | For the production JSON output | Logical path `index.json`; omitted for other output modules. |
 | `normalized_output_count` | Yes | Number of final normalized outputs. |
 | `rejected_output_count` | Yes | Number of recorded rejected outputs. |
-| `warning_count` | Yes | Number of final run warnings. |
+| `warning_group_count` | Yes | Exact number of actionable warning groups. |
+| `warning_occurrence_count` | Yes | Exact occurrences represented by actionable warning groups. |
+| `diagnostic_group_count` | Yes | Number of represented advisory and observation groups. |
+| `diagnostic_occurrence_count` | Yes | Advisory and observation occurrences, including unrepresented occurrences. |
+| `diagnostics_truncated` | Yes | Whether advisory/observation group representation was truncated. |
 | `validation_status` | Yes | The final run manifest validation status. |
+| `validation_summaries` | No | Bounded per-producer validation outcomes; present when producer work ran. |
 | `debug_directory` | No | Absolute path to the run-specific debug bundle when requested debug capture completed. |
 
 For the production `json` output module, `index_file` is present only when the
 completed run returned exactly one logical output file named `index.json`.
 For another output module, its absence does not indicate a failed run.
+`validation_status` is `approved`, `rejected`, or `incomplete`; `incomplete`
+means one or more otherwise accepted results advanced under validator-failure
+`warn_continue`.
 
 ```json
 {
-  "schema_version": "notarius.run-result.v1",
+  "schema_version": "notarius.run-result.v2",
   "run_id": "run-1770000000000000000-0123456789abcdef0123456789abcdef",
   "pipeline_id": "dnd-session",
   "output_directory": "/work/results/run-1770000000000000000-0123456789abcdef0123456789abcdef",
   "index_file": "index.json",
   "normalized_output_count": 6,
   "rejected_output_count": 2,
-  "warning_count": 1,
-  "validation_status": "rejected"
+  "warning_group_count": 1,
+  "warning_occurrence_count": 2,
+  "diagnostic_group_count": 3,
+  "diagnostic_occurrence_count": 5,
+  "diagnostics_truncated": false,
+  "validation_status": "incomplete",
+  "validation_summaries": [
+    {
+      "stage": "extract",
+      "lane_id": "spells",
+      "status": "incomplete",
+      "incomplete_validators": ["dnd/spells/source_refs"],
+      "producer_attempt_count": 1,
+      "terminal_action": "warn_continue"
+    }
+  ]
 }
 ```
 
@@ -48,8 +70,11 @@ For another output module, its absence does not indicate a failed run.
 paths. They identify the paths used by Notarius and do not resolve symlinks.
 `output_directory` is the run-specific bundle, not the configured output root.
 
-The receipt is a summary and discovery document. It does not contain lane
-descriptors, payloads, manifest data, rejections, warnings, or file contents.
+The receipt is a summary and discovery document. Its optional validation
+summaries contain only stable status, identity, validator names, reason codes,
+attempt counts, and terminal actions. It does not contain lane descriptors,
+payloads, manifest payloads, rejection messages, warnings, raw model responses,
+correction guidance, or file contents.
 For the production JSON output, resolve `index_file` beneath
 `output_directory`, reject path escapes, and use the
 [Published JSON Output contract](json-output.md) to discover logical files and

docs/internal/cli.md

@@ -24,19 +24,25 @@ preparation, and runner mechanics after their inputs are supplied.
 
 ## Dispatch And Configuration Handoff
 
-The root dispatcher handles help, configuration validation, pipeline listing,
-and a pipeline run. It normalizes injectable options before dispatch so that a
-missing production dependency fails as a command error rather than reaching
-execution.
+The root dispatcher handles help, version reporting, configuration validation,
+effective-configuration inspection, explanation inspection, pipeline listing,
+and a pipeline run. Version reporting resolves build
+information through `internal/buildinfo` before production composition, so the
+diagnostic remains available without configuration or runtime collaborators.
+The public syntax, streams, exit classes, and version semantics are defined by
+the [CLI reference](../cli.md). Other root commands normalize injectable
+options before dispatch so that a missing production dependency fails as a
+command error rather than reaching execution.
 
-Commands that need configuration use one shared loader. The CLI discovers the
-file, parses it through **internal/core/config**, starts from defaults, applies
-the file and supported environment overrides, and then validates it for the
-command. The configured discovery and precedence contract is in
+Commands that need configuration use one shared ConfigKit-backed loader. The
+CLI discovers the file, passes the requested profile and only explicit
+configuration-valued run overrides, and receives a **LoadedConfig**. The
+configured discovery and precedence contract is in
 [Configuration](../config.md), while the loading and resolution mechanics are
 in [Configuration Internals](configuration.md).
 
-Configuration validation without a selected pipeline checks structural
+`config effective` and `config explain` only render the loaded ConfigKit view;
+they do not resolve pipelines or construct providers. Configuration validation without a selected pipeline checks structural
 configuration only. Validation with a selected pipeline also builds the
 effective catalog, resolves the pipeline, and verifies every explicit effective
 PromptKit profile. Selected LLM-backed input, chunk, lane, output, and validator
@@ -80,8 +86,7 @@ field.
 After parsing and validating a run invocation, the CLI performs this ordered
 handoff:
 
-1. load and validate configuration, then apply command-level operational
-   overrides;
+1. load and validate configuration with ConfigKit command overrides;
 2. create and validate a safe run identity, then allocate a debug bundle only
    when requested;
 3. build the effective catalog, resolve requested reference changes, resolve
@@ -111,6 +116,16 @@ to checkpoint identity and `pipeline.RunInput`. The public flag and stability
 contract are defined by the [CLI reference](../cli.md#run); framework and LLM
 packages only transport the supplied value.
 
+The CLI also owns the scope grammar for reference flags. It enumerates the
+selected chunk and lane targets from registered module specifications, expands
+pipeline- and lane-scoped actions into exact stage-and-lane bindings, and
+resolves overlapping bind and unbind actions by specificity before calling
+configuration resolution. The generic pipeline therefore receives only exact
+`ReferenceBinding` and `ReferenceUnbind` values and has no knowledge of CLI
+selector syntax. Configuration resolution retains ownership of configured
+external/generated conflicts, required slots, and module compatibility; file
+materialization still occurs afterward.
+
 For `run --json`, the CLI constructs and encodes its private run-result receipt
 after a successful runner result is available, before it publishes logical
 output files. It writes the prepared receipt to standard output only after
@@ -162,8 +177,9 @@ is discoverable.
 - **internal/cli/production_contract_test.go** covers registrar composition,
   production catalog contents, assets, and representative configuration
   validation.
-- **internal/cli/reference_contract_test.go** covers CLI reference overrides,
-  origin separation, and materialization boundaries.
+- **internal/cli/reference_contract_test.go** covers scoped CLI reference
+  expansion, specificity, bind/unbind conflicts, generated-reference
+  protection, origin separation, and materialization boundaries.
 - **internal/cli/state_hardening_test.go** covers safe run identity, state
   roots, and failure ordering.
 

docs/internal/configuration.md

@@ -8,17 +8,17 @@ either interface.
 
 ## Boundary
 
-The configuration package turns a selected YAML file and supported environment
-values into a validated, independently owned configuration. It then resolves a
+The configuration package turns a selected ConfigKit-composed YAML document
+into a validated, independently owned configuration. It then resolves a
 requested pipeline against a module catalog before the framework prepares or
 runs anything.
 
 | Boundary | Inputs | Outputs | Does not own |
 | --- | --- | --- | --- |
-| Loading | Selected file path and environment lookup | Parsed file model and a populated **Config** | Choosing the file path or reporting a command result. |
+| Loading | Selected file path, profile, overrides, and invocation directory | **LoadedConfig** with ConfigKit inspection metadata and a populated **Config** | Choosing the file path or reporting a command result. |
 | Validation | **Config** | Structural configuration errors with pipeline, lane, or binding context | Module availability, capabilities, or construction. |
 | Resolution | Valid **Config**, selected pipeline and lanes, runtime reference changes, LLM override, and module catalog | **EffectiveConfig** with a **ResolvedPipeline** | Materializing reference bytes, preparing modules, execution, or filesystem state. |
-| Summary | **Config** or **EffectiveConfig** | Detached redacted payload suitable for debug summaries | Redacting arbitrary process state or provider traffic. |
+| Inspection | **LoadedConfig** and an optional JSON Pointer | Redacted effective YAML or explanation text | Pipeline selection, module preparation, or provider work. |
 
 The CLI discovers a configuration file, invokes this package, and supplies the
 result to the framework. Configuration never reads an input file, constructs a
@@ -28,23 +28,15 @@ at their respective [CLI](cli.md), [pipeline](pipeline.md), and
 
 ## Loading And Validation
 
-The CLI loads configuration in this order:
-
-1. parse the selected YAML file strictly into the file model;
-2. start from **Default**;
-3. apply the file model; and
-4. apply the supported environment overrides.
-
-This establishes the public precedence order without giving environment input a
-second file schema. Loading and application reject malformed YAML, unsupported
-file versions, unknown fields, invalid values, and identifiers that are empty
-or collide after whitespace normalization. The file application also makes the
-effective extraction-worker default follow the effective LLM limit. A present
-PromptKit local-backend object requires and trims its endpoint, defaults its
-omitted concurrency limit to zero, and is copied so the parsed file model
-cannot alias the populated **Config**. A pipeline `llm_profile` is
-presence-aware: omission remains empty, while a present blank value is
-rejected and a non-empty file value is trimmed before it reaches **Config**.
+ConfigKit acquires the selected root, imports, optional selected profile, and
+explicit CLI overrides in precedence order. Its document definition supplies
+typed defaults, source-aware path normalization, and structural validation
+before conversion to the independently owned runtime **Config**. Only
+`NOTARIUS_CONFIG` participates in discovery; retired field-level environment
+variables do not participate in loading. Version 5 decoding rejects scalar
+module bindings and scalar external references. ConfigKit retains redacted,
+historical inspection metadata while **LoadedConfig** keeps that implementation
+detail private behind effective and explanation writers.
 
 **Config.Validate** checks configuration-only invariants before resolution. It
 rejects incompatible profile sources, invalid state-surface values, unsupported
@@ -57,13 +49,14 @@ It deliberately does not require modules to be registered: that requires a
 catalog and belongs to resolution.
 
 The exact user-selectable values and validation rules are defined in
-[Configuration](../config.md). Keep additions to the file model, an
-environment override, its validation, and that reference in the same change.
+[Configuration](../config.md). Keep document schema, source-aware normalization,
+runtime conversion, and that reference aligned in the same change.
 
 ## Effective Resolution
 
-**Config.Resolve** first recomputes derived concurrency defaults and validates
-the configuration. It normalizes the requested pipeline ID, copies the selected
+**Config.Resolve** derives an absent extraction-worker value in an owned copy
+and validates the configuration. A programmatically supplied worker value is
+always explicit. It normalizes the requested pipeline ID, copies the selected
 profile, and passes the non-empty command-level LLM profile override, requested
 lane selection, and reference changes to the framework resolver.
 
@@ -74,10 +67,19 @@ profile-free, and no second inheritance decision occurs during execution. The
 public field definitions and precedence are owned by
 [Configuration](../config.md#pipelines).
 
+The resolver retains configured `validation_policy` overrides and derives one
+detached concrete terminal policy for the chunk producer and every lane's
+extract, merge, and normalize producers. That field-by-field inheritance is
+complete before preparation, and the effective values contribute to pipeline
+and checkpoint identity; execution does not interpret configuration defaults.
+
 The framework resolver supplies defaults, selects lanes, resolves validator
 chains, checks registered module and artifact compatibility, validates module
-options, and returns the fixed ordered pipeline shape. The resulting
-**EffectiveConfig** retains the selected ID, requested selection and reference
+options, and returns the fixed ordered pipeline shape. Positive validator retry
+budgets require an LLM-backed selected validator; deterministic validators are
+rejected during resolution. Eligible LLM-backed producer specifications also
+contribute their declared correction protocol to the resolved metadata. The
+resulting **EffectiveConfig** retains the selected ID, requested selection and reference
 changes, a clone of the input configuration, and the resolved pipeline.
 Callers may therefore retain or modify their input slices and maps without
 changing the resolved result, and later consumers cannot mutate the original
@@ -94,50 +96,44 @@ runtime error class described in the [CLI reference](../cli.md#output-streams-an
 
 The framework assigns the resolved pipeline a deterministic SHA-256 digest
 after defaults, lane selection, module bindings, reference bindings, validator
-chains, effective LLM profiles, and artifact schema identity have been
-resolved. The digest excludes
+chains, selected correction protocols, effective LLM profiles, and artifact
+schema identity have been resolved. The digest excludes
 its own stored value. It identifies resolved composition rather than raw YAML
 bytes, a debug payload, or all runtime state. The CLI records it as invocation
 provenance before execution; cache and checkpoint identity have additional
 owners in [Run State Internals](state.md).
 
-Configuration summaries must use **Redacted**, **RedactedSummaryPayload**, or
-**RedactedResolvedPipelinePayload**, never a direct configuration marshal.
-Those methods copy every binding and nested option container, replace values
-whose key is credential-shaped with **[REDACTED]**, and omit materialized
-reference content while retaining safe binding and reference provenance. The
-payload must not alias the source configuration or resolved pipeline.
-PromptKit's local endpoint and concurrency limit are preserved as non-secret
-configuration metadata in the independently owned summary; the object contains
-no credential value. This redaction is deliberately narrow: it protects
-configuration summaries and does not authorize recording arbitrary environment
-values or provider requests.
+Configuration publication uses ConfigKit's redacted effective YAML. Resolved
+pipeline publication uses **RedactedResolvedPipelinePayload**, which copies
+bindings and nested option containers, replaces credential-shaped values with
+**[REDACTED]**, and omits materialized reference content while retaining safe
+provenance. This redaction is deliberately narrow and does not authorize
+recording arbitrary environment values or provider requests.
 
 ## Invariants To Preserve
 
-- Defaults, YAML values, and environment values are applied in one direction;
-  later sources may override only their supported operational settings.
+- ConfigKit composition applies defaults, imports, an optional profile, and
+  explicit overrides in one direction with retained origin metadata.
 - A configuration is structurally valid before it is resolved, and a resolved
   pipeline is compatible with the supplied catalog before preparation begins.
 - Whitespace-normalized identifiers are unique wherever they identify a
   pipeline, step, lane, worker, or reference slot.
-- Resolution and summary generation return detached data. Redaction must cover
-  every configured and resolved binding, including nested validator bindings.
+- Runtime conversion and resolved-pipeline publication return detached data.
 - The resolved digest changes when resolved composition changes and never
   includes itself.
 
 ## Focused Tests
 
-- **internal/core/config/file_config_contract_test.go** covers strict file
-  parsing, normalization, file application, and structural rejection.
-- **internal/core/config/env_contract_test.go** covers supported operational
-  overrides and their precedence.
+- **internal/core/config/document_test.go** and
+  **operational_document_test.go** cover ConfigKit document loading,
+  normalization, imports, profiles, and overrides.
 - **internal/core/config/validation_contract_test.go** covers configuration
   invariants and contextual failures.
 - **internal/core/config/effective_config_contract_test.go** covers defaults,
   selections, overrides, resolution context, digest changes, and ownership.
-- **internal/core/config/redaction_test.go** covers recursive credential
-  redaction, reference-content exclusion, and non-aliasing payloads.
+- **internal/core/config/loaded_config_test.go** covers safe effective and
+  explanation rendering; **redaction_test.go** covers resolved-pipeline
+  redaction and reference-content exclusion.
 
 Run **go test ./internal/core/config** after changing this boundary. Changes to
 the handoff or resolved-composition semantics also need the focused framework

docs/internal/dnd.md

@@ -29,11 +29,17 @@ The D&D registrar registers the family’s artifact codecs, extractors, typed
 append-order mergers, normalizers, validators, prompt assets, fallback LLM
 profile asset, and default validator chains. Each extractor and normalizer has
 a stable module spec, explicit execution class, strict option decoding, and a
-typed builder. Scene chunking, every extractor, and NPC, location, and item-registry
-normalization are registered as `llm_backed`; the remaining current D&D mergers
-and normalizers are `deterministic`. The metadata is available to catalog inspection and
-resolved-pipeline debug data and determines which selected bindings inherit the
-pipeline profile. Configuration remains the canonical owner of the exact keys,
+typed builder. Scene chunking, every extractor, and NPC, location, and
+item-registry normalization are registered as `llm_backed`; the remaining
+current D&D mergers and normalizers are `deterministic`. The metadata is
+available to catalog inspection and resolved-pipeline debug data and determines
+which selected bindings inherit the pipeline profile. The registry normalizers
+use `single_response_v1`, forwarding corrections to their reconciliation
+completion and retaining the accepted raw proposal only as an owned model
+candidate. When deterministic proposal safety or typed application rejects a
+group, they provide separate model-facing prose for a corrective module retry;
+internal issue categories, reason codes, and operator messages remain
+diagnostic-only. Configuration remains the canonical owner of the exact keys,
 profile precedence, and validator order.
 
 Private structured-LLM response schemas are deliberately minimal. They reject
@@ -42,6 +48,13 @@ unknown fields, while preserving semantic candidates for deterministic
 validation. Do not promote a private response envelope into a durable schema;
 the contracts above define durable data.
 
+A D&D producer that declares `single_response_v1` forwards any supplied
+semantic correction to its structured completion and returns an owned copy of
+that completion's exact validated raw response as its model candidate. It does
+not serialize normalized artifacts to create that candidate, so deterministic
+identity, evidence, warning, and durable-schema behavior remains separate from
+the model transport material.
+
 ## Prompt Construction
 
 D&D LLM-facing content lives beneath `assets/dnd/`. Each module contributes a
@@ -108,14 +121,27 @@ presentation, and final ephemeral generic transcript windows. These orders and
 cache controls are prompt behavior; change them only through the owning
 manifest and prompt declaration.
 
+NPC, item, and location registry reconciliation translate shared proposal
+safety categories into bounded prose that references only the response-local
+duplicate-group ordinals and candidate handles. Item reconciliation appends its
+typed rule that currency may be consolidated only with aliases of the same
+denomination and never with non-currency items. The next normalize attempt
+receives that prose with the exact defective proposal under the shared
+replacement-request protocol. Structurally invalid output has no valid proposal
+candidate and receives a fresh feedback-free attempt instead. If the stage
+budget is exhausted, safe groups stay applied, unsafe groups stay separate, and
+one fallback warning summarizes the final defect without raw model content.
+
 ## Evidence, Candidates, And Normalization
 
 The current transcript is the only durable evidence source. Extractors assign
-the current source identity, preserve candidate evidence ranges for validators,
-and canonically order or remove exact duplicate ranges without asking the
-model to repair semantic errors. Campaign context and generated artifacts may
-ground names or control routing, but they never establish evidence for a D&D
-result.
+the current source identity and losslessly order any reversed range whose two
+endpoints resolve in that source, using transcript position rather than numeric
+unit-ID order. They then canonically order ranges and remove exact duplicates.
+This routine canonicalization does not request a retry or emit a warning.
+Unresolvable or wrong-source ranges remain unchanged for validators to reject.
+Campaign context and generated artifacts may ground names or control routing,
+but they never establish evidence for a D&D result.
 
 Default chains keep responsibilities separate: structural validators assess the
 candidate, source-reference validators resolve cited ranges against the current
@@ -125,15 +151,119 @@ relatedness validators report advisory evidence concerns. The configured order
 is documented in
 [Configuration](../config.md#production-validator-keys-and-default-chains).
 
+The optional `extract/dnd/scene-descriptions/combat_semantics` validator is the
+D&D family's LLM-backed review of only combat versus non-combat classification.
+It selects the shared combat-policy prompt fragment and asks the model to
+classify the current chunk independently as `combat` or `non_combat` without
+receiving the proposed scene kind. Deterministic code compares that
+classification with the proposed kind and either approves it or produces the
+appropriate correction guidance. This keeps every schema-valid classification
+interpretable and avoids anchoring the reviewer on the producer's answer. It
+does not assess titles, summaries, non-combat subtype, or scene boundaries;
+deferred boundary-coherence review remains separate. It is opt-in;
+[Configuration](../config.md) owns selection and retry/failure behavior.
+
+### Combat-semantics provider evaluation
+
+The human-reviewed corpus at
+`internal/modules/dnd/validate/scenedescriptions/combat_semantics/testdata/evaluation_cases.json`
+owns the proposed kind, expected combat classification, and reviewer rationale
+for each synthetic case. Its package test validates the fixture contract only.
+Provider evaluation remains an explicit maintainer operation and must not be
+added to the default offline test suite.
+
+Use the following protocol before proposing default-chain inclusion:
+
+1. Record the Notarius commit, prompt and schema fingerprints, provider, model,
+   profile settings, reasoning effort, structural-repair setting, number of
+   repetitions, and evaluation date before collecting results. Do not revise
+   expected classifications merely to agree with provider output; a substantive
+   corpus correction requires independent human review.
+2. Exercise the production validator construction and prompt assets from an
+   explicitly invoked, disposable evaluation driver or test in the validator
+   package. For each corpus case, construct transcript source units from the
+   listed IDs and text, assign matching per-unit source references, and use
+   `source.MaterializeChunkPlan` with one range spanning those units. Construct
+   exactly one scene whose ID and source range match that chunk and whose kind
+   is the case's `proposed_kind`; title and summary may use fixed placeholders
+   because the validator neither receives nor evaluates them. Invoke the typed
+   validator at extract stage through the production registry and scheduled LLM
+   client. Do not commit provider credentials, generated source material, or an
+   always-on live test.
+3. Compare the model classification with `expected_classification`, then verify
+   that its deterministic comparison with `proposed_kind` yields the expected
+   approval or rejection direction.
+   Record an unexpected approval of an expected rejection as a false
+   acceptance, an unexpected rejection of an expected approval as a false
+   rejection, and any validator execution failure separately from semantic
+   accuracy. Retain per-case results so repeated trials and systematic failure
+   modes remain visible.
+4. Evaluate producer correction separately with representative complete
+   scene-description runs configured as shown in
+   [Configuration](../config.md#production-validator-keys-and-default-chains).
+   For every initial semantic rejection, record whether the next producer
+   attempt returns the requested combat status and is approved. Do not count a
+   PromptKit structural repair as a producer-correction attempt.
+5. Run the correction evaluation with debug capture enabled and without reused
+   extraction checkpoints. Record added validator and producer calls, elapsed
+   latency, and cumulative prompt, completion, cached, and total token usage
+   from the debug attempt records. Compare these values with an otherwise
+   identical run whose scene-description chain omits the semantic validator.
+
+The default-chain review must consider classification error, false acceptance,
+false rejection, execution failure, producer-correction success, added calls,
+latency, and token use together. A structurally successful provider run alone
+is not evidence that the validator should become a default.
+
+Every producer-correctable D&D rejection describes all currently detectable
+corrections in transcript-grounded domain terms, using contextual names,
+model-owned artifact fields, and source segment ranges when useful. Validators
+collect independent record defects in one pass so one retry does not merely
+reveal the next issue. Shared D&D diagnostic helpers keep repeated rules and
+record descriptions stable, de-duplicated, and bounded; each artifact family
+continues to own the semantic rule and its prose.
+
+Operator diagnostics and model guidance are separate products of the same
+assessment. Operator messages may use typed paths, reason details, and opaque
+application identities. Correction guidance must not copy those messages or
+ask the model to reproduce durable entity IDs, hashes, validator module keys,
+reason codes, or Go field paths. A registry-normalization rejection instead
+speaks in terms of the duplicate-group proposal response the normalizer can
+actually revise. Normalization-only deterministic invariants retain useful
+operator detail but do not imply that a model controls derived ordering or
+identity. Only bounded actionable semantic guidance is eligible for a
+correction prompt.
+
+Private LLM schemas use simple enums for closed categorical fields when the
+provider-compatible shape can express the rule directly. Deterministic typed
+validators retain the same checks as defense in depth and for non-LLM
+producers. Private schemas keep every property required and avoid optional
+properties, `uniqueItems`, and conditional cross-field logic.
+
+Item-occurrence shape validation groups repeated holder mistakes by occurrence
+kind and gives the producer the required JSON null/non-null relationship. It
+identifies affected records by contextual item name and cited transcript range,
+never by the deterministically attached durable item ID. Holder mistakes remain
+semantic rejections rather than silent rewrites because changing a holder can
+also change the meaning of the occurrence kind.
+
 Enemy-event extraction additionally rejects a second `engaged` observation for
 the same comparison identity within one scene-scoped result. Normalization may
 combine results from distinct scenes, so it intentionally does not apply that
 rule. Configuration owns the exact validator key and chain position.
 
+Extraction source-reference validators share one full-span chunk-containment
+policy. After ordinary reference validity succeeds, the policy resolves both
+endpoints through document order and requires every source unit in the
+inclusive range to be present in the current chunk. It does not assume numeric
+unit-ID ordering, mutate input, or weaken wrong-source and unresolved-reference
+validation. Scene descriptions remain separate because their validator owns an
+exact one-scene range contract rather than general extraction containment.
+
 Normalizers are deterministic for spells, combat turns, item occurrences, NPC
 occurrences, scene descriptions, enemy events, and location occurrences. They
 canonicalize display values and evidence, use source-document order for stable
-output, and issue bounded warnings for changes or collapsed duplicates. NPC,
+output, and emit bounded normalization observations for changes or collapsed duplicates. NPC,
 item, and location registry normalizers are intentional exceptions: each first
 produces a deterministic candidate set, then may use a bounded structured-LLM
 proposal to reconcile identity groups.
@@ -142,13 +272,16 @@ proposal to reconcile identity groups.
 
 The three registry normalizers instantiate the domain-neutral
 `internal/framework/semanticreconcile` engine with default bounds. Each
-eligible candidate receives a contiguous, one-based `candidate_id` for that
-request. The model sees that handle, the candidate label and source-free
-evidence ranges, plus bounded transcript windows; it returns only duplicate
+eligible candidate receives a contiguous, one-based `candidate_number` for that
+request, and candidate material declares the exact inclusive range. The model
+sees that handle, the candidate label and source-free evidence ranges, plus
+bounded transcript windows; it returns only duplicate
 groups of supplied handles and one supplied canonical handle per group. It
 never returns names, evidence, durable IDs, or replacement records. Identical
 labels and evidence remain independently selectable because their handles are
-distinct.
+distinct. Transcript unit `id` values and evidence `start_unit_id` and
+`end_unit_id` values are source positions in a separate namespace and are
+never valid candidate numbers.
 
 The generic core owns the mandatory handle protocol, candidate and transcript
 presentation, the private response schema, source-reference validation,

docs/internal/llm.md

@@ -42,6 +42,22 @@ observability. The adapter returns PromptKit’s validated raw bytes rather than
 re-encoding the decoded target. An empty optional material is represented as
 one space so its named input is retained by PromptKit.
 
+When a request includes semantic correction material, the adapter validates and
+defensively copies it before preparation, then appends exactly two messages
+after the ordinarily rendered prompt: the prior response as an assistant
+message and the correction guidance as a user message. Requests without a
+correction do not add messages or introduce caller roles. Ordinary request
+summaries record correction byte counts and digests only; complete messages are
+available solely in an explicitly requested debug trace.
+
+The adapter leaves the ordinary rendered message prefix, named inputs,
+variables, session, profile, execution overrides, prepared-execution path, and
+PromptKit repair policy unchanged for a corrected request. It never imports a
+PromptKit message type into a module or pipeline contract. PromptKit reports
+actual structural repair count and cumulative token usage per completion; the
+pipeline's safe terminal debug record projects those values without copying
+message content.
+
 Client construction may also receive a run-wide reasoning-effort override from
 the CLI factory boundary. The adapter copies the caller-owned pointer and
 creates a fresh PromptKit execution override for each request: a nil pointer
@@ -75,7 +91,7 @@ backend membership as runtime without performing generation. Fallback assets
 are mounted only when at least one source is registered. The production D&D
 registrar contributes its `dnd-extraction` fallback, and the maintained D&D
 prompts select that logical ID by default. PromptKit owns source precedence and
-profile parsing: an operator-provided matching profile takes precedence over a
+profile parsing and inheritance: an operator-provided matching profile takes precedence over a
 fallback profile without Notarius merging either document.
 When the registration is absent, a profile selecting `backend: local` fails
 inspection instead of falling back to a built-in or endpoint-only target.
@@ -210,12 +226,32 @@ caller context takes precedence. The adapter does not retry capacity failures;
 the pipeline's existing binding attempt policy sees the operational error and
 decides whether to rerun the complete operation.
 
-Prompt-declared repair is executed within PromptKit’s structured-output flow.
-The current production D&D prompt manifests set repair attempts to zero. That
-setting does not replace pipeline retry behavior: a binding’s configured retry
-count reruns its stage attempt after an error or rejection, and an exhausted
-rejection is a recorded output rather than a provider error. The pipeline owns
-attempt lifecycle, validation chains, and retry diagnostics; see
+PromptKit executes structural repair within its structured-output flow. The
+maintained production prompt manifests declare one additional repair attempt.
+When a resolved binding supplies a repair value, the adapter inspects the
+prompt, copies its complete output contract, changes only the repair limit, and
+passes that complete replacement contract to PromptKit. This preserves the
+prompt's output format, validation mode, schema, and provider structured-output
+settings.
+
+A successful repair is an ordinary successful completion, not a warning. The
+adapter reports PromptKit's actual repair count and its cumulative usage
+directly, without adding the initial and corrective counts again. Debug prompt
+material records the configured complete contract; debug response material
+records the repaired response and actual validation result. If the repair
+budget is exhausted, the adapter retains the final raw bytes and debug material
+and reports `ErrInvalidStructuredOutput`. Generation failures during an initial
+or corrective call remain provider-neutral operational errors with the same
+redaction boundary.
+
+Structural repair does not replace pipeline retry behavior: a binding's
+configured retry count reruns its complete stage attempt after an operational
+or structural error, module-requested retry, or actionable semantic rejection.
+An actionable module-requested retry and a validator rejection both use the
+same correction payload when the producer exposes an exact latest response;
+feedback-free module retries reconstruct the ordinary request without appended
+messages. The pipeline owns attempt lifecycle, validation chains, and retry
+diagnostics; see
 [Pipeline Internals](pipeline.md#validation-retries-and-output) and the
 [binding reference](../config.md#module-bindings-and-validators).
 
@@ -247,6 +283,13 @@ surfaced; when the completion already failed, its call error remains the
 result. Debug-bundle location, retention, and handling are operational concerns
 documented in [Operations](../operations.md#debug-bundles).
 
+The attempt-terminal summary is a separate safe trace record: it contains
+attempt kinds, validator outcome counts and reason codes, effective policy,
+terminal action, and repair/usage references. It excludes raw assistant
+responses and correction text. Those values can appear only in the explicitly
+requested detailed prompt and response artifacts, which require sensitive-data
+handling.
+
 Run manifests receive selected profile summaries, including optional effective
 backend and reasoning provenance, and component identities—not prompt, schema,
 source, reference, or response content. The published field semantics belong
@@ -256,6 +299,9 @@ redacted before it crosses the runtime boundary. Known-secret redaction is
 available to other runtime collaborators; it does not make prompt or response
 contents safe for general logging.
 
+Generation failures expose an application-owned category and optional HTTP
+status. Provider code, type, and message remain debug-only, after redaction.
+
 ## Failure Boundaries
 
 - Construction fails for missing asset registries, mutually exclusive profile

docs/internal/modules.md

@@ -22,6 +22,15 @@ to bindings whose declared execution class is `llm_backed` and rejects a
 binding-specific profile on a deterministic module. The user-facing precedence
 contract belongs in [Configuration](../config.md#pipelines).
 
+An eligible LLM-backed chunk, extract, merge, or normalize producer may also
+declare correction protocol `single_response_v1`. That declaration is a
+promise that the implementation accepts one attempt-local semantic correction
+and returns an owned copy of the exact one model response that directly
+controlled the candidate. It must forward correction only to its structured
+completion request; it must not manufacture prior-response material by
+serializing a normalized artifact or expose opaque application IDs. Input,
+output, validator, and deterministic specs cannot declare the protocol.
+
 Implementations that accept options must provide both an option validator and
 a builder. The validator is used while resolving configuration; the builder
 decodes the same options and constructs the implementation from the prepared
@@ -36,6 +45,15 @@ they need, register each leaf implementation, and add any family-owned assets
 or default validator chains. They return contextual errors so production
 composition fails at startup rather than at the first run.
 
+A validator that returns a completed rejection must supply two separate
+bounded values: a stable `ReasonCode` for provenance and actionable
+`CorrectionGuidance` for the producer. Guidance identifies the semantic defect
+and the constraints on one complete corrected replacement. It must not contain
+validator keys, diagnostic paths, opaque application IDs, or other internal
+identifiers. An operator-facing `Message` may explain the same event, but the
+framework never copies it into a model request. Missing or invalid guidance is
+a validator contract failure.
+
 An artifact family can register an optional typed evidence projector alongside
 its codec. The projector returns defensive copies of the artifact's direct
 generic source references and must use the codec's exact Go type. It does not
@@ -77,19 +95,40 @@ source-backed artifact-family normalizer projects its deterministic records
 into contextual candidates and owned typed record envelopes, supplies its
 chosen prompt identity and resolved LLM profile, and constructs an engine with
 explicit limits. The core filters invalid evidence, assigns contiguous
-request-local integer handles, renders bounded candidate and transcript
-materials, invokes the structured-completion boundary, and assesses the
-returned duplicate groups into a stable non-overlapping plan.
+one-based request-local candidate numbers, renders bounded candidate and
+transcript materials, invokes the structured-completion boundary, and assesses
+the returned duplicate groups into a stable non-overlapping plan. Candidate
+material declares the exact inclusive number range for the request. Transcript
+unit IDs and evidence range endpoints remain source positions in a separate
+namespace and are never valid candidate numbers.
 
 The normalizer then applies that plan through a typed `ApplicationPolicy`. The
 core preserves ungrouped records, contribution order, and provenance while the
 artifact family owns group guards, field and evidence consolidation, durable
-ID derivation, retry and fallback presentation, warnings, and postconditions.
-Request-local handles do not enter the typed value or durable artifact. Fewer
-than two eligible candidates skips model invocation; exceeding a candidate or
-combined-material bound preserves the deterministic result under the family's
-fallback policy. Provider, transport, cancellation, and context-construction
-failures remain execution errors.
+ID derivation, retry and fallback presentation, classified diagnostics, and
+postconditions.
+These candidate numbers are the concrete private representation of ADR-0013's
+request-local handles; they do not enter the typed value or durable artifact.
+Fewer than two eligible candidates skips model invocation; exceeding a
+candidate or combined-material bound preserves the deterministic result under
+the family's fallback policy. Provider, transport, cancellation, and
+context-construction failures remain execution errors.
+
+When the engine actually makes a proposal call, its typed result carries the
+owned exact proposal response under the same correction contract as other
+eligible producers. Deterministic skip, limit, and fallback outcomes carry no
+model candidate, so a later rejection applies terminal policy without spending
+an ineffective semantic retry.
+
+When proposal assessment or typed group application rejects a structurally
+valid group, the normalizer may return its safe partial value with a
+module-requested retry. A feedback-capable directive supplies bounded
+model-facing correction guidance separately from operator diagnostics and
+retains the exact proposal response as its candidate. The shared stage retry
+mechanism appends that response and guidance to a fresh complete request. A
+structurally invalid completion has no valid candidate and therefore requests a
+feedback-free fresh attempt. On exhaustion, only the final safe fallback and
+its bounded process diagnostic advance to validation.
 
 The core supplies a conservative generic prompt and the single private
 response schema. A domain prompt may substitute its semantic instructions but
@@ -111,9 +150,13 @@ its domain prompt.
 3. Implement strict option decoding, construction, and the typed stage
    interface. Preserve caller ownership: do not retain mutable request data
    and return defensive copies where an implementation exposes stored data.
+   If declaring correction capability, forward the request correction and
+   retain only the exact validated response that controlled the result.
 4. Register the module through its typed registry helper and add it to the
    owning family registrar. Add a default validator chain only when that
    family owns the behavior; otherwise require an explicit compatible chain.
+   Every rejection path in a validator must provide actionable correction
+   guidance while retaining its stable internal reason code.
 5. Update the selectable-key and chain reference in
    [Configuration](../config.md#production-module-keys), the applicable
    integration contract, and focused tests. Keep the configuration document

docs/internal/overview.md

@@ -25,7 +25,8 @@ physical state roots.
 | Area | Implemented owners | Responsibility |
 | --- | --- | --- |
 | Executable and command boundary | **cmd/notarius**, **internal/cli** | Process entry, command dispatch, configuration discovery, production composition, runtime collaborator setup, durable file placement, and user-facing reporting. |
-| Configuration | **internal/core/config** | Defaults, strict YAML parsing, environment overrides, structural validation, effective resolution, redaction, and resolved-composition summaries. |
+| Build information | **internal/buildinfo** | Resolves a stable linked release tag or build metadata for the diagnostic root version command. |
+| Configuration | **internal/core/config** | ConfigKit document definition, source-aware normalization, runtime conversion, structural validation, safe inspection, effective resolution, and resolved-composition redaction. |
 | Generic models | **internal/core/source**, **internal/core/artifacts**, **internal/framework/contracts** | Source documents and chunks, manifests and provenance, plus typed artifact, reference, validation, output, and structured-completion contracts. |
 | Pipeline framework | **internal/framework/pipeline** | Registries, profile and reference resolution, typed preparation, validation, retry coordination, ordered execution, handoff, and result assembly. |
 | LLM and prompt runtime | **internal/framework/llm**, **internal/framework/promptfs** | Provider-neutral structured completions, scheduling, profile recording, prompt assets, schema registration, and credential-shaped-value redaction. |

docs/internal/pipeline.md

@@ -12,7 +12,7 @@ own durable output shapes. Concrete production extensions are covered by
 The pipeline framework accepts a resolved composition, registries, shared
 dependencies, input bytes, a supplied prompt session, and state/debug
 collaborators. It returns logical output files, normalized artifacts, recorded
-rejections and warnings, manifest provenance, and checkpoint decisions. The
+rejections, grouped diagnostics, manifest provenance, and checkpoint decisions. The
 CLI owns process arguments, configuration discovery, session resolution,
 physical roots, and placement of returned output files.
 
@@ -32,13 +32,26 @@ Resolution turns a configured pipeline profile into a **ResolvedPipeline**.
 It normalizes the pipeline and lane identities, applies stage defaults, selects
 requested lanes where that is supported, resolves validator chains, checks
 module capabilities and typed artifact compatibility, validates options, and
-assigns a deterministic resolved-composition digest. The resolved pipeline
-contains bindings and declared reference targets, not external reference bytes.
+assigns a deterministic resolved-composition digest. A correction protocol is
+selected from each eligible LLM-backed producer specification and becomes part
+of that resolved identity; only `single_response_v1` is currently supported.
+Preparation rejects an LLM-backed producer that combines a non-empty validator
+chain with positive producer retries unless it declares that protocol. Producers
+without validators or without retries remain valid without correction support.
+The resolved pipeline contains bindings and declared reference targets, not
+external reference bytes.
 After selection, the resolver applies command, binding, and pipeline profile
 precedence to LLM-backed bindings and validators only; prompt defaults remain
-an empty resolved binding profile. Deterministic bindings remain profile-free.
-These effective values are part of the digest, so execution and checkpoint
-consumers do not repeat profile inheritance.
+an empty resolved binding profile. It resolves structural output repair
+separately: a binding's `structured_output_repair_attempts` value wins, then a
+pipeline value applies to LLM-backed bindings and validators, and omission
+leaves the prompt-owned policy intact. An explicit repair value on a
+deterministic binding is rejected. Resolved bindings own copied repair values,
+and these effective values are part of the digest, so execution and checkpoint
+consumers do not repeat profile inheritance or configuration resolution.
+Each LLM request receives its own copy of that resolved value. PromptKit spends
+it only for structural correction inside one completion; the runner's binding
+retry policy remains the separate outer budget for complete stage attempts.
 Configuration resolution supplies the selected profile and catalog; see
 [Configuration Internals](configuration.md).
 
@@ -52,9 +65,11 @@ remains declared but has no bytes until its producing step completes.
 
 Preparation is the construction boundary. It validates the resolved shape and
 registry set, clones the resolved data, then constructs the input adapter,
-chunker, stage-local validators, every typed lane, and output encoder. Each
-registered builder receives its own cloned build request immediately before its
-module-owned code runs. Preparation also collects stable checkpoint
+chunker, stage-local validators, every typed lane, and output encoder. The
+prepared producer metadata preserves each selected correction protocol, and
+the resolved digest carrying that metadata participates in checkpoint identity.
+Each registered builder receives its own cloned build request immediately
+before its module-owned code runs. Preparation also collects stable checkpoint
 fingerprints. Missing registrations, incompatible typed entries, nil
 implementations, and constructor failures are reported before source parsing
 or any stage operation begins.
@@ -82,6 +97,13 @@ codec, checks its complete schema and media identity, and records a content
 digest plus bounded producer provenance. A missing, ambiguous, invalid, or
 incompatible producer prevents the consumer step from starting.
 
+Resolution receives only exact stage-and-lane operational reference overrides.
+The CLI may offer broader pipeline- or lane-scoped selectors, but expands and
+arbitrates those before entering the framework. External overrides are applied
+after configured external defaults and local bindings. They cannot coexist
+with a generated binding for the same target and slot, and external unbinds do
+not remove generated handoffs.
+
 ## Execution And Ordering
 
 The runner validates its input, installs no-op state collaborators when none
@@ -121,17 +143,55 @@ their target: chunks, codec-decoded typed candidates, or serialized codec
 bytes. Each typed validator receives a newly decoded value from the one
 candidate serialization for that attempt, while serialized validators receive
 separately owned representation bytes and schema metadata. They may approve,
-approve with warnings, reject, or fail. A rejection is an ordinary pipeline
-result; a validator error is a framework error.
+approve with warnings, reject, fail, or be skipped when a runtime prerequisite
+is unavailable. The shared executor settles every configured validator in
+order. A failed LLM-backed validator retries only itself against the same
+immutable candidate; it does not regenerate the producer or alter the
+validator request. Rejections stop that validator, while other configured
+validators still run. The executor retains ordered results, bounded
+deduplicated correction guidance from rejections, and only the final exhausted
+failure outcome for each validator. The correction builder keeps first
+occurrence order, omits internal reason codes, validator names, and operator
+messages, and requests one complete replacement. Missing guidance or an
+oversized aggregate is a framework contract error; guidance is never inferred
+or truncated.
 
 The runner applies the binding's retry policy around a stage operation and its
-complete validation chain. It preserves warnings only from the final accepted
-or rejected attempt. Cancellation stops retries. Normalizer-specific retry
-directives consume this same budget and validate any final safe fallback through
-the normalizer chain.
+complete validation chain. It preserves terminal diagnostics only from the final accepted
+or rejected attempt, plus one fixed validation-incomplete warning per validator whose execution
+budget was exhausted under `warn_continue`. Cancellation stops retries.
+Normalizer-specific retry directives consume this same budget and validate any
+final safe fallback through the normalizer chain. A directive may carry bounded
+correction guidance only when it also exposes the exact latest
+`single_response_v1` candidate. The state machine then uses the same replacement
+request shape as validator correction. A directive without guidance clears any
+prior correction and starts a fresh attempt, which preserves structural retry
+behavior when no valid response exists.
+
+The artifact-neutral producer-attempt state machine owns that shared budget,
+attempt provenance, semantic-correction material, and terminal-policy
+selection. It accepts producer and complete-validation closures, so artifact
+materialization, cache handling, checkpoints, and debug output stay at the
+operation boundary. It distinguishes operational, structural, module-requested,
+and validator-semantic retries. Model feedback from either semantic source is
+available only for a valid latest `single_response_v1` candidate. A
+deterministic or no-model validator rejection instead settles the semantic
+policy immediately, while a feedback-free module directive remains an ordinary
+fresh retry. Structural-output errors alone use the structural policy, and
+validation failure without rejection settles the validator-failure policy
+without regenerating the producer.
+
+Chunk planning uses this state machine for generated plans. A rejected or
+validation-incomplete automatic cache hit is not model material and therefore
+falls through to a fresh initial generation at producer attempt one; it neither
+receives a correction, consumes retry budget, promotes cached-candidate
+diagnostics, nor overwrites the stored record. An incomplete cache validation
+under `fail_run` terminates instead. Only a newly generated, completely
+validated plan is published to the chunk-plan store. Rejected plans never
+advance, and validation-incomplete plans remain unpublishable.
 
 After terminal lane work, the runner assembles manifest provenance, normalized
-artifacts, rejections, warnings, and an optional accepted chunk map. When an
+artifacts, rejections, final grouped diagnostics, and an optional accepted chunk map. When an
 output policy selected evidence lanes, it decodes accepted serialized normalize
 outputs through their registered codecs and invokes the prepared typed
 projectors. Rejected or absent lanes contribute nothing. This reconstruction is
@@ -142,6 +202,15 @@ The CLI publishes those files only after the runner returns without a framework
 error. Logical file names and schemas are defined by the [output integration
 contracts](../integrations/).
 
+For every completed producer disposition, the runner projects one bounded
+validation summary to the manifest, the affected rejection when present, and
+the CLI result receipt. The summary records status, configured-order rejecting
+validators and reason codes, incomplete validators, producer-attempt count,
+and terminal action. It contains no operator message, correction guidance, or
+model response. `complete`, `rejected`, and `incomplete` describe the final
+candidate disposition; a run-level `incomplete` status indicates at least one
+current-run output advanced under `warn_continue`.
+
 ## Checkpoint And Debug Hooks
 
 The runner receives checkpoint and debug interfaces rather than roots. It
@@ -151,10 +220,26 @@ handoff. Generated-reference dependencies participate in checkpoint decisions.
 Selective recomputation can require a canonical accepted normalized predecessor
 before a dependent lane starts.
 
+The runner writes successful checkpoint artifacts only after complete accepted
+validation. Chunk plans follow the same rule for publication. A rejection,
+invalid structured response, or incomplete validation is never reusable state;
+the current run may still hand off an otherwise valid `warn_continue` result
+according to its terminal policy. The runner carries private reuse eligibility
+through extract, merge, normalize, and generated-reference handoff. Any stage
+derived from incomplete validation skips both checkpoint lookup and all
+checkpoint publication even when that stage's own validation completes.
+External references and fully validated generated references remain eligible.
+Attempt debug records retain safe kind,
+validator, repair-usage, policy, and terminal-decision provenance. Full
+assistant and correction content remains confined to the requested detailed
+LLM trace.
+
 Debug recording is attempt-scoped and application-owned. A failure to persist
 required debug data is a framework error. State roots, persistence, reason-code
 meanings, resume, and cleanup are intentionally owned by
 [Run State Internals](state.md) and [Operations](../operations.md).
+Extract-validator trace scopes include the current chunk ordinal so concurrent
+chunks cannot overwrite one another's validator attempts or LLM artifacts.
 
 ## Invariants To Preserve
 

docs/internal/state.md

@@ -32,11 +32,18 @@ owned by the [Accepted Chunk Map contract](../integrations/chunk-map.md).
 
 `internal/framework/chunkplan` owns source-addressed plan storage, validation,
 and atomic publication. Its store is constructed only when the selected mode is
-not `bypass`.
+not `bypass`. It preserves rooted entry checks while creating new cache roots
+and source-digest directories with requested mode `0777`, then creates an
+atomic temporary file through rooted `O_CREATE|O_EXCL` with requested mode
+`0666` before sync and rename.
 
 `internal/framework/checkpoint` owns checkpoint identity, manifests, payload
 codecs, loader, and recorder. The CLI constructs a recorder whenever checkpoint
 recording is enabled and constructs a loader only for a `--resume` invocation.
+The recorder writes manifest and payload JSON through
+`fileio.WriteJSONWithProcessPermissions`, whose confined atomic writer creates
+new directories with requested mode `0777` and temporary files with requested
+mode `0666` before publication.
 Identity incorporates explicit stable semantic fingerprints collected from
 prepared modules and validators in addition to configuration, input,
 references, runtime overrides, observed LLM profiles, and the LLM runtime's
@@ -64,12 +71,13 @@ Ordinary resume loads extract, merge, and normalize checkpoints progressively
 and may execute later lane stages after an earlier cache miss. Selective
 recomputation instead asks the loader for the required producer's accepted
 normalize artifact. That lookup reuses the existing normalize files, requires
-workspace schema v3 plus an exact non-empty invocation identity, and deliberately
+workspace schema v4 plus an exact non-empty invocation identity, and deliberately
 does not require extract or merge checkpoint files or dependency fingerprints.
 The runner performs canonical codec and producer-provenance validation before
 cloning the artifact into normal step output. Success restores only stored
-normalize warnings and emits one normalize decision; failure retains the files,
-records the decision, and stops without executing the producer or consumer.
+normalize diagnostics and emits one normalize decision; failure retains the
+files, records the decision, and stops without executing the producer or
+consumer.
 
 The loader assigns a typed category and reason code at each validation site;
 diagnostic prose is not classified after the fact. The runner then applies
@@ -85,7 +93,10 @@ and debug-summary JSON.
 canonical operator-facing reason-code reference.
 
 `internal/core/fileio` provides confined atomic file writes used by state
-collaborators. The chunk-plan store retains its stronger entry validation.
+collaborators. Its exact-mode writer remains available to output and debug
+callers; its process-permission writer is used by checkpoints. The chunk-plan
+store retains its stronger rooted entry validation. [Operations](../operations.md#cache-provisioning)
+owns cache mode and provisioning guidance.
 
 The CLI constructs selective-recomputation policy from resolved generated
 artifact dependencies. It forces the selected step and transitive consumers,
@@ -97,7 +108,7 @@ owns the operator workflow and stable reason-code meanings.
 
 `internal/core/debugbundle` allocates an explicitly requested per-run bundle
 with `summary/` and `trace/` roots. `SummaryWriter` persists redacted command,
-resolution, run, warning, and failure artifacts. `internal/framework/debug`
+resolution, run, final grouped diagnostic, and failure artifacts. `internal/framework/debug`
 implements the pipeline-facing trace recorder under the trace root.
 
 The CLI allocates a bundle before pipeline resolution and treats requested

docs/operations.md

@@ -5,6 +5,23 @@ This is the canonical guide for operating Notarius runtime state. The
 [Configuration](config.md) owns fields, defaults, and precedence. Maintainers
 who need implementation mechanics should read [Run State Internals](internal/state.md).
 
+## Source Deployment
+
+Linux is the supported deployment platform. Install a pinned source release
+with the Go version declared in `go.mod` (currently Go 1.25.5):
+
+~~~sh
+GOWORK=off go install \
+  gitea.maximumdirect.net/eric/notarius/cmd/notarius@vMAJOR.MINOR.PATCH
+~~~
+
+Pin the exact tag in deployment automation rather than following a branch.
+Use [`notarius --version`](cli.md#command-summary) as a diagnostic after
+installation; its syntax and semantics are owned by the [CLI reference](cli.md).
+The maintainer publication process, including tag guards and verification,
+belongs to [Source Releases](release.md). macOS builds are best-effort for
+development, and Windows is unsupported.
+
 ## State Surfaces
 
 Each run can use independent roots with different retention and access-control
@@ -33,11 +50,45 @@ directory selects a per-user root:
 
 The field definitions and configuration examples are in [Configuration](config.md).
 On supported Unix systems, output directories and files are created with
-requested modes **0755** and **0644**. Chunk-plan, checkpoint, and debug
-directories and files use **0700** and **0600**. The operating system's umask
-may impose stricter output modes. Cache and debug roots may contain sensitive
-source-derived data, so provision them for one trusted account or service. An
-output bundle can also contain source content when its JSON output enables
+requested modes **0755** and **0644**; debug directories and files use
+**0700** and **0600**. The operating system's umask may impose stricter output
+or debug modes. Chunk-plan and checkpoint caches instead request maximum modes
+of **0777** for new directories and **0666** for new files, including atomic
+temporary files. The kernel applies the service's umask and applicable default
+ACLs when those entries are created.
+
+## Cache Provisioning
+
+Cache locations select where Notarius writes, not who can access them. A
+private service can use a restrictive umask:
+
+~~~sh
+umask 0077
+notarius run example --config /etc/notarius/config.yml --input /srv/input.txt
+~~~
+
+A trusted-group deployment can provision a setgid cache root and use a
+group-oriented umask (or an equivalent default ACL):
+
+~~~sh
+install -d -m 2770 -g notarius-operators /srv/notarius/chunk-plans /srv/notarius/checkpoints
+umask 0007
+notarius run example --config /etc/notarius/config.yml --input /srv/input.txt
+~~~
+
+Setgid and default ACLs control inherited group ownership and access policy;
+Notarius does not copy a parent directory's literal mode bits. Cache state is
+reconstructible but can contain source-derived data, so operators own its
+access and retention policy. Existing cache inodes retain their prior modes
+after an upgrade. Recreate a cache beneath correctly provisioned parents, or
+stop writers and perform a deliberate administrative migration when a
+whole-tree change is required.
+
+Cache publication remains atomic and stored records remain validated. Those
+guarantees do not add cross-process locking or transactional coordination for
+simultaneous writers to one checkpoint identity.
+
+Output bundles can also contain source content when JSON output enables
 evidence publication. Apply an appropriate umask and output-root access policy
 before enabling that option; the requested output modes alone may not be
 suitable for transcript-bearing bundles.
@@ -64,7 +115,7 @@ precedence rules are owned by [Configuration](config.md#promptkit-profiles).
 Use a profile source owned by the service account, keep it readable only by
 the intended operator, and supply provider credentials through the service
 environment—not in the Notarius configuration or profile YAML. The maintained
-[operator profile](../examples/profiles/dnd-extraction.yml) is secret-free and
+[LLM profile](../examples/llm-profiles/dnd-extraction.yml) is secret-free and
 can be copied as a format starting point. Validate a deployment without a
 provider call or credentials:
 
@@ -72,11 +123,14 @@ provider call or credentials:
 notarius config validate --config /etc/notarius/config.yml --pipeline dnd-session
 ~~~
 
-Profile paths are currently resolved from the process working directory, not
-from the configuration file. The complete example's
-`./examples/profiles/dnd-extraction.yml` path is valid for a repository-root
-invocation only. Use absolute paths such as
-`/etc/notarius/profiles/dnd-extraction.yml` for services and containers.
+An unset optional `api_key_env` reaches the provider without authorization and
+may receive a 401 or 403 response.
+
+Configured PromptKit profile paths resolve from the file that supplies their
+surviving declaration. The complete example's
+`./llm-profiles/dnd-extraction.yml` path is therefore relative to that example.
+Use absolute paths such as `/etc/notarius/profiles/dnd-extraction.yml` for
+services and containers when that is clearer for the deployment.
 
 ## Run Lifecycle
 
@@ -85,10 +139,40 @@ and resolves configuration before module preparation and source parsing. It
 then performs any permitted cache lookup, executes the pipeline, and publishes
 logical output files only after a successful runner result.
 
-On success, the command reports the output bundle path. A warning-bearing run
-still succeeds and reports its warning count on standard error. Errors and
+On success, the command reports the output bundle path. A run with actionable
+process warnings still succeeds and reports warning-group and occurrence counts
+on standard error; advisory and observation findings do not produce a warning
+line. Errors and
 their exit classes are defined in the [CLI reference](cli.md#output-streams-and-exit-statuses).
 
+## Validation Retries And Terminal Outcomes
+
+Each producer binding has one outer **retries** budget. It covers complete
+producer attempts for operational failures, invalid structured output,
+normalizer fallback retry, and semantic correction. It is independent from
+PromptKit's structural-repair calls inside one completion and from an
+LLM-backed validator's own retry budget. A semantic correction rebuilds the
+ordinary producer request and supplies only the latest rejected model response
+plus aggregated validator guidance; it is not a conversation replay.
+
+After the applicable budgets are exhausted, the resolved
+[`validation_policy`](config.md#pipelines) determines the result. Structural
+failure and semantic rejection normally fail the run; an explicit
+`reject_output` records a rejection and allows unrelated work to finish. A
+validator execution failure normally uses `warn_continue`, which keeps an
+otherwise accepted result in the current run with `incomplete` validation
+provenance. It emits one bounded warning for every validator whose execution
+budget was exhausted. A corrected result that later passes validation does not
+retain diagnostics from abandoned attempts.
+
+Treat a successful process exit as a completed run, not as proof that every
+candidate was fully validated. Inspect the receipt's `validation_status`,
+`validation_summaries`, rejection count, and warning group and occurrence
+counts when an orchestrator requires complete validation. The durable fields
+and their meanings are owned by the
+[run-result receipt](integrations/run-result.md) and
+[published JSON output contract](integrations/json-output.md).
+
 ## Output Bundles
 
 Each successful run receives a generated safe run identifier and writes beneath:
@@ -138,7 +222,9 @@ The configured cache mode controls one invocation:
 A reused plan is still materialized and validated against the current source.
 If a prior plan no longer gives acceptable results, use a refresh run rather
 than editing cache files. Deleting a plan is recoverable but can repeat costly
-chunking work.
+chunking work. A plan accepted only under incomplete validation is not
+published, and a rejected cache hit falls through to ordinary generation rather
+than becoming a correction candidate.
 
 ## Checkpoint Recording, Resume, And Recompute
 
@@ -163,6 +249,12 @@ Reasoning-effort inheritance, replacement, and explicit clearing are distinct
 runtime identities, so checkpoints created under one state are not reused by
 either of the others.
 
+Only accepted, completely validated chunk, extract, merge, and normalize
+results are checkpointed for reuse. Rejected, structurally invalid, and
+validation-incomplete producer results remain non-reusable, even when a
+`warn_continue` result advanced during its original run. A resumed invocation
+therefore reruns that producer rather than treating degraded state as accepted.
+
 Checkpoint state is confined below an identity-specific path:
 
 ~~~
@@ -216,14 +308,19 @@ Only a [debug-enabled run](cli.md#run) creates a bundle:
   trace/
 ~~~
 
-The summary contains redacted invocation and resolution information plus run,
-warning, checkpoint, chunk-plan, and terminal reporting artifacts. The trace
-contains allowlisted application diagnostic records and can include source or
-derived application data. Neither surface is a cache input. Do not treat a
-debug bundle as safe to share merely because its configuration summary is
-redacted. Invocation metadata omits reasoning effort when it is inherited,
-records the replacement value when one is supplied, and records an empty value
-when inherited reasoning was explicitly cleared.
+The summary contains redacted invocation and resolution information, including
+`effective-config.yml` and the separately redacted resolved pipeline, plus run,
+final grouped diagnostic, checkpoint, chunk-plan, and terminal reporting artifacts. Attempt
+terminal records contain bounded attempt kinds, validator outcomes, policy,
+decision, PromptKit repair count, and usage; they do not contain assistant
+responses or complete correction messages. The trace contains allowlisted
+application diagnostic records and can include source, model, and correction
+content. Neither surface is a cache input. Do not treat a debug bundle as safe
+to share merely because its configuration summary is redacted. Invocation
+metadata omits reasoning effort when it is inherited, records the replacement
+value when one is supplied, and records an empty value when inherited reasoning
+was explicitly cleared. It records `config_profile` only when a non-base
+configuration overlay was selected.
 
 Notarius never creates debug state without an explicit request and never
 automatically deletes a requested bundle. If allocation succeeds, the command
@@ -255,14 +352,29 @@ Provider execution settings and the generation timeout come from the selected
 PromptKit profile. The invocation-only **--reasoning-effort** and
 **--clear-reasoning-effort** controls may replace or clear that profile setting
 for all LLM-backed calls in one run without changing the profile. PromptKit
-v0.5.0 does not add a provider retry loop. Notarius binding retries rerun the
-complete module operation and validation chain as defined by
-[module bindings](config.md#module-bindings-and-validators).
+structural output repair happens within one structured-completion call. Its
+effective `structured_output_repair_attempts` limit is resolved from the
+selected binding, then the pipeline, then the prompt declaration; see
+[module bindings](config.md#module-bindings-and-validators). This is distinct
+from Notarius binding **retries**, which rerun the complete module operation
+and validation chain and do not consume or replenish the structural-repair
+limit. The maintained production prompts declare one repair attempt, paid only
+after a structural failure. One structured completion with repair budget **R**
+makes at most **R + 1** serial provider calls. If one stage attempt performs
+**C** structured completions, a binding with **retries: N** has a maximum of
+**(N + 1) * C * (R + 1)** provider calls; LLM-backed validators have their own
+corresponding invocation counts and budgets. This is an upper bound, not a
+promise that every call reaches a provider.
 
 Timeouts are layered. Caller cancellation is the outer authority. A positive
 effective generation timeout adds an inner request deadline, while zero
 disables only that generation deadline. The HTTP client timeout remains a
 transport-wide cap. Notarius does not add another timeout around PromptKit.
+Repairs are serial within the same caller context, so their worst-case latency
+and cost follow the provider-call bound above; provision run deadlines and
+provider budgets accordingly. Credentials remain optional unless the selected
+PromptKit profile requires one, in which case preparation fails before a
+provider call when its configured credential is unavailable.
 The pinned upstream boundary and profile-format links are in
 [PromptKit Integration](integrations/pkg-promptkit.md).
 
@@ -279,6 +391,11 @@ positive value makes the effective active local-generation bound the smaller
 of **total_llm** and that local limit, so a local limit of four permits no more
 than four active local generations.
 
+The Notarius scheduler admits one logical structured completion and holds that
+permit while PromptKit performs its serial corrective calls. PromptKit applies
+its selected-backend admission to each provider call; Notarius does not
+reacquire a permit or add another scheduler for a repair.
+
 For a positive local limit, PromptKit owns its default waiting capacity and
 admission behavior. When a PromptKit backend has admitted all active and queued
 work, a new call fails as capacity exhaustion before generation. The adapter

docs/policy/architecture.md

@@ -38,6 +38,36 @@ Notarius is contract-first without being abstraction-heavy. Interfaces and
 extension points should protect demonstrated boundaries. New abstraction is not
 itself an architectural goal.
 
+## Implementation Discipline
+
+Simplify the mechanism, not the user's choices.
+
+**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.
+
+**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. Existing
+explicit contracts remain authoritative; changing or weakening one requires a
+deliberate decision rather than being inferred from this principle.
+
+[ADR-0017](../adr/0017-defer-cache-access-policy-to-deployment.md) illustrates
+these principles. Notarius previously elevated private cache creation modes
+into application policy even though Unix deployments already express access
+policy through directory permissions, process umask, setgid inheritance, and
+default ACLs. The revised design retains application-owned correctness
+boundaries while leaving deployment security policy to the operator.
+
 ## Layers And Dependency Direction
 
 The application boundary is the composition root and may depend on concrete
@@ -146,8 +176,9 @@ lanes, validators, and LLM profile: the canonical source digest selects the
 plan, while the current run still applies its configured chunk validators to
 the materialized chunks.
 
-The framework owns orchestration and handoff provenance. Modules return logical
-results and warnings; they do not own CLI reporting, physical output, cache, or
+The framework owns orchestration, origin enrichment, aggregation, and handoff
+provenance. Modules return logical results and classified diagnostics; they do
+not own CLI reporting, physical output, cache, or
 debug roots, durable file placement, or checkpoint and debug lifecycle.
 
 After pipeline-wide chunking, extraction uses bounded framework concurrency.
@@ -159,26 +190,53 @@ may overlap. The framework must not create unbounded goroutines per lane or
 chunk.
 
 Completion timing does not choose public ordering or errors. The coordinator
-orders accepted artifacts, warnings, rejections, checkpoint events, and
+orders accepted artifacts, grouped diagnostics, rejections, checkpoint events, and
 framework errors by stable pipeline scope. Rejections do not cancel unrelated
 work. A framework error cancels derived work, prevents undispatched work from
 starting, waits for started work, and prevents output encoding.
 
+Warnings are process-only signals: configuration degradation, approved fallback,
+or incomplete configured validation. Quality uncertainty and grounding findings
+are advisories; successful canonicalization and cleanup are observations.
+Modules choose that semantic classification, while the framework attaches
+origin, aggregates groups, enforces bounds, and presents final collections.
+An ordinary successful run therefore has zero warnings. See
+[ADR-0015](../adr/0015-separate-process-warnings-from-quality-diagnostics.md)
+for the decision rationale.
+
 ## Validation
 
 Validation is a framework-managed boundary around outputs from chunk, extract,
-merge, and normalize stages. Validators receive immutable stage output
-and make an explicit whole-output decision: approve, approve with warnings, or
-reject.
+merge, and normalize stages. Validators receive immutable stage output and
+make an explicit whole-output decision: approve, reject, fail, or skip when a
+runtime prerequisite is unavailable.
 
 Typed artifact validators receive the domain value directly. Chunk validators
 receive source-zone chunks, while serialized validators receive immutable
 representation bytes and declared schema metadata. A validator registered for
 one target or artifact kind cannot satisfy an incompatible selection.
 
-Rejection is a recorded pipeline outcome, not a framework execution error.
-Validator execution failures are framework errors. Rejected output does not
-advance to the next stage.
+The framework runs every applicable validator sequentially in configured order.
+It aggregates rejections, exhausted validator failures, and skips before the
+producer policy chooses a disposition. A completed rejection never advances.
+With no rejection, an exhausted validator failure may fail the run or, under
+the configured `warn_continue` policy, advance a structurally valid candidate
+with explicit incomplete-validation provenance. Validators report findings;
+they do not choose candidate disposition.
+
+A completed rejection supplies a stable reason code for internal provenance
+and bounded actionable correction guidance for the candidate producer. Reason
+codes, validator keys, and operator-facing messages remain diagnostic data;
+they are not model instructions. The framework constructs model-facing retry
+text only from the semantic guidance and fails the contract rather than
+inventing or truncating missing guidance.
+
+An LLM-backed module may also request a feedback-aware retry when its own
+deterministic translation or typed safety policy rejects a structurally valid
+model response. It must supply model-facing guidance separately from its
+reason code, operator message, and fallback diagnostics, together with the
+exact `single_response_v1` candidate. A feedback-free module retry remains
+valid when no exact candidate exists.
 
 Default validator chains are production composition policy and are registered
 centrally by stage and module. Configuration may replace a stage-local default,
@@ -195,6 +253,21 @@ The caller of the LLM owns prompt selection, prompt inputs, response schema,
 and interpretation of structured output. Provider adapters do not own source-
 or domain-specific prompt logic.
 
+PromptKit owns bounded structural correction within one structured completion.
+Notarius owns outer stage attempts, semantic validation, and acceptance policy;
+the two budgets must remain separate.
+
+An LLM-backed producer can participate in semantic correction only when it
+declares `single_response_v1` and returns the exact one response that directly
+controlled its candidate. On an actionable validator rejection or
+feedback-capable module retry, the framework rebuilds the ordinary request and
+appends only the latest defective response as an `assistant` message plus one
+aggregated `user` correction message. This is a fresh replacement request, not
+a growing conversation. The retry budgets, terminal policy, and sensitive-data
+rationale are recorded in
+[ADR-0014](../adr/0014-feedback-aware-validation-retries.md) and
+[ADR-0016](../adr/0016-feedback-aware-module-requested-retries.md).
+
 When a model selects an application entity, callers must supply a contextual
 selection and deterministically attach the opaque application identity whenever
 the selection resolves exactly. Models do not receive or reproduce opaque
@@ -218,9 +291,11 @@ beneath this mandatory application-wide scheduler.
 
 ## Configuration And Provenance
 
-Configuration loading, precedence, defaults, environment overrides, redaction,
-and validation are centralized. Named pipeline definitions make structural
-composition explicit and discoverable. Operational overrides are permitted
+Configuration loading, precedence, defaults, redaction, and validation are
+centralized through the ConfigKit-backed boundary. `NOTARIUS_CONFIG` is limited
+to discovery; field-level environment overlays are not configuration input.
+Named pipeline definitions make structural composition explicit and discoverable.
+Explicit command overrides are permitted
 when they do not obscure the configured pipeline structure.
 
 Run preparation fails before stage execution when statically discoverable
@@ -229,7 +304,8 @@ invalid or incompatible.
 
 Run manifests record enough resolved pipeline, module, source, reference, and
 LLM provenance to make a run auditable after configuration changes. Manifests
-record identities and summaries rather than secret or large payload content.
+record identities and bounded validation summaries rather than secret, raw
+model, correction, or large payload content.
 
 ## State, Output, And Safety
 
@@ -249,18 +325,48 @@ an invocation that explicitly requests resume. Debug is never a cache input and
 is never created without an explicit request. Pipeline modules receive
 collaborator interfaces and never physical roots.
 
+Deployments own cache access policy. Cache writers request conventional maximum
+creation modes and let the process umask, setgid inheritance, and default ACLs
+determine effective access. Notarius owns the separate safety boundary:
+confined paths, symlink rejection, atomic publication, validation, and
+recovery. It does not chmod, chown, chgrp, or migrate cache entries to impose
+an application policy. This rule applies to chunk-plan and checkpoint cache
+only; output and debug retain their separate contracts. [Operations](../operations.md)
+owns deployment provisioning and retention guidance.
+
+Only accepted, completely validated producer output is reusable checkpoint or
+chunk-plan state. Rejected, structurally invalid, and validation-incomplete
+results cannot become cache or checkpoint inputs, even when a
+`warn_continue` result is allowed to advance in the current run. This
+ineligibility follows derived merge and normalize results and generated
+references for the remainder of the run: current-run handoff remains allowed,
+but no dependent cache or checkpoint may be loaded or published.
+
 Writes are atomic where practical. Paths for writes, moves, overwrites, and
 deletion must be narrow and explicit. Notarius never automatically deletes
 output or requested debug bundles; cache cleanup is explicit and recoverable.
 
 Secrets must not appear in errors, logs, output, cache, debug summaries,
-traces, manifests, documentation, examples, or redacted configuration. Debug
+traces, manifests, documentation, examples, or redacted configuration. Raw
+assistant responses and complete correction messages are attempt-local and are
+excluded from ordinary durable records and summaries; the requested detailed
+debug trace is the sole diagnostic surface allowed to retain them. Debug
 collection is allowlisted to application-owned payloads and must not capture
 unrelated process environment values or filesystem content. Trace data may
 contain application data and therefore inherits its sensitivity; operators own
 access controls and retention. Physical layout and operation are defined in
 [Operations](../operations.md).
 
+## Platform And Distribution
+
+Linux is the supported deployment platform. macOS is supported only as a
+best-effort development and compilation environment, while Windows is
+unsupported. Notarius distributes source releases only: an immutable source
+tag and its checked-in release note identify a release. The project does not
+publish executable binaries, archives, installers, container images,
+checksums, signatures, or package-manager entries. Maintainer release commands
+and tag guards belong to [Source Releases](../release.md).
+
 ## Architectural Non-Goals
 
 Notarius does not aim to provide:

docs/policy/documentation.md

@@ -30,6 +30,18 @@ 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.
 
+### Implementation Details And Policy
+
+Documentation must distinguish an explicit contract from an incidental
+implementation choice. Do not describe a restriction, guarantee, security
+policy, default, or limit as authoritative merely because the current code
+implements it. Documentation may accurately describe current behavior without
+making it a permanent requirement; distinguish descriptive behavior from
+normative policy. The
+[implementation-discipline rules](architecture.md#implementation-discipline)
+own the criteria for adopting product restrictions and for elevating behavior
+into repository policy.
+
 ### Audience And Detail
 
 Write for the document's stated audience and include only the detail needed for
@@ -65,6 +77,8 @@ secret values.
 | 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, resume, cleanup, permissions, recovery, and operational limits. | CLI flag syntax, configuration field definitions, logical output schemas, implementation mechanics. |
+| Source release procedure | `docs/release.md` | Maintainer release selection, candidate validation, tagging, publication guards, verification, and immutable-tag recovery. | Product installation summary, CLI version semantics, historical release summaries, CI implementation detail. |
+| 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. |
 | Public HTTP contract, if introduced | `docs/api.md` | Routes, authentication, media types, request and response schemas, status codes, pagination, caching, idempotency, rate limits, and HTTP retry semantics. | Client walkthroughs, upstream or downstream integration internals, implementation detail. |
 | Consumer guidance, if a public package or API is introduced | `docs/consumers/` | Task-oriented use of the public interface, minimal client examples, and consumer responsibilities. | HTTP wire semantics, external protocol contracts, internal implementation detail. |
 | External and durable integration contracts | `docs/integrations/` | 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. |
@@ -95,6 +109,14 @@ 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
+
+`docs/release.md` owns the source-release procedure. Release notes are
+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 note to defer
+or replace current documentation updates.
+
 ### Contracts And Implementation
 
 Integration and API documents define externally observable shapes and

docs/policy/testing.md

@@ -87,6 +87,13 @@ Do not force all behavior through oversized end-to-end tests. Do not test every
 
 A test should protect a decision, contract, or invariant—not memorialize the current implementation.
 
+Tests must not create a stronger product contract than the requirement they
+protect. In particular, do not turn an incidental limit, restriction,
+permission choice, validation rule, or implementation default into policy by
+asserting it without an expressly selected contract. Follow the
+[implementation-discipline rules](architecture.md#implementation-discipline)
+when deciding what behavior deserves durable protection.
+
 Before adding or retaining a test, ask:
 
 > What realistic defect would this test catch?

docs/release.md

@@ -0,0 +1,164 @@
+# Source Releases
+
+This procedure is for maintainers publishing Notarius source releases. A
+release is an immutable lightweight `vMAJOR.MINOR.PATCH` tag on `main` together
+with its checked-in `docs/releases/<tag>.md` note. Tag CI validates that source
+candidate after publication; it does not publish or repair a release.
+
+Notarius publishes no binaries, archives, checksums, signatures, containers,
+package-manager entries, or Gitea release objects. Windows is not supported.
+Do not create retrospective notes for the pre-procedure `v0.1.0`, `v0.2.0`, or
+`v0.3.0` tags.
+
+## Select And Describe The Release
+
+Choose an unused stable semantic version in the form `vMAJOR.MINOR.PATCH`.
+Prereleases are not supported. Before `v1.0.0`, a minor release may change a
+documented CLI, configuration, durable artifact, integration, or operating
+contract when its note explains the impact and required operator action. A
+patch release must not intentionally break those documented contracts within
+its minor line.
+
+Create the version-matched note as part of the candidate. Every new note uses
+this structure, with concise, truthful content in each section:
+
+```markdown
+# Notarius vMAJOR.MINOR.PATCH
+
+This release ...
+
+## Summary
+
+## Compatibility
+
+## Upgrade
+
+## Changes
+```
+
+The note is a historical summary. Link to current canonical documentation for
+exact behavior, and update that documentation in the candidate rather than
+using the note as a substitute.
+
+## Prepare The Candidate
+
+Set the selected release version and disable Go workspace use for every
+candidate command:
+
+```sh
+RELEASE_VERSION=vMAJOR.MINOR.PATCH
+export RELEASE_VERSION GOWORK=off
+```
+
+Run the shared source-candidate checks from the repository. They cover module
+hygiene, tests, race tests, vet, builds, formatting, whitespace, maintained
+configuration validation, and the Linux and Darwin command-build matrix:
+
+```sh
+./scripts/check-release-source.sh "$RELEASE_VERSION"
+```
+
+Before committing, manually follow every changed local Markdown link and
+review the candidate for unintended files, generated output, credentials, or
+other unrelated changes. Commit the release note and all affected current
+documentation, then run the shared checker against that exact candidate. Push
+the candidate commit to `main` only after it succeeds. Record the exact commit
+only after that push:
+
+```sh
+RELEASE_COMMIT=$(git rev-parse 'HEAD^{commit}')
+export RELEASE_COMMIT
+```
+
+For private-module installation, configure standard `GOPRIVATE` matching this
+module and ordinary Git authentication for the hosting service before running
+the verification below. The exact authentication mechanism belongs to the
+maintainer environment; never record credentials or environment dumps in a
+release note, command history, or repository file.
+
+## Guard And Publish The Tag
+
+Fetch current remote references, then run this guard without editing the
+candidate. It requires `main`, a clean worktree and index, disabled workspace
+use, a stable release version, the recorded and pushed commit, a matching note,
+and unused local and remote tags:
+
+```sh
+git fetch origin main --tags
+
+if ! printf '%s\n' "$RELEASE_VERSION" |
+  grep -E -x 'v(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)' >/dev/null
+then
+  printf '%s\n' "invalid release version: $RELEASE_VERSION" >&2
+  exit 1
+fi
+test "$GOWORK" = off
+test "$(git branch --show-current)" = main
+test -z "$(git status --porcelain)"
+test "$RELEASE_COMMIT" = "$(git rev-parse 'HEAD^{commit}')"
+test "$RELEASE_COMMIT" = "$(git rev-parse 'origin/main^{commit}')"
+test -s "docs/releases/$RELEASE_VERSION.md"
+grep -F -x "# Notarius $RELEASE_VERSION" "docs/releases/$RELEASE_VERSION.md"
+for heading in '## Summary' '## Compatibility' '## Upgrade' '## Changes'; do
+  grep -F -x "$heading" "docs/releases/$RELEASE_VERSION.md"
+done
+if git rev-parse -q --verify "refs/tags/$RELEASE_VERSION" >/dev/null; then
+  printf '%s\n' "local tag already exists: $RELEASE_VERSION" >&2
+  exit 1
+fi
+if git ls-remote --exit-code --tags origin "refs/tags/$RELEASE_VERSION" >/dev/null 2>&1; then
+  printf '%s\n' "remote tag already exists: $RELEASE_VERSION" >&2
+  exit 1
+fi
+```
+
+Create an explicitly lightweight tag against the guarded commit, verify its
+target, and push only that tag ref:
+
+```sh
+git -c tag.gpgSign=false tag "$RELEASE_VERSION" "$RELEASE_COMMIT"
+test "$(git cat-file -t "$RELEASE_VERSION")" = commit
+test "$(git rev-parse "$RELEASE_VERSION^{commit}")" = "$RELEASE_COMMIT"
+git push origin "refs/tags/$RELEASE_VERSION:refs/tags/$RELEASE_VERSION"
+```
+
+Never use `git push --tags`, move a published tag, or delete a published tag.
+
+## Verify The Published Release
+
+Confirm that the remote tag still points at the guarded commit and that the
+note is available from the tagged tree:
+
+```sh
+REMOTE_TAG_COMMIT=$(git ls-remote origin "refs/tags/$RELEASE_VERSION" | awk '{print $1}')
+test "$REMOTE_TAG_COMMIT" = "$RELEASE_COMMIT"
+git show "$RELEASE_VERSION:docs/releases/$RELEASE_VERSION.md" >/dev/null
+```
+
+Verify a fresh source installation and its diagnostic version. The temporary
+directory confines the installed command to this check:
+
+```sh
+release_verification_dir=$(mktemp -d)
+trap 'rm -rf "$release_verification_dir"' 0 HUP INT TERM
+mkdir -p "$release_verification_dir/bin"
+GOWORK=off GOBIN="$release_verification_dir/bin" go install \
+  "gitea.maximumdirect.net/eric/notarius/cmd/notarius@$RELEASE_VERSION"
+test "$("$release_verification_dir/bin/notarius" --version)" = "notarius $RELEASE_VERSION"
+```
+
+An exact fresh checkout and `GOWORK=off go build ./cmd/notarius` is an
+equivalent source verification when local installation policy requires it.
+`notarius --version` is diagnostic only; downstream compatibility remains
+defined by the published receipt and artifact contracts.
+
+## Failure And Correction Policy
+
+If candidate validation fails before publication, fix the candidate on `main`,
+rerun the shared checker, and repeat the guards. An unpublished local tag may
+be deleted after inspection.
+
+If the remote tag or tag CI reveals a defect, leave the published tag intact.
+Fix the defect on `main`, choose a new patch version, write a new matching
+note, and repeat this procedure. Do not weaken tag immutability or add release
+assets as a workaround.

docs/releases/v0.4.0.md

@@ -0,0 +1,83 @@
+# Notarius v0.4.0
+
+This release strengthens LLM reliability and validation throughout the
+configured pipeline, upgrades the PromptKit integration, and establishes the
+source-release and downstream-consumer workflows needed for broader D&D
+pipeline integration.
+
+## Summary
+
+Notarius now distinguishes PromptKit structural-output repair from
+application-owned semantic validation retries. Producer candidates can run
+through complete deterministic validator chains, receive bounded semantic
+correction guidance, and retry under explicit stage policies. Final run
+receipts and manifests preserve bounded validation provenance, while outputs
+that advance with incomplete validation remain available to the current run
+without entering reusable checkpoint state.
+
+The release also adds a maintained complete D&D subprocess-consumer workflow,
+diagnostic build versions, and the source-only release procedure used to
+publish this version.
+
+## Compatibility
+
+- Configuration files must use schema version 4. Version 3 is not decoded or
+  rewritten; rename the top-level `scriptorium` section to `promptkit` when
+  migrating. See [Configuration](../config.md#migrating-version-3-configuration).
+- PromptKit is pinned to v0.9.0. Operator profile files use PromptKit's v0.9.0
+  format and may use its profile-inheritance support. Notarius continues to
+  resolve operator profiles before embedded fallbacks.
+- Structural-output repair and semantic stage retries are separate bounded
+  mechanisms. Maintained production prompts request one structural repair by
+  default; explicit configuration can override the supported repair count.
+- Validation policy can now fail a run, reject an output, or permit an
+  otherwise valid candidate to advance with incomplete-validation provenance.
+  The application defaults are documented in
+  [Configuration](../config.md#pipelines).
+- The `notarius.run-result.v1` receipt remains at schema version 1 and adds
+  optional validation summaries plus a required validation-status field.
+  Consumers of this pre-release contract should follow the current
+  [run-result receipt](../integrations/run-result.md).
+- Existing D&D artifact schema identities remain unchanged. Validation and
+  producer-policy changes can nevertheless cause previously accepted weak
+  candidates to retry, reject, or fail instead.
+
+## Upgrade
+
+1. Migrate every Notarius configuration to version 4 and rename `scriptorium`
+   to `promptkit`.
+2. Review deployed PromptKit profiles against the pinned v0.9.0 profile format
+   and ensure their credential environment variables are available at run
+   time.
+3. Run `notarius config validate --config <path> --pipeline <id>` before the
+   first production invocation.
+4. Review `structured_output_repair_attempts`, producer retry counts, and
+   `validation_policy` wherever the deployment needs behavior different from
+   the documented defaults.
+5. Update subprocess consumers to inspect receipt `validation_status` and to
+   tolerate the optional bounded `validation_summaries` field. A consumer that
+   requires fully validated artifacts should require `approved`.
+
+## Changes
+
+- Upgraded PromptKit from v0.5.0 through v0.9.0 and adopted profile
+  inheritance, structured-output repair, typed error classification, and the
+  correction-aware completion protocol.
+- Added pipeline and binding configuration for structural repair and terminal
+  validation policy, with strict startup validation and effective-setting
+  provenance.
+- Added feedback-aware retries for chunking, extraction, merge, normalize, and
+  semantic reconciliation producers. Retry prompts contain the exact defective
+  response and actionable semantic correction guidance without exposing
+  internal reason codes or opaque entity identifiers.
+- Added complete validator-chain execution, validator retry handling, bounded
+  warnings, terminal dispositions, and durable validation summaries.
+- Prevented validation-incomplete artifacts and all derived lineage from
+  loading or publishing reusable checkpoints while preserving same-run
+  generated-reference handoff.
+- Tightened cached chunk-plan validation so only completely validated plans are
+  reused or replace stored plans.
+- Added a machine-readable subprocess receipt workflow and complete D&D
+  consumer documentation covering all maintained artifacts.
+- Added source-release checks, immutable lightweight-tag guidance, Linux and
+  Darwin build verification, and diagnostic `notarius --version` output.

docs/releases/v0.5.0.md

@@ -0,0 +1,85 @@
+# Notarius v0.5.0
+
+This release separates actionable process warnings from extraction-quality
+advisories and routine normalization observations, giving operators a quiet
+warning channel without discarding durable diagnostic detail.
+
+## Summary
+
+Notarius now carries one validated, origin-aware diagnostic contract from
+producers and validators through retries, reusable state, output publication,
+debug summaries, run receipts, and CLI presentation. Warnings are reserved for
+process degradation or incomplete configured work. Data-quality findings are
+advisories, and successful deterministic cleanup is recorded as observations.
+An ordinary successful run therefore reports zero warnings while retaining
+bounded diagnostic provenance for later review.
+
+The framework aggregates findings deterministically by their stable identity
+and complete pipeline origin, preserves exact occurrence counts, and retains
+bounded representative samples. Warning groups fail rather than truncate;
+advisory and observation representation is bounded with explicit truncation
+metadata and exact unrepresented-occurrence counts.
+
+## Compatibility
+
+- `warnings.json` now uses the incompatible grouped
+  `notarius.warnings.v2` envelope and contains process warnings only. Consumers
+  of the former flat warning payload must migrate to the current
+  [JSON output contract](../integrations/json-output.md).
+- The new `diagnostics.json` file uses `notarius.diagnostics.v1` and contains
+  advisory and observation groups. Production `index.json` files always expose
+  both `warnings_file` and `diagnostics_file`.
+- The machine-readable run receipt is now `notarius.run-result.v2`. It replaces
+  `warning_count` with exact warning group and occurrence counts and adds
+  advisory/observation group, occurrence, and truncation fields. See the
+  current [run-result receipt](../integrations/run-result.md).
+- Custom output modules must return their complete logical file set or an
+  error. The former `OutputResult.Warnings` field has been removed; an output
+  module cannot report a warning after serializing its output.
+- Reusable state now uses `notarius.workspace.v4` and chunk-plan records use
+  `notarius.chunk-plan.v3` so they can preserve structured diagnostics. Older
+  pre-release reusable state is not reused under these contracts; start with
+  clean state when deterministic continuity with an older workspace is not
+  required.
+- Validation acceptance, semantic retry budgets, rejection policy, and D&D
+  artifact schema identities are unchanged by this release.
+
+## Upgrade
+
+1. Update subprocess consumers to require `notarius.run-result.v2` and read
+   `warning_group_count`, `warning_occurrence_count`,
+   `diagnostic_group_count`, `diagnostic_occurrence_count`, and
+   `diagnostics_truncated`.
+2. Update output-bundle consumers to decode `notarius.warnings.v2`, discover
+   `diagnostics.json` through `index.json`, and treat diagnostics as review
+   information rather than process warnings.
+3. Update any custom output module for the removal of
+   `OutputResult.Warnings`; return an error when encoding cannot complete.
+4. Clear pre-release reusable state before the first upgraded production run
+   when deterministic continuity with an older workspace is not required.
+5. Run `notarius config validate --config <path> --pipeline <id>` and perform
+   one representative run before promoting the release in an automated
+   pipeline.
+
+## Changes
+
+- Added validated diagnostic dispositions, categories, origins, stable reason
+  codes, exact occurrence counts, and bounded representative samples.
+- Added deterministic run-level aggregation with separate limits for
+  actionable warning groups and advisory/observation groups.
+- Reclassified D&D source-relatedness and unresolved-identity findings as
+  data-quality advisories and routine normalization changes as observations.
+- Preserved structured diagnostics across producer retries, validation,
+  generated-reference handoff, checkpoints, chunk-plan reuse, and debug
+  summaries while discarding superseded-attempt findings.
+- Added grouped `warnings.json`, a new grouped `diagnostics.json`, and the
+  corresponding production index entries.
+- Upgraded the machine-readable run receipt and human CLI summary to report
+  exact warning and diagnostic counts without allowing advisory volume to
+  create warning output.
+- Removed post-encoding output warnings and hardened diagnostic validation,
+  overflow handling, aggregate memory bounds, and warning-file path
+  presentation.
+- Documented diagnostic ownership, classification, operator interpretation,
+  durable contracts, and architectural invariants in ADR-0015 and the
+  canonical CLI, operations, integration, and internal documentation.

docs/releases/v0.6.0.md

@@ -0,0 +1,84 @@
+# Notarius v0.6.0
+
+This release improves the reliability and ergonomics of unattended,
+subprocess-driven D&D extraction pipelines.
+
+## Summary
+
+Notarius now gives models more precise, semantically useful correction guidance
+when deterministic or model-backed validation rejects an otherwise structured
+candidate. Semantic reconciliation retries identify candidates with compact,
+request-local numbers, preserve valid candidates when a proposal cannot be
+repaired, and report bounded process warnings when reconciliation is
+incomplete. D&D extraction also canonicalizes safely resolvable reversed source
+ranges and applies more consistent schema, prompt, and validator constraints,
+improving successful operation with smaller models without weakening evidence
+validation.
+
+An optional LLM-backed combat-semantics validator can review whether scene
+descriptions correctly consolidate combat and apply the `combat` kind. For
+subprocess callers, command-line references now follow the pipeline-scoped
+configuration model: a shared reference can be supplied once and automatically
+reach every compatible selected target, while lane- and binding-specific forms
+remain available for exceptions.
+
+## Compatibility
+
+- The meaning of an unqualified `--reference slot=path` or
+  `--without-reference slot` selector has changed. It now applies to every
+  selected pipeline target that declares the slot instead of requiring exactly
+  one matching target. Review callers that relied on ambiguity rejection or a
+  unique implicit target; the current selector contract is documented in the
+  [CLI reference](../cli.md).
+- The stage-wide `merge.slot` CLI reference shorthand has been removed. Use
+  `lane.merge.slot` for an exact merge binding, `lane.slot` for all compatible
+  bindings in one lane, or an unqualified `slot` for pipeline scope.
+- Command-line file references cannot replace, remove, or coexist with a
+  generated artifact handoff for the same concrete target and slot. Resolution
+  reports the conflict so the caller can narrow or remove the selector.
+- The private LLM-facing semantic-reconciliation contract now uses contiguous
+  `candidate_number` values rather than application identities. This is not a
+  durable or operator-configurable contract and requires no operator action.
+- Durable D&D artifact schemas, the `notarius.run-result.v2` receipt,
+  `notarius.warnings.v2`, and `notarius.diagnostics.v1` are unchanged from
+  `v0.5.0`.
+
+## Upgrade
+
+1. Update subprocess commands to provide shared inputs once with
+   `--reference slot=path`, and convert any stage-wide `merge.slot` selector to
+   an explicit supported scope. The [subprocess consumer guide](../consumers/subprocess.md)
+   and [complete D&D consumer guide](../consumers/dnd-pipeline.md) show the
+   current invocation pattern.
+2. Review CLI reference overrides that overlap generated same-run references;
+   narrow or remove an external selector rather than attempting to replace the
+   generated handoff.
+3. Run `notarius config validate --config <path> --pipeline <id>` for every
+   deployed pipeline configuration.
+4. Optionally add
+   `extract/dnd/scene-descriptions/combat_semantics` to the scene-description
+   validator chain where the deployment wants LLM-backed combat-scene review;
+   see the [configuration reference](../config.md).
+5. Perform a representative run with the deployed model profile and inspect
+   the machine-readable receipt, warnings, and diagnostics before promotion.
+
+## Changes
+
+- Added feedback-aware, module-requested retry support for normalizers while
+  keeping retry budgets bounded and preserving attempt diagnostics.
+- Reworked semantic reconciliation around compact, contiguous candidate
+  numbers, explicit namespace guidance, complete proposal validation, and
+  model-facing retry feedback that omits opaque internal identities.
+- Preserved candidates separately when an exhausted semantic proposal cannot
+  be safely applied, with a bounded process warning describing the fallback.
+- Canonicalized safely resolvable reversed D&D source-reference endpoints
+  before deterministic coverage validation and clarified the shared evidence
+  prompt.
+- Improved D&D schemas, prompts, and validators with closed-value constraints,
+  contextual correction guidance, shared source-range coverage logic, and
+  clearer item-holder transition rules.
+- Added the optional D&D scene-description combat-semantics validator, shared
+  combat policy assets, evaluation fixtures, and retry-path coverage.
+- Made unqualified CLI references pipeline-scoped, added hierarchical lane and
+  exact-binding selectors, defined deterministic override precedence, and
+  documented first-class subprocess use.

docs/releases/v0.7.0.md

@@ -0,0 +1,88 @@
+# Notarius v0.7.0
+
+This release simplifies configuration management with ConfigKit and makes
+cache access permissions a deployment responsibility.
+
+## Summary
+
+Notarius adopts ConfigKit v0.3.0 for strict configuration loading, imports,
+configuration profiles, provenance, and redacted inspection. Operators can
+compose whole pipeline definitions from separate files and select development
+overlays without duplicating the base configuration.
+
+Configured relative filesystem paths now resolve against their declaring
+files, including imported definitions and profile replacements. The effective
+configuration and its explanation reflect the paths used at runtime.
+
+## Compatibility
+
+- Configuration version 5 replaces version 4; there is no compatibility loader.
+  Numeric `version: 5` is the canonical spelling; quoted `"5"` also works.
+- Module and validator bindings require objects with `module:`; external
+  references require objects with `path:`. Lane-level reference aliases are
+  removed; use binding-local references instead.
+- Strict ConfigKit YAML parsing rejects aliases, merge keys, and unsupported
+  scalar forms. See the [configuration reference](../config.md) for the
+  complete contract and migration guidance.
+- The seven operational environment overrides are removed:
+  `NOTARIUS_TOTAL_LLM_CONCURRENCY`, `NOTARIUS_STAGE_WORKERS_EXTRACT`,
+  `NOTARIUS_OUTPUT_DIR`, `NOTARIUS_CACHE_CHUNK_PLANS_MODE`,
+  `NOTARIUS_CACHE_CHUNK_PLANS_DIR`, `NOTARIUS_CACHE_CHECKPOINTS_DIR`, and
+  `NOTARIUS_DEBUG_DIR`. `NOTARIUS_CONFIG` still selects the configuration
+  file; PromptKit credential environment variables remain supported.
+- Configured relative reference, PromptKit profile, output, cache, and debug
+  paths use the canonical declaring file's directory. Built-in relative
+  defaults use the main configuration directory. CLI-relative paths still use
+  the invocation directory; absolute paths are unchanged.
+- Requested debug bundles now contain `effective-config.yml` instead of
+  `effective-config.json`. Inspection output is redacted and is not promised
+  to be reloadable configuration.
+- Chunk-plan and checkpoint cache writers respect deployment permissions
+  through ordinary creation modes, umask, setgid inheritance, and default ACLs;
+  they no longer impose private permissions. Existing entries are not chmodded
+  or migrated. Output and debug access contracts are unchanged.
+- Durable extraction artifact schemas and subprocess receipt formats are
+  unchanged by this release.
+
+## Upgrade
+
+1. Migrate deployed configurations to version 5 using the
+   [configuration migration guide](../config.md#version-5-migration). Expand
+   shorthand bindings and references, and move retired environment settings
+   into YAML or supported CLI overrides.
+2. Review relative paths, particularly output/cache/debug directories and
+   PromptKit profile sources. Use absolute paths when preserving an existing
+   deployment location is important.
+3. Review deployment access settings for cache directories. Provision umask,
+   groups, and ACLs as appropriate; Notarius does not choose the installation's
+   access policy. See [Operations](../operations.md).
+4. Use `config validate --config <path> --pipeline <id>` for each pipeline,
+   including selected configuration profiles. Inspect resolved settings with
+   `config effective` and their provenance with `config explain`; see the
+   [CLI reference](../cli.md).
+5. Update any debug-bundle tooling that consumes the former effective JSON
+   file. Subprocess callers may optionally select an overlay through
+   `--config-profile`; this is separate from PromptKit's `--llm-profile`.
+
+The maintained [minimal](../../examples/dnd-minimal.config.yml) and
+[complete](../../examples/dnd-complete.config.yml) D&D examples demonstrate
+self-contained configuration and composed pipelines with a development profile.
+Build from the source tag using the [source-release procedure](../release.md).
+
+## Changes
+
+- Replaced custom YAML parsing and environment-overlay machinery with a
+  ConfigKit-backed typed document boundary and detached runtime conversion.
+- Added imports, configuration profiles, source-aware normalization, safe
+  effective/explanation commands, and source-file legends for load errors.
+- Preserved explicit zeroes and empty validator overrides at supported stages,
+  rejected invalid reference branches before conversion, and retained
+  whitespace normalization for worker identifiers.
+- Corrected normalization recording for ordered steps and validator sequences,
+  with regression coverage for runtime values and provenance.
+- Accepted ConfigKit's integer representation at module-option boundaries
+  through checked shared conversion.
+- Updated examples, configuration and subprocess guidance, and debug rendering.
+- Reworked cache creation to respect deployment-selected permissions while
+  retaining confined paths, symlink rejection, atomic publication, validation,
+  and recovery.

docs/releases/v0.7.1.md

@@ -0,0 +1,44 @@
+# Notarius v0.7.1
+
+This release updates the prompt engine and its provider catalogs while retaining
+Notarius's application-owned D&D defaults and subprocess contracts.
+
+## Summary
+
+- Upgrade PromptKit from v0.9.0 to v0.10.0 and adopt its selected provider catalogs.
+- Refresh the profile-source checkpoint identity for the updated catalog.
+- Document profile ownership and upstream catalog compatibility.
+
+## Compatibility
+
+Notarius's CLI, configuration schema, durable artifact formats, and subprocess
+receipt are unchanged. The `dnd-extraction` fallback retains its explicit model
+and settings; the upgrade does not switch it to an upstream default profile.
+
+Upstream catalog selections require review: `deepseek-3-2`, `deepseek-4-flash`,
+and `deepseek-4-pro` are no longer present. Operators selecting or inheriting
+those profiles must supply a complete operator-owned definition or select an
+available profile. `deepseek-flash-latest` and `deepseek-pro-latest` are not
+equivalent aliases and need not preserve previous settings. See the
+[PromptKit integration contract](../integrations/pkg-promptkit.md#profile-sources-and-compatibility).
+
+Existing checkpoints with the previous profile-source identity are not reused.
+Source-keyed chunk plans and existing output artifacts are unaffected.
+
+## Upgrade
+
+Review selected LLM profile IDs and inherited bases before rebuilding from this
+tag. No Notarius configuration migration is required for the maintained examples
+or application-owned D&D defaults. For controlled model settings, use a complete
+[operator profile](../config.md#promptkit-profiles), rather than inheriting a
+volatile catalog entry. Follow the [source release procedure](../release.md)
+for source installation and verification.
+
+## Changes
+
+- PromptKit v0.10.0 preserves the public engine API and adds two opt-in maintained
+  profile names; Notarius continues using its existing adapter and defaults.
+- The dependency graph selects OpenRouter catalog v2.20260922.1 and Rakestrawhome
+  catalog v2.20260922.0.
+- Profile checkpoint fingerprinting and current integration documentation now
+  identify the updated engine.

docs/roadmap/archive/warning-signal-and-presentation-audit.md

@@ -0,0 +1,648 @@
+# Warning Signal And Presentation Audit
+
+## Executive Assessment
+
+Notarius warning execution is mechanically stronger than its operator-facing
+presentation. Terminal-attempt promotion, stable ordering after concurrent
+work, checkpoint replay, validation summaries, and debug retention are all
+substantially correct. The audit found no general duplicate-append defect in
+the extract, merge, or normalize handoffs and no leakage of abandoned-attempt
+warnings into a successful result.
+
+The warning channel itself is not coherent. One flat `contracts.Warning` type
+currently represents at least four materially different concepts:
+
+- actionable degradation or incomplete validation;
+- heuristic data-quality doubt;
+- successful but potentially reviewable fallback; and
+- routine canonicalization, ordering, and deduplication observations.
+
+That conflation is the primary reason successful runs produce a count that is
+large but operationally weak. The maintained complete D&D example demonstrates
+the problem without a live provider: an approved run with no rejected outputs
+published 12 warning records, all from three advisory relatedness checks. An
+operator separately reported a successful complete D&D run with 10 outputs,
+one rejection, and 85 warnings. The production bundle for that run was not
+available in this environment, so its reason-code distribution could not be
+measured.
+
+The current implementation also has four correctness or robustness gaps:
+
+1. warning records lose stage, step, lane, module, validator, and chunk
+   provenance when promoted, which makes safe aggregation and diagnosis
+   impossible from `warnings.json` alone;
+2. there is no framework-level validation or aggregate bound, and the NPC- and
+   spell-relatedness validators bypass the D&D warning limiter entirely;
+3. warnings returned by an output encoder are added after `warnings.json` has
+   already been encoded, so the receipt, stderr, debug bundle, and published
+   warning file can disagree; and
+4. a skipped validator contributes to `incomplete` validation but does not
+   receive the warning generated for an exhausted validator failure.
+
+The recommended end state is a structured diagnostic contract with explicit
+disposition, category, origin, occurrence count, and bounded samples. Warnings
+are reserved for process-level degradation or incompleteness. LLM-judged or
+deterministically inferred extraction-quality signals are advisories, never
+warnings, and routine normalization observations remain inspectable without
+being reported as top-level warnings. An ordinary successful run in which all
+configured work completes normally should therefore report zero warnings. This
+is an architectural and durable-contract change, not merely revised CLI prose.
+
+## Evidence And Limits
+
+The audit used:
+
+- a complete static search of production `contracts.Warning` constructors,
+  reason-code constants, result fields, and promotion sites under `internal/`;
+- call-path inspection through producer attempts, validators, lane
+  coordination, chunk-plan reuse, checkpoints, output encoding, debug output,
+  CLI presentation, and run-result construction;
+- the maintained complete and minimal D&D examples with offline fake LLMs;
+- focused deterministic tests for warning bounds, semantic-reconciliation
+  fallback, `warn_continue`, semantic retries, terminal rejection, concurrency
+  ordering, and checkpoint reuse; and
+- the operator-provided observation of an 85-warning complete D&D run.
+
+No provider-backed production run was attempted because this environment has
+no API key. Consequently, the audit can establish warning mechanics, possible
+multiplicity, synthetic volume, and obvious heuristic limitations, but cannot
+estimate production frequency or the real false-positive rate of individual
+D&D advisories. Those measurements are not required to choose the recommended
+architecture; they are required before strengthening any heuristic advisory
+into a rejection or setting a numerical production acceptance target.
+
+## Complete Warning-Producer Inventory
+
+### Framework And Generic Boundaries
+
+| Producer | Reason code | Trigger and consequence | Multiplicity and bound | Current surfaces and coverage |
+| --- | --- | --- | --- | --- |
+| Reference materialization in `internal/framework/pipeline/references.go` | `empty_reference` | A bound external reference is a valid, accepted media type but contains zero bytes. The prompt may receive materially incomplete context. | One per empty bound file; finite by configuration but no shared run-level cap. | Enters `RunInput.Warnings`; reference tests protect contextual scope. |
+| Producer-attempt policy in `internal/framework/pipeline/producer_attempts.go` | `validator_execution_incomplete` | An applicable validator exhausted its execution budget and `warn_continue` accepted the otherwise valid candidate. | One per failed validator on each terminal candidate. An extract chain can multiply this by chunks and lanes. There is no global cap. | Durable warning, receipt count, stderr, debug, and validation summary. `TestWarnContinueRecordsOneWarningForEachExhaustedValidator` covers failures. |
+| Chunk, extract, merge, normalize, and output module result contracts | Module-defined | A module may return arbitrary warnings with its successful candidate. | No contract validation, message limit, per-result cap, or global cap. Current production modules are inventoried below. | Terminal-attempt filtering and concurrency ordering are well tested. |
+| Production JSON output encoder | None | The encoder copies incoming warnings into `warnings.json`; it does not currently create warnings. | Same count as its input. | JSON encoder and assembled-pipeline tests compare the incoming run warnings with the published file. |
+| Output encoder result contract | Module-defined | Any output encoder may return warnings discovered during encoding. | Unbounded by contract. No production encoder currently exercises this capability. | Appended to final `RunOutput` only after logical files were encoded; this is the cross-surface defect described in AUD-WARN-004. |
+
+Input adapters and production mergers do not currently have independent
+warning producers. Chunk-plan and checkpoint decisions are structured manifest
+or debug provenance rather than warnings. Cancellation and hard persistence,
+reference, parsing, serialization, and provider failures remain errors.
+
+### D&D Extraction Gates
+
+| Producer | Reason code | Trigger and consequence | Multiplicity and bound |
+| --- | --- | --- | --- |
+| `dnd/combat-turns` extractor | `scene_classification_unavailable` | The chunk has no exact matching scene-description classification. The extractor returns an empty accepted result and skips the LLM, so combat-turn output may be incomplete. | At most one per chunk for this lane. |
+| `dnd/enemy-events` extractor | `scene_classification_unavailable` | The same missing or mismatched scene gate causes accepted empty enemy-event output. | At most one per chunk for this lane. |
+
+An exact non-combat classification produces an intentional empty result without
+a warning. An exact combat classification proceeds normally. The two producers
+share a code and operator consequence but use different messages; their module
+origins are not retained in the final warning record.
+
+### D&D Source-Relatedness Validators
+
+All ten relatedness validators are deterministic advisories: they approve the
+candidate and warn when contextual prose or an entity name is not lexically
+present in cited text. Shape and source-reference failures are deliberately
+left to blocking validators earlier in the chain. The same relatedness
+validator is registered in both the extract and normalize default chain for
+each artifact family in `internal/modules/dnd/register/chains.go`.
+
+| Artifact family | Warning reason | Per-record trigger | Local bound | Omission reason |
+| --- | --- | --- | --- | --- |
+| Combat turns | `combat_turn_not_near_source` | Actor token sequence absent | 20 per validator invocation | `combat_turn_relatedness_warnings_omitted` |
+| Enemy events | `enemy_event_not_near_source` | Subject token sequence absent | 20 | `enemy_event_relatedness_warnings_omitted` |
+| Item occurrences | `item_occurrence_source_unrelated` | Item name token sequence absent | 20 | `item_occurrence_relatedness_warnings_omitted` |
+| Item registry | `item_not_near_source` | Item name token sequence absent | 20 | `item_relatedness_warnings_omitted` |
+| Location occurrences | `location_occurrence_not_near_source` | Location name token sequence absent | 20 | `location_occurrence_relatedness_warnings_omitted` |
+| Location registry | `location_not_near_source` | Location name token sequence absent | 20 | `location_relatedness_warnings_omitted` |
+| NPC occurrences | `npc_occurrence_not_near_source` | NPC name token sequence absent | 20 | `npc_occurrence_relatedness_warnings_omitted` |
+| NPC registry | `npc_not_near_source` | NPC name token sequence absent | **Unbounded** | None |
+| Scene descriptions | `scene_description_not_near_source` | No significant title or summary token appears; up to two findings per scene | 20 | `scene_description_relatedness_warnings_omitted` |
+| Spells | `spell_not_near_source` | Spell-name token sequence absent | **Unbounded** | None |
+
+The eight limiter-generated omission records are presentation artifacts, not
+new source-relatedness conditions. They occupy a warning slot and make list
+length differ from the actual occurrence count.
+
+### D&D Normalizers
+
+Every production D&D normalizer bounds its returned warning slice to 20 through
+`internal/modules/dnd/shared/diagnostics`, including a final omission record
+when needed. Registry semantic retries reserve space for their fallback
+warning. The following table is complete by semantically distinct condition;
+codes listed together are parallel artifact-family variants.
+
+| Condition | Reason codes | Result impact | Current classification assessment |
+| --- | --- | --- | --- |
+| Display or field whitespace/name canonicalization | `npc_fields_normalized`, `item_fields_normalized`, `location_fields_normalized`, `spell_name_canonicalized`, `combat_actor_canonicalized`, `enemy_event_name_canonicalized`, `item_occurrence_name_canonicalized`, `location_occurrence_name_canonicalized`, `scene_description_prose_normalized` | Deterministic successful mutation. The item-occurrence code can also describe `from`/`to` whitespace, not only the item name. | Routine observation. |
+| Durable ID recomputation | `npc_id_recomputed`, `item_id_recomputed`, `location_id_recomputed` | Restores the deterministic name-derived ID. | Routine observation; invalid identity is separately rejected by default chains. |
+| Source-reference sorting or deduplication | `source_references_normalized` | Sorts and removes exact duplicate references while deliberately preserving invalid references for their validators. | Routine observation. Shared code is useful but ambiguous without producer origin. |
+| Canonical record ordering | `combat_turns_reordered`, `enemy_events_reordered`, `item_occurrences_reordered`, `location_occurrences_reordered`, `npc_occurrences_reordered`, `scene_description_order_normalized` | Deterministic order changes only. | Routine observation. |
+| Exact or approved semantic duplicate consolidation | `duplicate_npc_collapsed`, `duplicate_item_collapsed`, `duplicate_location_collapsed`, `duplicate_spell_cast_collapsed`, `duplicate_combat_turn_collapsed`, `duplicate_enemy_event_collapsed`, `duplicate_item_occurrence_collapsed`, `duplicate_location_occurrence_collapsed`, `duplicate_npc_occurrence_collapsed`, `scene_description_duplicate_collapsed` | Removes duplicate records and preserves or combines canonical evidence according to the artifact policy. Registry codes cover both exact and accepted semantic consolidation. | Durable normalization observation; not normally operator-actionable. |
+| Unresolved external membership | `spell_name_unresolved`, `item_occurrence_unknown_item_id`, `location_occurrence_unknown_location_id` | The value is preserved but is not grounded in the effective catalog or registry. Default chains normally reject the same condition before normalization; it remains reachable with validator overrides or defensive direct use. | Actionable data-quality warning. |
+| Unsafe currency consolidation proposal | `item_semantic_proposal_invalid` | The proposed group is rejected and all records are preserved because denominations or currency/non-currency members are incompatible. The same code is also used internally as a retry reason. | Advisory about model proposal quality; no accepted-data loss. The control and diagnostic meanings should be separated. |
+| Semantic reconciliation unavailable or exhausted | `npc_semantic_reconciliation_exhausted`, `item_semantic_reconciliation_exhausted`, `location_semantic_reconciliation_exhausted` | The safe deterministic result is accepted, but possible semantic duplicates remain. | Actionable fallback warning. |
+| Local warning truncation | `npc_normalization_warnings_omitted`, `item_normalization_warnings_omitted`, `location_normalization_warnings_omitted`, `spell_normalization_warnings_omitted`, `combat_turn_normalization_warnings_omitted`, `enemy_event_normalization_warnings_omitted`, `item_occurrence_normalization_warnings_omitted`, `location_occurrence_normalization_warnings_omitted`, `npc_occurrence_normalization_warnings_omitted`, `scene_description_normalization_warnings_omitted` | Reports that individual records were omitted from presentation. | Group metadata, not an independent warning. |
+
+No production D&D normalization warning exposes raw model responses,
+correction guidance, or provider errors. Most dynamic names are quoted and
+truncated by the shared helper. That local discipline is not enforced by the
+generic warning contract, and the spell relatedness message does not use the
+shared truncation helper.
+
+## Warning Propagation And Surface Map
+
+```text
+external-reference warnings -----------------------------+
+                                                           |
+module candidate warnings -> validation chain warnings     |
+             |                      |                       |
+             +---- producer-attempt terminal policy -------+
+                                  |                        |
+                   accepted / terminal rejection only      |
+                                  |                        |
+              chunk or lane result in canonical order      |
+                                  |                        |
+       checkpoint record/replay and ordered step merge     |
+                                  |                        |
+                            RunOutput.Warnings <------------+
+                                  |
+                   OutputRequest -> output encoder
+                                  |             |
+                         warnings.json     OutputResult.Warnings
+                                                |
+                              appended to final RunOutput only
+                                  |
+                 receipt, stderr, final debug warning summary
+```
+
+### Attempts And Validation
+
+- `runProducerAttempts` promotes only the terminal accepted or terminal
+  rejected candidate's module and completed-validator warnings. Operational,
+  structural, semantic, and module-directed attempts that are superseded are
+  retained in attempt debug artifacts but not in the final collection.
+- A module-directed semantic-reconciliation retry adds its fallback warning
+  only when no retry remains. Earlier attempt warnings are discarded.
+- On `warn_continue`, warnings from the otherwise accepted candidate and
+  completed approved or rejected validators are retained. One fixed,
+  non-sensitive `validator_execution_incomplete` warning is added for every
+  failed validator. Skipped validators affect the validation summary and final
+  `incomplete` status but do not receive such a warning.
+- A semantic terminal rejection retains only warnings from that rejected
+  attempt. Structural rejection after producer failure cannot retain a
+  candidate warning because no valid candidate result exists.
+
+These behaviors are protected by the producer-attempt, extract-handoff,
+rejection-warning, normalize-retry, and attempt-debug tests. They are the right
+foundation for the redesign and should not be replaced with early-exit or
+all-attempt accumulation.
+
+### Concurrency And Ordering
+
+Extract jobs are dispatched chunk-first and lane-second. Results are stored by
+chunk index, finalized in ascending chunk order, and lane continuations are
+merged into a slice indexed by configured lane order. Pipeline steps run in
+configured order. The resulting public order is therefore:
+
+1. pre-run reference warnings;
+2. chunk-stage warnings;
+3. step order;
+4. configured lane order within each step;
+5. chunk order within extract;
+6. merge warnings; then
+7. normalize warnings; followed by any output-result warnings.
+
+`TestRunnerBoundsExtractJobsAndStabilizesReverseCompletion` exercises warning
+order under reversed completion. No completion-order leak was found.
+
+### Chunk Plans And Checkpoints
+
+- A reusable chunk plan stores producer warnings only. Current validators run
+  again, and their current warnings are appended. Warnings from a cached plan
+  candidate that fails current validation are discarded before regeneration.
+- Accepted extract, merge, and normalize checkpoints store the terminal
+  warnings for their stage. Reuse loads and appends those warnings once at the
+  same logical handoff. Tests compare fresh and resumed warning collections and
+  preserve their order.
+- Validation-incomplete accepted outputs are not reusable, preventing a later
+  run from silently treating incomplete validation as complete.
+- Required accepted-normalize hydration replays that normalize checkpoint's
+  warnings; checkpoint decisions separately expose that reuse occurred.
+
+The recommended redesign should keep fresh and resumed logical diagnostics
+equivalent. Whether a result was reused belongs in checkpoint provenance, not
+in the diagnostic grouping key; adding a `reused` distinction would fragment
+groups and make equivalent runs present differently.
+
+### Terminal Surfaces
+
+| Surface | Current content | Audience | Audit result |
+| --- | --- | --- | --- |
+| `RunOutput.Warnings` | Flat final slice | Framework and CLI | Canonical in-memory list, but lacks origin and bounds. |
+| Published `warnings.json` | Object containing the warnings passed into the output encoder | Durable consumers | Exact for the production JSON encoder unless the encoder itself returns warnings. |
+| `index.json` | Path to `warnings.json` | Durable consumers | Stable discovery path; no separate diagnostic-detail path. |
+| Run-result v1 | `warning_count = len(final RunOutput.Warnings)` | Subprocess callers | Count only; no group/occurrence distinction. |
+| Human stderr | `run completed with N warning(s)` | Operators | Count only and no direct detail path. Successful exit remains zero. |
+| Manifest | Validation and rejection summaries, no warning collection | Durable provenance | Correctly avoids duplicating the flat list. |
+| Debug summary `warnings.json` | Raw final warning array | Operators/developers | Includes final output-result warnings and can therefore differ from published `warnings.json`. |
+| Debug run report | Final warning count | Operators/developers | Same final slice length as receipt and stderr. |
+| Attempt/stage debug | Candidate-local warning detail and origin in path/envelope | Forensics | Sufficient to diagnose provenance, but debug capture is optional and is not a durable consumer contract. |
+
+## Empirical Measurements
+
+### Offline And Synthetic Runs
+
+| Scenario | Result | What it establishes |
+| --- | --- | --- |
+| Maintained complete D&D config and transcript with the repository's offline fake LLM | Approved, 10 normalized outputs, 0 rejected outputs, 12 warnings; receipt, stderr, and published file all reported 12 | An ordinary structurally successful workflow can be noisy without fallback or incomplete validation. |
+| Same complete run, grouped after publication | Three reason codes, seven exact `(reason, scope, message)` tuples, maximum exact-tuple repetition of three | The flat count materially overstates distinct operator conditions. Scope resets within chunks and does not identify origin. |
+| Maintained focused scene-description workflow | Approved, one normalized output, 0 warnings | The warning channel can be quiet when synthetic model text is lexically grounded. |
+| Generic warning publication contract | One warning reaches successful stderr, durable output, and debug summary | The ordinary pre-output path is consistent. |
+| NPC semantic-reconciliation candidate-limit fallback | No LLM call, all records preserved, one exhaustion warning, total warnings no greater than 20 | Fallback is bounded and materially different from routine normalization. |
+| `warn_continue` with two failed validators and one skipped validator | Validation status contains all three incomplete validators; warning slice contains two execution-incomplete records | Current warning count does not describe all incomplete validation. |
+| Retrying extract candidate | Two producer attempts; only the accepted attempt's one warning is final | Retry does not amplify abandoned warnings. |
+| Terminal semantic rejection | Only the final rejected attempt's operation and validator warnings are final | Rejection diagnostics are retained without retaining superseded warnings. |
+| Fresh versus reused extract checkpoint | Warning collections are deeply equal | Checkpoint replay does not itself amplify warnings. |
+| Spell normalizer with 21 unresolved entries | 20 records: 19 samples plus one omission record saying two additional warnings were omitted | `warning_count` is neither exact occurrence count nor distinct-condition count. |
+
+The focused audit tests passed in `internal/cli`,
+`internal/framework/pipeline`, the NPC-registry and spell normalizers, and all
+D&D packages.
+
+### Bounded Sample Review
+
+The complete offline D&D run produced:
+
+| Reason | Count | Sample | Review |
+| --- | ---: | --- | --- |
+| `location_not_near_source` | 4 | `Moon Gate` was absent from cited text | Correctly identifies deliberately unsupported fake output. Three records shared the same exact tuple because chunk and stage origin were lost. |
+| `location_occurrence_not_near_source` | 4 | A `Moon Gate` visit was absent from cited text | Correctly identifies the same unsupported registry-driven occurrence, but repeats the same operator concern across extraction and normalization. |
+| `scene_description_not_near_source` | 4 | A title or summary had no significant exact token in cited text | Mixed value. Generic `session scene` prose is ungrounded, while `Arrival` versus transcript `arrive` illustrates an expected lexical false positive. |
+
+This fake workflow is an integration fixture, not a model-quality benchmark.
+It nonetheless proves that the checks carry useful evidence while being too
+imprecise and repetitive to serve as one-warning-per-record operator alerts.
+
+### Production Evidence Still Needed
+
+The reported 85-warning run establishes that high volume occurs in practice,
+but the following remain unknown:
+
+- dominant production reason codes and stage/lane sources;
+- unique group count versus repeated occurrence count;
+- false-positive rate for each relatedness family;
+- how much volume comes from normalization observations versus advisories;
+- whether fresh and resumed production runs remain equivalent; and
+- a defensible numerical acceptance target.
+
+If further data is worthwhile, the operator can supply the v1 receipt,
+`warnings.json`, and manifest validation summaries without supplying transcript
+or lane artifacts. An initial privacy-preserving report should group by reason
+code and normalized scope family, count exact repeated tuples, and omit message
+text. Reviewing heuristic precision requires a separately approved bounded
+sample with its cited source context.
+
+## Classification Of Current Conditions
+
+| Target disposition | Current families | Result impact | Operator action | Durable placement |
+| --- | --- | --- | --- | --- |
+| **Warning** | Empty reference; validator failure or skip accepted under `warn_continue`; unavailable required scene classification; exhausted semantic reconciliation | A configured process completed under an allowed degraded or incomplete policy rather than completing normally | Correct reference/configuration, inspect provider/validator, or rerun | Actionable `warnings.json`, receipt summary, stderr summary, debug |
+| **Advisory** | Source-relatedness heuristics; unresolved spell or registry membership; guarded invalid semantic proposal; any future LLM-judged uncertainty or extraction-quality signal | Uncertain data quality or poor model proposal, but accepted data is structurally valid and deterministic guards prevented unsafe mutation | Optional model/source review; no routine action for every record | Durable diagnostic detail and debug; never a top-level warning |
+| **Observation** | Whitespace/name/ID/source-reference canonicalization; canonical ordering; exact and approved semantic duplicate consolidation | Successful intended normalization | None under normal operation | Durable bounded normalization diagnostics or debug; no stderr warning |
+| **Not a diagnostic** | Rejection, invalid structure, cancellation, persistence error, provider failure under fail-run policy | Candidate or run did not complete according to policy | Inspect rejection/error and retry or correct input/configuration | Existing rejection, validation summary, error, and debug contracts |
+
+Exact and semantic duplicate consolidation should remain distinguishable in
+category or reason metadata even though both are observations. Semantic
+reconciliation exhaustion remains a warning because a capability was not
+applied; successful approved consolidation is an observation because it is the
+normalizer's intended work.
+
+## Ranked Findings
+
+### AUD-WARN-001 — The Flat Warning Type Destroys Signal Quality
+
+- **Priority:** High operator impact; high implementation leverage.
+- **Evidence:** `contracts.Warning` has only scope, reason, and message. Routine
+  normalizer changes, heuristic doubt, fallback, and incomplete validation all
+  enter the same slice and the same CLI count. The offline complete run's 12
+  records were all advisories; the operator observed 85 records in a successful
+  real run.
+- **Impact:** Operators cannot tell whether a warning requires a rerun, a
+  configuration repair, optional review, or no action. Repeated routine output
+  trains them to ignore the channel.
+- **Recommendation:** Replace the flat semantic contract with explicit
+  `warning`, `advisory`, and `observation` dispositions plus a small category
+  vocabulary. Do not infer disposition from message text or require every
+  downstream consumer to maintain a reason-code policy table.
+
+### AUD-WARN-002 — Warning Records Lose The Origin Needed For Diagnosis And Aggregation
+
+- **Priority:** High correctness and usability impact.
+- **Evidence:** The runner knows stage, step, lane, module, validator, chunk ID,
+  and chunk index at promotion time, but `terminalWarnings` flattens module and
+  validator records into `[]contracts.Warning`. Per-chunk scopes such as
+  `locations[0]` and `occurrences[0]` then repeat without identifying their
+  chunk or producer. `source_references_normalized` is intentionally shared
+  across families and is therefore especially ambiguous.
+- **Impact:** `warnings.json` cannot answer which stage or module produced a
+  record. Message- or scope-based deduplication would merge unrelated findings
+  or retain accidental duplicates.
+- **Recommendation:** Keep module findings free of framework context, then have
+  the framework add a structured origin envelope before promotion. Validator
+  findings must retain validator identity instead of passing through
+  `validationReport.Warnings()` as a flat slice.
+
+### AUD-WARN-003 — Warning Volume Is Not End-To-End Bounded Or Validated
+
+- **Priority:** High robustness impact; medium immediate likelihood.
+- **Evidence:** D&D's `LimitWarnings` caps most individual producers at 20, but
+  NPC- and spell-relatedness return one warning per record without the helper.
+  Every extract validator is invoked per chunk, all ten relatedness checks run
+  again after normalization, and there is no run-level collector. The generic
+  contract validates neither disposition nor reason/message size, UTF-8,
+  blankness, or total records.
+- **Impact:** Warning memory and output grow with chunks, lanes, configured
+  validators, and record counts. Local omission records lose exact occurrence
+  semantics while still incrementing `warning_count`.
+- **Recommendation:** Add a generic bounded diagnostic collector that preserves
+  exact occurrence counts and bounded samples. Validate all diagnostic fields
+  at the module/framework boundary. Immediately bring NPC and spell
+  relatedness under the existing cap if the full redesign is staged.
+
+### AUD-WARN-004 — Output Encoder Warnings Make Durable Surfaces Disagree
+
+- **Priority:** Medium current impact; high contract correctness risk.
+- **Evidence:** `Runner.Run` passes existing warnings to `encoder.Encode`, then
+  the production encoder serializes `warnings.json`. Only after encoding does
+  the runner append `OutputResult.Warnings`. The receipt, stderr, debug summary,
+  and debug run report see the final slice; the already-created published file
+  cannot. No production encoder currently returns a warning, so ordinary JSON
+  runs do not trigger the defect.
+- **Impact:** A valid output-module implementation can violate the documented
+  claim that `warning_count` describes the published warning collection.
+- **Recommendation:** Remove successful output warnings from the output-module
+  contract unless a demonstrated use case requires them; encoding failures
+  should be errors and optional encoder observations should be debug data. A
+  two-phase finalize API is the viable but more complex alternative.
+
+### AUD-WARN-005 — Validation Skips Are Incomplete But Not Warned
+
+- **Priority:** Medium operator/correctness impact.
+- **Evidence:** `firstIncompleteValidation` treats failed and skipped validators
+  alike, and validation summaries include both. `incompleteValidationWarnings`
+  emits records only for `validationFailed`. The focused test demonstrates
+  three incomplete validators but two warnings.
+- **Impact:** A successful run can have `validation_status: incomplete` while
+  its warning count understates or even omits the affected validators. A caller
+  that checks only warnings receives a weaker signal than the manifest and
+  receipt status.
+- **Recommendation:** Produce one aggregated incomplete-validation warning
+  group whose occurrences cover both failure and skip, while retaining typed
+  outcome and safe reason metadata in the validation summary. Do not expose
+  provider errors or arbitrary skip prose in model or operator messages.
+
+### AUD-WARN-006 — Relatedness Checks Are Useful But Repetitive And Lexically Weak
+
+- **Priority:** Medium operator impact; low acceptance-policy urgency.
+- **Evidence:** Every family runs the advisory in both extract and normalize
+  chains. The complete fixture contains exact repeated tuples, and the checks
+  rely on exact normalized token sequences or a minimal significant-token
+  overlap. Reason naming drifts between `*_not_near_source` and
+  `*_source_unrelated`.
+- **Impact:** The checks can catch unsupported entities, but aliases, pronouns,
+  inflection, and generic scene prose create predictable false positives. Flat
+  per-record presentation magnifies them.
+- **Recommendation:** Retain the validators and their stage-local execution,
+  but classify and aggregate them as advisories. Normalize reason-code naming
+  when the diagnostic contract changes. Do not strengthen them into rejection
+  rules without a human-reviewed production evaluation.
+
+### AUD-WARN-007 — `warning_count` Has No Stable Operational Meaning
+
+- **Priority:** High downstream-contract impact.
+- **Evidence:** The receipt and CLI use `len(output.Warnings)`. One list element
+  can be an omission summary representing several hidden occurrences; repeated
+  records can represent the same condition; and skipped validators can be
+  absent. A 21-occurrence spell test produces a list length of 20.
+- **Impact:** The value is neither an exact occurrence count nor a distinct
+  warning-group count. Consumers cannot set policy or present a trustworthy
+  summary from it.
+- **Recommendation:** Introduce explicit warning-group and warning-occurrence
+  counts in a versioned receipt. Do not silently redefine the v1 field.
+
+## Recommended Target Contract And Presentation Model
+
+### Diagnostic Model
+
+Use one validated internal diagnostic model with these concepts:
+
+- **disposition:** `warning`, `advisory`, or `observation`;
+- **category:** a small enum such as `configuration`, `degradation`,
+  `validation_incomplete`, `data_quality`, `fallback`, or `normalization`;
+- **reason code:** stable semantic identity owned by the producer;
+- **origin:** framework-added phase/stage, step ID, lane ID, module key,
+  validator name, chunk ID, and chunk index when applicable;
+- **occurrence count:** exact number of matching findings;
+- **samples:** a small deterministic list of bounded scope/message pairs; and
+- **omitted sample count:** `occurrence_count - len(samples)`, represented as
+  metadata rather than another diagnostic record.
+
+Errors and rejected outputs must not become diagnostic dispositions. A warning
+means that the run completed under policy despite a process-level degradation
+or incomplete configured operation. Advisory and observation dispositions can
+describe accepted artifact quality and transformation provenance, but no
+LLM-judged extraction-quality signal may be promoted to a warning. Validation
+status remains authoritative for approval, rejection, and incomplete
+validation.
+
+### Aggregation
+
+The framework runner should own aggregation after it enriches findings with
+origin and before public output construction. Modules and validators retain
+semantic ownership of disposition, category, reason, scope, and message; they
+must not own CLI or file presentation.
+
+The default stable key should be:
+
+```text
+disposition + category + reason_code
++ phase/stage + step_id + lane_id + module_key + validator_name
+```
+
+Chunk, record scope, and message text belong in samples and must not be part of
+the group key. This groups repeated per-chunk findings without merging the same
+code across distinct producers or pipeline locations. Group order should be
+the first occurrence in the runner's existing canonical order; sample order
+should follow the same order. A final canonical sort by the complete origin key
+is also viable, but completion timing must never choose either order.
+
+Aggregation must be incremental and bounded. Producers should use a shared
+collector that counts every occurrence while retaining only bounded samples;
+the framework then merges producer groups without reconstructing counts from
+omission prose. A global maximum group count is also required, with overflow
+represented by structured aggregate metadata and with actionable groups given
+priority over lower dispositions.
+
+### Durable Files
+
+Keep one canonical home for each class:
+
+- `warnings.json` should contain versioned, grouped actionable warnings only;
+- a new `diagnostics.json` should contain versioned advisory and observation
+  groups only, avoiding duplication of warning groups;
+- `index.json` should link both files;
+- `rejected.json` and manifest validation summaries should retain their current
+  separate responsibilities; and
+- debug bundles should retain candidate-attempt detail plus the final grouped
+  projections.
+
+This is preferable to keeping all detail in `warnings.json` and filtering only
+the CLI: downstream consumers would otherwise continue to receive a semantically
+mixed warning contract, and routine observations would still dominate the
+durable file.
+
+### CLI And Receipt
+
+For a successful human run with actionable warnings, print a concise summary
+such as:
+
+```text
+notarius: run completed with 2 warning groups (7 occurrences); details=/.../warnings.json
+```
+
+Advisories and observations should not produce the warning line. Their durable
+path remains discoverable through `index.json`; a concise non-warning count can
+be added to the ordinary success line only if operator testing shows value. An
+ordinary successful run with no process degradation should write nothing to
+the warning stream even when it publishes quality advisories or normalization
+observations.
+
+Create `notarius.run-result.v2` rather than redefining v1. It should expose at
+least:
+
+- `warning_group_count`;
+- `warning_occurrence_count`; and
+- `diagnostic_group_count` for non-warning durable groups.
+
+The receipt should continue to expose validation status, validation summaries,
+and rejected-output count independently. Process exit behavior should not
+change as part of warning presentation reform.
+
+### Checkpoint Semantics
+
+Store the structured terminal diagnostic groups with accepted checkpoints and
+replay them exactly once at their logical stage. Fresh and reused runs should
+produce the same public groups and counts. Checkpoint events and debug records,
+not diagnostic identity, should disclose whether computation was reused.
+
+### Output Encoder Boundary
+
+Prefer removing `OutputResult.Warnings`. A successful output encoder should
+either return the complete logical files or fail. If future encoders genuinely
+need to produce durable post-encoding warnings, introduce an explicit
+two-phase prepare/finalize contract so those warnings can be included in the
+same published collection. Do not retain the current self-inconsistent
+one-pass capability.
+
+## Resolution Of Required Design Questions
+
+| Question | Recommendation | Viable alternative and tradeoff |
+| --- | --- | --- |
+| Explicit severity/disposition or external reason mapping? | Put validated disposition and category in the contract. | A central reason-code registry avoids payload fields but makes new modules depend on a second synchronized policy table and leaves downstream meaning implicit. |
+| Keep all detail in `warnings.json` or separate it? | Separate grouped actionable warnings from grouped advisories/observations in `diagnostics.json`. | Keep the flat durable list and aggregate only CLI output; simpler migration, but it preserves the noisy downstream contract and ambiguous count. |
+| Who owns aggregation? | Framework runner/coordinator after origin enrichment. | Output module aggregation keeps framework types smaller but duplicates policy across encoders and cannot repair missing validator origin. |
+| Stable aggregation key? | Disposition, category, reason code, and full producer origin; exclude chunk/scope/message. | Explicit producer-supplied grouping keys offer flexibility but add another identity that can drift from reason codes. Message-template grouping is brittle and unsafe. |
+| Samples and omissions? | Exact occurrence count plus deterministic bounded samples and numeric omitted-sample count. | Omission warning records preserve the current representation but inflate group counts and require prose parsing. |
+| `warning_count` semantics? | Version receipt and replace ambiguity with group and occurrence counts. | Keep v1 count as published record length and add optional fields; compatible, but two competing warning counts remain easy to misuse. |
+| Which normalization changes remain warnings? | Only exhausted process fallback. Unresolved membership is a data-quality advisory; successful canonicalization, reordering, ID repair, source-ref dedupe, and duplicate consolidation are observations. | Treat unresolved membership or semantic duplicate consolidation as warnings because they affect grounding or cardinality; more conservative, but it violates the process-only warning rule and reports accepted artifact quality as an operational failure. |
+| Source-relatedness disposition? | Grouped advisory by default; preserve current approve behavior. | Retain warning disposition or make rejection configurable. Rejection requires production precision evidence; current lexical rules are not strong enough. |
+| Checkpoint-loaded warnings? | Present the same logical groups as fresh execution and use checkpoint events for reuse provenance. | Mark groups as replayed; aids forensics but fragments aggregation and makes semantically equivalent runs differ. |
+| ADR and schema versions? | Add an ADR and version the run receipt and diagnostic files. | Treat the work as CLI-only presentation and avoid an ADR; insufficient because module contracts, output files, checkpoint payloads, and downstream fields change. |
+
+## Compatibility, Documentation, And ADR Implications
+
+The target alters public and internal contracts enough to require a new ADR.
+It should record:
+
+- the distinction among warnings, advisories, observations, rejections, and
+  errors;
+- the invariant that warnings are process-level signals, LLM-judged extraction
+  quality is never a warning, and ordinary non-degraded success has zero
+  warnings;
+- module semantic ownership versus framework origin/aggregation ownership;
+- bounded group and sample semantics;
+- fresh/checkpoint equivalence; and
+- the output-encoder decision.
+
+Implementation should introduce `notarius.run-result.v2`. The grouped warning
+and diagnostic envelopes should each carry their own schema version. Because
+the content of `warnings.json` changes incompatibly from a flat array wrapper
+to groups, release notes and the published JSON integration contract must call
+out the migration. `index.json` gains the diagnostic file path.
+
+Canonical documentation updates belong in:
+
+- `docs/cli.md` for stderr presentation only;
+- `docs/operations.md` for operator review and debug workflow;
+- `docs/integrations/json-output.md` for warning and diagnostic file schemas;
+- `docs/integrations/run-result.md` for v2 fields and compatibility;
+- `docs/consumers/subprocess.md` and `docs/consumers/dnd-pipeline.md` for
+  downstream policy checks;
+- `docs/internal/pipeline.md` for promotion, aggregation, retry, and checkpoint
+  mechanics;
+- `docs/internal/modules.md` and `docs/internal/dnd.md` for producer rules and
+  the D&D classification matrix; and
+- `docs/policy/architecture.md` for the durable ownership invariant after the
+  ADR is accepted and implemented.
+
+No configuration knob is required for the first implementation. A fixed,
+well-documented taxonomy is easier to reason about than per-reason display
+overrides. Configurable escalation or suppression can be considered only after
+production review demonstrates a concrete operator need.
+
+## Test Coverage Assessment
+
+Existing coverage worth preserving includes:
+
+- accepted-attempt and terminal-rejection warning promotion;
+- validator failure retry exhaustion and `warn_continue`;
+- module semantic retry fallback;
+- deterministic warning order under concurrent lane completion;
+- chunk-plan invalidation and discarded-cache warning behavior;
+- fresh/checkpoint warning equivalence;
+- local D&D warning caps and safe dynamic-message quoting;
+- JSON warning-file publication; and
+- CLI stderr, debug, and receipt counts.
+
+Material gaps are:
+
+- no bound test for NPC- or spell-relatedness warnings;
+- no generic warning-field or result-size validation;
+- no test for output encoder warnings versus published `warnings.json`;
+- no operator-level aggregation or bounded-sample tests;
+- no fresh/resume test for grouped counts because groups do not yet exist; and
+- no production evaluation of advisory precision.
+
+Tests should protect the semantic relationships: exact occurrence counts,
+bounded samples, deterministic group order, actionable-only warning
+presentation, and cross-surface equality. They should not assert one exact
+warning count for every complete D&D run or treat message wording as a public
+API unless the wording itself enforces a security boundary.
+
+## Audit Conclusion
+
+The application is in a good position for warning reform. Its retry,
+validation, checkpoint, and concurrency mechanics provide reliable points at
+which to attach structured diagnostics. The most valuable change is not to
+suppress individual reason codes; it is to replace the semantically flat,
+origin-free collection with bounded typed groups and to reserve the word
+“warning” for conditions that merit operator attention.
+
+Provider-backed runs would improve prioritization and help tune the D&D
+advisories, but they are not necessary to conclude that routine normalization
+and heuristic doubt should not dominate stderr or the durable warning
+contract. They should be gathered before changing heuristic acceptance policy
+or adopting a numerical production warning-volume target.

docs/roadmap/future.md

@@ -5,13 +5,26 @@ configuration, operations, internal, and integration docs. This roadmap records
 future work only. Items are ordered roughly by current value and specificity,
 not as committed release dates.
 
+## Near-Term Validation And LLM Reliability
+
+PromptKit now owns structural output repair within one completion. Notarius
+owns stage candidates, validator chains, semantic rejection policy, bounded
+feedback-aware stage retries, validation provenance, and reusable-state
+eligibility, and the separation of actionable process warnings from quality
+diagnostics. The remaining near-term work applies those completed foundations
+to domain review and empirical evaluation.
+
+Near-term reliability work should now be selected from the concrete evaluation
+and extension opportunities below. The retry, validation, and subprocess
+foundations described above are implemented current behavior.
+
 ## Near-Term D&D Pipeline
 
 ### Evaluate Spell Extraction And Normalization
 
 - Evaluate ordinary extraction retries and the completed normalization path
-  against a human-reviewed transcript set before adding repair-aware retries or
-  an LLM-backed semantic validator.
+  against a human-reviewed transcript set before and after adopting the shared
+  PromptKit repair and Notarius validation-retry policies above.
 - Maintain a small set of human-reviewed transcripts and outputs for prompt,
   validator, and normalizer development. Treat model-quality review as an
   iterative human evaluation aid, not a deterministic correctness gate.
@@ -28,8 +41,7 @@ The implemented source-backed core and initial D&D registry adoption are
 described by [Module Internals](../internal/modules.md#semantic-reconciliation)
 and
 [D&D Module Internals](../internal/dnd.md#semantic-registry-reconciliation).
-The [Semantic Reconciliation Roadmap](semantic-reconciliation.md) retains the
-original feature scope; the sections below keep broader extensions deferred.
+The sections below keep broader extensions deferred.
 
 ### Large-Collection Semantic Reconciliation
 
@@ -153,7 +165,6 @@ section only after a concrete workflow, contract, and priority emerge.
 ### Distribution And Operations
 
 - Packaged release artifacts for alpha distribution.
-- A documented versioning and release process.
 - Optional generated example-output fixtures with a regeneration procedure.
 - Additional diagnostics or reporting views.
 

examples/dnd-complete.config.yml

@@ -1,6 +1,9 @@
-version: 4
+version: 5
+configkit:
+  imports:
+    - path: pipelines/dnd-session.yml
 promptkit:
-  profile_file: ./examples/profiles/dnd-extraction.yml
+  profile_file: ./llm-profiles/dnd-extraction.yml
 concurrency:
   total_llm: 2
   stage_workers:
@@ -16,145 +19,3 @@ cache:
     directory: ./notarius-cache/checkpoints
 debug:
   directory: ./notarius-debug
-pipelines:
-  dnd-session:
-    llm_profile: dnd-extraction
-    input: seriatim
-    # Stable campaign context is shared by every module that accepts these slots.
-    references:
-      party: ./dnd-party.txt
-      glossary: ./dnd-glossary.txt
-    chunk:
-      module: dnd/scenes
-      retries: 2
-    output:
-      module: json
-      options:
-        include_chunk_map: true
-        evidence_context:
-          enabled: true
-          window_units: 3
-          lanes:
-            - item-occurrences
-            - item-registry
-            - location-registry
-            - location-occurrences
-            - npc-registry
-            - spells
-            - combat-turns
-            - npc-occurrences
-            - enemy-events
-    steps:
-      # Establish session-wide reference artifacts before their consumers.
-      - id: describe-session
-        artifacts:
-          item-registry:
-            extract:
-              module: dnd/item-registry
-              retries: 2
-            merge: appendorder
-            normalize: dnd/item-registry
-          npc-registry:
-            extract:
-              module: dnd/npc-registry
-              retries: 2
-            merge: appendorder
-            normalize:
-              module: dnd/npc-registry
-              retries: 2
-          location-registry:
-            extract:
-              module: dnd/location-registry
-              retries: 2
-            merge: appendorder
-            normalize:
-              module: dnd/location-registry
-              retries: 2
-          scene-descriptions:
-            extract:
-              module: dnd/scene-descriptions
-              retries: 2
-            merge: appendorder
-            normalize: dnd/scene-descriptions
-      - id: extract-events
-        # Accepted registry artifacts and scene-description eligibility artifacts
-        # are supplied in memory to their compatible consumers in this step.
-        references:
-          location_registry:
-            artifact:
-              step: describe-session
-              lane: location-registry
-          npc_registry:
-            artifact:
-              step: describe-session
-              lane: npc-registry
-          scene_descriptions:
-            artifact:
-              step: describe-session
-              lane: scene-descriptions
-          item_registry:
-            artifact:
-              step: describe-session
-              lane: item-registry
-        artifacts:
-          item-occurrences:
-            extract:
-              module: dnd/item-occurrences
-              retries: 2
-            merge: appendorder
-            normalize: dnd/item-occurrences
-          spells:
-            extract:
-              module: dnd/spells
-              retries: 2
-              references:
-                spell_catalog: ./dnd-spell-catalog.json
-            merge: appendorder
-            # Stage-local file references are intentionally bound at each stage.
-            normalize:
-              module: dnd/spells
-              references:
-                spell_catalog: ./dnd-spell-catalog.json
-          combat-turns:
-            extract:
-              module: dnd/combat-turns
-              retries: 2
-            merge: appendorder
-            normalize: dnd/combat-turns
-          npc-occurrences:
-            extract:
-              module: dnd/npc-occurrences
-              retries: 2
-            merge: appendorder
-            normalize: dnd/npc-occurrences
-          location-occurrences:
-            extract:
-              module: dnd/location-occurrences
-              retries: 2
-            merge: appendorder
-            normalize: dnd/location-occurrences
-      - id: track-enemies
-        references:
-          npc_registry:
-            artifact:
-              step: describe-session
-              lane: npc-registry
-          scene_descriptions:
-            artifact:
-              step: describe-session
-              lane: scene-descriptions
-          combat_turns:
-            artifact:
-              step: extract-events
-              lane: combat-turns
-          npc_occurrences:
-            artifact:
-              step: extract-events
-              lane: npc-occurrences
-        artifacts:
-          enemy-events:
-            extract:
-              module: dnd/enemy-events
-              retries: 2
-            merge: appendorder
-            normalize: dnd/enemy-events

examples/dnd-minimal.config.yml

@@ -1,9 +1,12 @@
-version: 4
+version: 5
 pipelines:
   dnd-session:
     llm_profile: dnd-extraction
-    input: seriatim
+    input:
+      module: seriatim
     artifacts:
       spells:
-        extract: dnd/spells
-        normalize: dnd/spells
+        extract:
+          module: dnd/spells
+        normalize:
+          module: dnd/spells

examples/pipelines/dnd-session.yml

@@ -0,0 +1,164 @@
+pipelines:
+  dnd-session:
+    llm_profile: dnd-extraction
+    input:
+      module: seriatim
+    # Stable campaign context is shared by every module that accepts these slots.
+    references:
+      party:
+        path: ../dnd-party.txt
+      glossary:
+        path: ../dnd-glossary.txt
+    chunk:
+      module: dnd/scenes
+      retries: 2
+    output:
+      module: json
+      options:
+        include_chunk_map: true
+        evidence_context:
+          enabled: true
+          window_units: 3
+          lanes:
+            - item-occurrences
+            - item-registry
+            - location-registry
+            - location-occurrences
+            - npc-registry
+            - spells
+            - combat-turns
+            - npc-occurrences
+            - enemy-events
+    steps:
+      # Establish session-wide reference artifacts before their consumers.
+      - id: describe-session
+        artifacts:
+          item-registry:
+            extract:
+              module: dnd/item-registry
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/item-registry
+          npc-registry:
+            extract:
+              module: dnd/npc-registry
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/npc-registry
+              retries: 2
+          location-registry:
+            extract:
+              module: dnd/location-registry
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/location-registry
+              retries: 2
+          scene-descriptions:
+            extract:
+              module: dnd/scene-descriptions
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/scene-descriptions
+      - id: extract-events
+        # Accepted registry artifacts and scene-description eligibility artifacts
+        # are supplied in memory to their compatible consumers in this step.
+        references:
+          location_registry:
+            artifact:
+              step: describe-session
+              lane: location-registry
+          npc_registry:
+            artifact:
+              step: describe-session
+              lane: npc-registry
+          scene_descriptions:
+            artifact:
+              step: describe-session
+              lane: scene-descriptions
+          item_registry:
+            artifact:
+              step: describe-session
+              lane: item-registry
+        artifacts:
+          item-occurrences:
+            extract:
+              module: dnd/item-occurrences
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/item-occurrences
+          spells:
+            extract:
+              module: dnd/spells
+              retries: 2
+              references:
+                spell_catalog:
+                  path: ../dnd-spell-catalog.json
+            merge:
+              module: appendorder
+            # Stage-local file references are intentionally bound at each stage.
+            normalize:
+              module: dnd/spells
+              references:
+                spell_catalog:
+                  path: ../dnd-spell-catalog.json
+          combat-turns:
+            extract:
+              module: dnd/combat-turns
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/combat-turns
+          npc-occurrences:
+            extract:
+              module: dnd/npc-occurrences
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/npc-occurrences
+          location-occurrences:
+            extract:
+              module: dnd/location-occurrences
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/location-occurrences
+      - id: track-enemies
+        references:
+          npc_registry:
+            artifact:
+              step: describe-session
+              lane: npc-registry
+          scene_descriptions:
+            artifact:
+              step: describe-session
+              lane: scene-descriptions
+          combat_turns:
+            artifact:
+              step: extract-events
+              lane: combat-turns
+          npc_occurrences:
+            artifact:
+              step: extract-events
+              lane: npc-occurrences
+        artifacts:
+          enemy-events:
+            extract:
+              module: dnd/enemy-events
+              retries: 2
+            merge:
+              module: appendorder
+            normalize:
+              module: dnd/enemy-events

examples/profiles/development.yml

@@ -0,0 +1,8 @@
+# This selects operational settings only; choose a different LLM profile with
+# the pipeline's llm_profile field or the run command's --llm-profile flag.
+concurrency:
+  total_llm: 1
+  stage_workers:
+    extract: 1
+output:
+  directory: ./notarius-output-development

go.mod

@@ -3,9 +3,16 @@ module gitea.maximumdirect.net/eric/notarius
 go 1.25.5
 
 require (
-	gitea.maximumdirect.net/eric/promptkit v0.5.0
+	gitea.maximumdirect.net/eric/ConfigKit v0.3.0
+	gitea.maximumdirect.net/eric/promptkit v0.10.0
 	github.com/santhosh-tekuri/jsonschema/v6 v6.0.2
-	gopkg.in/yaml.v3 v3.0.1
+	gopkg.in/yaml.v3 v3.0.1 // indirect
 )
 
 require golang.org/x/text v0.40.0
+
+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
+	go.yaml.in/yaml/v4 v4.0.0-rc.6 // indirect
+)

go.sum

@@ -1,9 +1,17 @@
-gitea.maximumdirect.net/eric/promptkit v0.5.0 h1:jnpazLyyNhWrB2xzwwtUkNUfktkTdkENTwuSPnKiYrc=
-gitea.maximumdirect.net/eric/promptkit v0.5.0/go.mod h1:R95NM6fbMDGDC0/UomgnSBP6ui2ns+8SZb8bESNvrDQ=
+gitea.maximumdirect.net/eric/ConfigKit v0.3.0 h1:LgXEUjR0KsgwGBznOWa8FJTQ+iElctHtI1via5vKdu4=
+gitea.maximumdirect.net/eric/ConfigKit v0.3.0/go.mod h1:tYWpsrbLyu05r8+1MapWp5fN6ZrX9I8hU2xEPedYmYk=
+gitea.maximumdirect.net/eric/promptkit v0.10.0 h1:Nyi+YbqvO7tkRjen+Opc79wuzuKIIwcSfkriixPPwrE=
+gitea.maximumdirect.net/eric/promptkit v0.10.0/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/dlclark/regexp2 v1.11.0 h1:G/nrcoOa7ZXlpoa/91N3X7mM3r8eIlMBBJZvsz/mxKI=
 github.com/dlclark/regexp2 v1.11.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
 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 v0.40.0 h1:Ub2Z6/xjgF1WrYQz2nuITOEegKFtiIy+rieRJ5lHZKs=
 golang.org/x/text v0.40.0/go.mod h1:hpnzDAfGV753zIKo+wk3u1bVKCGPbrnF7+7LBF/UHVY=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=

internal/buildinfo/buildinfo.go

@@ -0,0 +1,36 @@
+// Package buildinfo resolves the product version embedded in a Notarius build.
+package buildinfo
+
+import (
+	"fmt"
+	"regexp"
+	"runtime/debug"
+)
+
+var stableVersion = regexp.MustCompile(`^v(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)$`)
+
+// Override is set at link time for controlled builds.
+var Override string
+
+// Version returns the release version embedded in the build, or development
+// when the build does not carry a stable release tag.
+func Version() (string, error) {
+	buildVersion := ""
+	if info, ok := debug.ReadBuildInfo(); ok {
+		buildVersion = info.Main.Version
+	}
+	return resolve(Override, buildVersion)
+}
+
+func resolve(override, buildVersion string) (string, error) {
+	if override != "" {
+		if !stableVersion.MatchString(override) {
+			return "", fmt.Errorf("build version override is not a stable release tag")
+		}
+		return override, nil
+	}
+	if stableVersion.MatchString(buildVersion) {
+		return buildVersion, nil
+	}
+	return "development", nil
+}

internal/buildinfo/buildinfo_test.go

@@ -0,0 +1,45 @@
+package buildinfo
+
+import "testing"
+
+func TestResolve(t *testing.T) {
+	tests := []struct {
+		name         string
+		override     string
+		buildVersion string
+		want         string
+		wantErr      bool
+	}{
+		{name: "stable main module version", buildVersion: "v1.2.3", want: "v1.2.3"},
+		{name: "zero version", buildVersion: "v0.0.0", want: "v0.0.0"},
+		{name: "override takes precedence", override: "v2.3.4", buildVersion: "v1.2.3", want: "v2.3.4"},
+		{name: "invalid override", override: "version", buildVersion: "v1.2.3", wantErr: true},
+		{name: "override with whitespace", override: " v1.2.3", wantErr: true},
+		{name: "leading zero major", buildVersion: "v01.2.3", want: "development"},
+		{name: "leading zero minor", buildVersion: "v1.02.3", want: "development"},
+		{name: "leading zero patch", buildVersion: "v1.2.03", want: "development"},
+		{name: "build version with whitespace", buildVersion: "v1.2.3 ", want: "development"},
+		{name: "prerelease", buildVersion: "v1.2.3-rc.1", want: "development"},
+		{name: "build suffix", buildVersion: "v1.2.3+build.1", want: "development"},
+		{name: "pseudo version", buildVersion: "v0.0.0-20260102030405-abcdef123456", want: "development"},
+		{name: "development build", buildVersion: "(devel)", want: "development"},
+		{name: "missing build information", want: "development"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got, err := resolve(tt.override, tt.buildVersion)
+			if tt.wantErr {
+				if err == nil {
+					t.Fatal("resolve() error = nil, want error")
+				}
+				return
+			}
+			if err != nil {
+				t.Fatalf("resolve() error = %v", err)
+			}
+			if got != tt.want {
+				t.Fatalf("resolve() = %q, want %q", got, tt.want)
+			}
+		})
+	}
+}

internal/cli/assembled_spell_pipeline_contract_test.go

@@ -21,6 +21,10 @@ import (
 
 const assembledSpellExtractorKey = "test/dnd/spell-casts"
 
+const assembledCorrectingSpellExtractorKey = "test/dnd/correcting-spell-casts"
+
+const assembledDirectSpellValidatorKey = "test/dnd/direct-spell-correction"
+
 func TestAssembledSpellPipelineNormalizesMergedCasts(t *testing.T) {
 	registries, resolved, extractor := assembledSpellPipeline(t, assembledSpellPipelineOptions{})
 	prepared, err := pipeline.Prepare(resolved, registries, pipeline.ModuleDependencies{})
@@ -67,28 +71,39 @@ func TestAssembledSpellPipelineNormalizesMergedCasts(t *testing.T) {
 		t.Fatalf("distinct cast = %#v, want separate evidence event", distinct)
 	}
 
-	wantWarningReasons := []string{
+	wantDiagnosticReasons := []string{
 		spellnormalize.ReasonCodeSpellNameCanonicalized,
 		spellnormalize.ReasonCodeSourceReferencesNormalized,
 		spellnormalize.ReasonCodeDuplicateSpellCastCollapsed,
 		"spell_not_near_source",
 	}
-	gotWarningReasons := make([]string, len(output.Warnings))
-	for index, warning := range output.Warnings {
-		gotWarningReasons[index] = warning.ReasonCode
+	gotDiagnosticReasons := make([]string, len(output.Diagnostics.Groups))
+	for index, group := range output.Diagnostics.Groups {
+		gotDiagnosticReasons[index] = group.ReasonCode
 	}
-	if !reflect.DeepEqual(gotWarningReasons, wantWarningReasons) {
-		t.Fatalf("warnings = %#v, want deterministic normalize and validation warnings", output.Warnings)
+	if !reflect.DeepEqual(gotDiagnosticReasons, wantDiagnosticReasons) {
+		t.Fatalf("diagnostics = %#v, want deterministic normalize and validation diagnostics", output.Diagnostics)
 	}
-	if output.Warnings[2].Scope != "spell_casts[0]" || !strings.Contains(output.Warnings[2].Message, "retained input index 0") || !strings.Contains(output.Warnings[2].Message, "removed input indices [1]") {
-		t.Fatalf("duplicate warning = %#v, want retained and removed merged indices", output.Warnings[2])
+	if output.Diagnostics.Groups[2].Samples[0].Scope != "spell_casts[0]" || !strings.Contains(output.Diagnostics.Groups[2].Samples[0].Message, "retained input index 0") || !strings.Contains(output.Diagnostics.Groups[2].Samples[0].Message, "removed input indices [1]") {
+		t.Fatalf("duplicate diagnostic = %#v, want retained and removed merged indices", output.Diagnostics.Groups[2])
 	}
 
 	warningsFile := decodeAssembledOutput[struct {
-		Warnings []contracts.Warning `json:"warnings"`
+		Groups []contracts.DiagnosticGroup `json:"groups"`
 	}](t, output.OutputFiles, "warnings.json")
-	if !reflect.DeepEqual(warningsFile.Warnings, output.Warnings) {
-		t.Fatalf("warnings file = %#v, run warnings = %#v, want manifest output path to preserve warnings", warningsFile.Warnings, output.Warnings)
+	if len(warningsFile.Groups) != 0 {
+		t.Fatalf("warnings file = %#v, want no process warnings for advisory-only diagnostics", warningsFile.Groups)
+	}
+	diagnosticsFile := decodeAssembledOutput[struct {
+		SchemaVersion                string                      `json:"schema_version"`
+		GroupCount                   int                         `json:"group_count"`
+		OccurrenceCount              int                         `json:"occurrence_count"`
+		Truncated                    bool                        `json:"truncated"`
+		UnrepresentedOccurrenceCount int                         `json:"unrepresented_occurrence_count"`
+		Groups                       []contracts.DiagnosticGroup `json:"groups"`
+	}](t, output.OutputFiles, "diagnostics.json")
+	if diagnosticsFile.SchemaVersion != "notarius.diagnostics.v1" || diagnosticsFile.GroupCount != len(output.Diagnostics.Groups) || !reflect.DeepEqual(diagnosticsFile.Groups, output.Diagnostics.Groups) || diagnosticsFile.OccurrenceCount != diagnosticGroupOccurrences(output.Diagnostics.Groups)+output.Diagnostics.UnrepresentedOccurrenceCount || diagnosticsFile.Truncated != output.Diagnostics.Truncated || diagnosticsFile.UnrepresentedOccurrenceCount != output.Diagnostics.UnrepresentedOccurrenceCount {
+		t.Fatalf("diagnostics file = %#v, run diagnostics = %#v", diagnosticsFile, output.Diagnostics)
 	}
 	manifest := decodeAssembledOutput[artifacts.RunManifest](t, output.OutputFiles, "manifest.json")
 	if len(manifest.ArtifactLanes) != 1 || manifest.ArtifactLanes[0].Normalizer != spellnormalize.Key {
@@ -101,6 +116,30 @@ func TestAssembledSpellPipelineNormalizesMergedCasts(t *testing.T) {
 	}
 }
 
+func TestAssembledSpellPipelineCorrectsRejectedDirectExtraction(t *testing.T) {
+	registries, resolved, extractor := assembledCorrectingSpellPipeline(t)
+	prepared, err := pipeline.Prepare(resolved, registries, pipeline.ModuleDependencies{})
+	if err != nil {
+		t.Fatalf("Prepare() error = %v, want nil", err)
+	}
+
+	output, err := pipeline.New().Run(context.Background(), pipeline.RunInput{
+		Prepared:       prepared,
+		RawInput:       readRepositoryFile(t, "examples", "seriatim-minimal-transcript.json"),
+		ChunkCacheMode: pipeline.ChunkCacheBypass,
+	})
+	if err != nil {
+		t.Fatalf("Run() error = %v, want nil", err)
+	}
+	if len(output.Rejected) != 0 || output.Manifest.ValidationStatus != "approved" || len(output.NormalizeOutputs) != 1 {
+		t.Fatalf("run output = %#v, want corrected accepted spell output", output)
+	}
+	correction := extractor.correctionSnapshot()
+	if correction == nil || string(correction.AssistantResponse) != `{"spell":"Mysterious Burst"}` || !strings.Contains(correction.UserGuidance, "use a known spell name") || !strings.Contains(correction.UserGuidance, "complete corrected replacement") || strings.Contains(correction.UserGuidance, "unknown_spell") || strings.Contains(correction.UserGuidance, "spell is not in the catalog") {
+		t.Fatalf("extract correction = %#v, want exact rejected model response and semantic replacement guidance only", correction)
+	}
+}
+
 func TestAssembledSpellPipelineHonorsNormalizeValidatorOverride(t *testing.T) {
 	registries, resolved, _ := assembledSpellPipeline(t, assembledSpellPipelineOptions{normalizeValidatorOverride: true})
 	var normalizeChain *pipeline.ResolvedValidatorChain
@@ -127,15 +166,16 @@ func TestAssembledSpellPipelineHonorsNormalizeValidatorOverride(t *testing.T) {
 	if err != nil || output.Manifest.ValidationStatus != "approved" || len(output.Rejected) != 0 || len(output.NormalizeOutputs) != 1 {
 		t.Fatalf("Run() error = %v output = %#v, want approved override run", err, output)
 	}
-	for _, warning := range output.Warnings {
-		if warning.ReasonCode == "spell_not_near_source" {
-			t.Fatalf("warnings = %#v, want explicit validator override to replace default relatedness chain", output.Warnings)
+	for _, group := range output.Diagnostics.Groups {
+		if group.ReasonCode == "spell_not_near_source" {
+			t.Fatalf("diagnostics = %#v, want explicit validator override to replace default relatedness chain", output.Diagnostics)
 		}
 	}
 }
 
-func TestAssembledSpellPipelinePromotesTerminalUnknownSpellWarning(t *testing.T) {
+func TestAssembledSpellPipelinePromotesTerminalUnknownSpellDiagnostics(t *testing.T) {
 	registries, resolved, _ := assembledSpellPipeline(t, assembledSpellPipelineOptions{unknownSpell: true})
+	resolved.Steps[0].ArtifactLanes[0].NormalizeValidationPolicy.SemanticRejection = pipeline.SemanticRejectionRejectOutput
 	prepared, err := pipeline.Prepare(resolved, registries, pipeline.ModuleDependencies{})
 	if err != nil {
 		t.Fatalf("Prepare() error = %v, want nil", err)
@@ -161,12 +201,12 @@ func TestAssembledSpellPipelinePromotesTerminalUnknownSpellWarning(t *testing.T)
 	if !reflect.DeepEqual(rejectedFile.Rejected, output.Rejected) {
 		t.Fatalf("rejected file = %#v, run rejections = %#v, want durable rejection diagnostic", rejectedFile.Rejected, output.Rejected)
 	}
-	if len(output.Warnings) != 1 || output.Warnings[0].ReasonCode != spellnormalize.ReasonCodeSpellNameUnresolved || output.Warnings[0].Scope != "spell_casts[0]" {
-		t.Fatalf("warnings = %#v, want terminal normalize catalog warning", output.Warnings)
+	if len(output.Diagnostics.Groups) != 2 || output.Diagnostics.Groups[0].ReasonCode != spellnormalize.ReasonCodeSpellNameUnresolved || output.Diagnostics.Groups[0].Samples[0].Scope != "spell_casts[0]" || output.Diagnostics.Groups[1].ReasonCode != "spell_not_near_source" {
+		t.Fatalf("diagnostics = %#v, want complete terminal normalize validation diagnostics", output.Diagnostics)
 	}
 }
 
-func TestAssembledSpellPipelinePromotesUnknownSpellWarningWhenOverrideAccepts(t *testing.T) {
+func TestAssembledSpellPipelinePromotesUnknownSpellAdvisoryWhenOverrideAccepts(t *testing.T) {
 	registries, resolved, _ := assembledSpellPipeline(t, assembledSpellPipelineOptions{normalizeValidatorOverride: true, unknownSpell: true})
 	prepared, err := pipeline.Prepare(resolved, registries, pipeline.ModuleDependencies{})
 	if err != nil {
@@ -190,14 +230,20 @@ func TestAssembledSpellPipelinePromotesUnknownSpellWarningWhenOverrideAccepts(t
 	if len(normalized.SpellCasts) != 1 || normalized.SpellCasts[0].Spell != "Mysterious Burst" {
 		t.Fatalf("normalized casts = %#v, want unresolved name preserved", normalized.SpellCasts)
 	}
-	if len(output.Warnings) != 1 || output.Warnings[0].ReasonCode != spellnormalize.ReasonCodeSpellNameUnresolved || output.Warnings[0].Scope != "spell_casts[0]" {
-		t.Fatalf("warnings = %#v, want promoted scoped unresolved-name warning", output.Warnings)
+	if len(output.Diagnostics.Groups) != 1 || output.Diagnostics.Groups[0].ReasonCode != spellnormalize.ReasonCodeSpellNameUnresolved || output.Diagnostics.Groups[0].Samples[0].Scope != "spell_casts[0]" {
+		t.Fatalf("diagnostics = %#v, want promoted scoped unresolved-name diagnostic", output.Diagnostics)
 	}
 	warningsFile := decodeAssembledOutput[struct {
-		Warnings []contracts.Warning `json:"warnings"`
+		Groups []contracts.DiagnosticGroup `json:"groups"`
 	}](t, output.OutputFiles, "warnings.json")
-	if !reflect.DeepEqual(warningsFile.Warnings, output.Warnings) {
-		t.Fatalf("warnings file = %#v, run warnings = %#v, want durable unresolved-name warning", warningsFile.Warnings, output.Warnings)
+	if len(warningsFile.Groups) != 0 {
+		t.Fatalf("warnings file = %#v, want no process warnings for an advisory diagnostic", warningsFile.Groups)
+	}
+	diagnosticsFile := decodeAssembledOutput[struct {
+		Groups []contracts.DiagnosticGroup `json:"groups"`
+	}](t, output.OutputFiles, "diagnostics.json")
+	if !reflect.DeepEqual(diagnosticsFile.Groups, output.Diagnostics.Groups) {
+		t.Fatalf("diagnostics file = %#v, run diagnostics = %#v", diagnosticsFile.Groups, output.Diagnostics)
 	}
 }
 
@@ -206,6 +252,47 @@ type assembledSpellPipelineOptions struct {
 	unknownSpell               bool
 }
 
+func assembledCorrectingSpellPipeline(t *testing.T) (pipeline.Registries, pipeline.ResolvedPipeline, *assembledCorrectingSpellExtractor) {
+	t.Helper()
+	components := productionTestComponents(t)
+	extractor := &assembledCorrectingSpellExtractor{}
+	if err := pipeline.RegisterExtractor[dnd.SpellList](components.registries.Extractors, pipeline.ModuleSpec{
+		Key:                assembledCorrectingSpellExtractorKey,
+		Stage:              pipeline.StageExtract,
+		ExecutionClass:     contracts.ExecutionClassLLMBacked,
+		CorrectionProtocol: contracts.CorrectionProtocolSingleResponseV1,
+		Requires:           []string{"chunks", "source.transcript"},
+		Provides:           []string{"dnd.spell_casts"},
+		ArtifactKind:       dnd.SpellListKind,
+	}, func() (contracts.Extractor[dnd.SpellList], error) {
+		return extractor, nil
+	}); err != nil {
+		t.Fatalf("register correcting extractor: %v", err)
+	}
+	if err := pipeline.RegisterTypedValidator[dnd.SpellList](components.registries.Validators, dnd.SpellListKind, pipeline.ValidatorSpec{Key: assembledDirectSpellValidatorKey, ExecutionClass: contracts.ExecutionClassDeterministic}, func() (contracts.TypedValidator[dnd.SpellList], error) {
+		return assembledDirectSpellValidator{}, nil
+	}); err != nil {
+		t.Fatalf("register direct spell validator: %v", err)
+	}
+
+	extract := pipeline.Binding(assembledCorrectingSpellExtractorKey)
+	extract.Retries = 1
+	extract.Validators = pipeline.ValidatorOverride{Set: true, Validators: []pipeline.ModuleBinding{{Module: assembledDirectSpellValidatorKey}}}
+	resolved, err := pipeline.ResolvePipeline(pipeline.PipelineProfile{
+		ID:    "assembled-dnd-correcting-spells",
+		Input: pipeline.Binding("seriatim"),
+		Chunk: pipeline.ModuleBinding{Module: "generic", Options: map[string]any{"max_units": 1}},
+		Artifacts: map[string]pipeline.ArtifactLaneProfile{
+			"spells": {Extract: extract, Normalize: pipeline.Binding(spellnormalize.Key)},
+		},
+		Output: pipeline.Binding("json"),
+	}, pipeline.ResolveOptions{}, catalogFromRegistries(components.registries))
+	if err != nil {
+		t.Fatalf("ResolvePipeline() error = %v, want nil", err)
+	}
+	return components.registries, resolved, extractor
+}
+
 func assembledSpellPipeline(t *testing.T, options assembledSpellPipelineOptions) (pipeline.Registries, pipeline.ResolvedPipeline, *assembledSpellExtractor) {
 	t.Helper()
 	components := productionTestComponents(t)
@@ -251,6 +338,69 @@ type assembledSpellExtractor struct {
 	unknownSpell bool
 }
 
+type assembledCorrectingSpellExtractor struct {
+	mu         sync.Mutex
+	correction *contracts.SemanticCorrection
+}
+
+func (*assembledCorrectingSpellExtractor) Key() string { return assembledCorrectingSpellExtractorKey }
+
+func (*assembledCorrectingSpellExtractor) ReferenceSlots() []contracts.ReferenceSlot { return nil }
+
+func (e *assembledCorrectingSpellExtractor) Extract(_ context.Context, req contracts.TypedExtractionRequest) (contracts.TypedExtractionResult[dnd.SpellList], error) {
+	if req.Source == nil || req.Chunk == nil {
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, fmt.Errorf("correcting assembled extractor requires source and chunk")
+	}
+	response := `{"spell":"accepted"}`
+	value := dnd.SpellList{SpellCasts: []dnd.SpellCast{}}
+	if req.Chunk.Index == 0 && req.Correction == nil {
+		response = `{"spell":"Mysterious Burst"}`
+		value.SpellCasts = []dnd.SpellCast{{Caster: "Aria", Spell: "Mysterious Burst", SourceRefs: []source.SourceRef{{SourceID: req.Source.ID, StartUnitID: 1, EndUnitID: 1}}}}
+	}
+	if req.Chunk.Index == 0 && req.Correction != nil {
+		correction, err := contracts.CloneSemanticCorrection(req.Correction)
+		if err != nil {
+			return contracts.TypedExtractionResult[dnd.SpellList]{}, err
+		}
+		e.mu.Lock()
+		e.correction = correction
+		e.mu.Unlock()
+		value.SpellCasts = []dnd.SpellCast{{Caster: "Aria", Spell: "Cure Wounds", SourceRefs: []source.SourceRef{{SourceID: req.Source.ID, StartUnitID: 1, EndUnitID: 1}}}}
+	}
+	candidate, err := contracts.NewModelCandidate([]byte(response), contracts.CorrectionProtocolSingleResponseV1)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, err
+	}
+	return contracts.TypedExtractionResult[dnd.SpellList]{Value: value, ModelCandidate: candidate}, nil
+}
+
+func (e *assembledCorrectingSpellExtractor) correctionSnapshot() *contracts.SemanticCorrection {
+	e.mu.Lock()
+	defer e.mu.Unlock()
+	correction, err := contracts.CloneSemanticCorrection(e.correction)
+	if err != nil {
+		return nil
+	}
+	return correction
+}
+
+type assembledDirectSpellValidator struct{}
+
+func (assembledDirectSpellValidator) Name() string { return assembledDirectSpellValidatorKey }
+
+func (assembledDirectSpellValidator) ExecutionClass() contracts.ExecutionClass {
+	return contracts.ExecutionClassDeterministic
+}
+
+func (assembledDirectSpellValidator) Validate(_ context.Context, req contracts.TypedValidationRequest[dnd.SpellList]) (contracts.ValidationResult, error) {
+	for _, cast := range req.Value.SpellCasts {
+		if cast.Spell == "Mysterious Burst" {
+			return contracts.ValidationResult{Approved: false, ReasonCode: "unknown_spell", Message: "spell is not in the catalog", CorrectionGuidance: "use a known spell name"}, nil
+		}
+	}
+	return contracts.ValidationResult{Approved: true}, nil
+}
+
 func (e *assembledSpellExtractor) Key() string { return assembledSpellExtractorKey }
 
 func (*assembledSpellExtractor) ReferenceSlots() []contracts.ReferenceSlot { return nil }
@@ -296,6 +446,14 @@ func (e *assembledSpellExtractor) chunkIndexesSnapshot() []int {
 	return append([]int(nil), e.chunkIndexes...)
 }
 
+func diagnosticGroupOccurrences(groups []contracts.DiagnosticGroup) int {
+	count := 0
+	for _, group := range groups {
+		count += group.OccurrenceCount
+	}
+	return count
+}
+
 func decodeAssembledOutput[T any](t *testing.T, files []contracts.OutputFile, name string) T {
 	t.Helper()
 	for _, file := range files {

internal/cli/cache_contract_test.go

@@ -23,7 +23,7 @@ func TestRunChunkPlanModePrecedenceAndValidation(t *testing.T) {
 	}{
 		{name: "default", wantStores: 1},
 		{name: "file", fileMode: "bypass"},
-		{name: "environment", envMode: "bypass"},
+		{name: "retired environment has no effect", envMode: "bypass", wantStores: 1},
 		{name: "cli", envMode: "refresh", cliMode: "bypass"},
 	}
 	for _, tt := range tests {
@@ -80,25 +80,14 @@ func TestRunChunkPlanModePrecedenceAndValidation(t *testing.T) {
 	})
 
 	for _, tt := range []struct {
-		name       string
-		fileConfig bool
+		name string
 	}{
-		{name: "invalid environment mode"},
-		{name: "invalid file mode", fileConfig: true},
+		{name: "invalid file mode"},
 	} {
 		t.Run(tt.name, func(t *testing.T) {
 			roots := newStateTestRoots(t)
 			opts := newStateTestHarness().options()
-			if tt.fileConfig {
-				replaceStateTestConfigLine(t, roots.config, "    mode: auto\n", "    mode: invalid\n")
-			} else {
-				opts.LookupEnv = func(name string) (string, bool) {
-					if name == "NOTARIUS_CACHE_CHUNK_PLANS_MODE" {
-						return "invalid", true
-					}
-					return "", false
-				}
-			}
+			replaceStateTestConfigLine(t, roots.config, "    mode: auto\n", "    mode: invalid\n")
 			var stdout, stderr bytes.Buffer
 			code := RunWithOptions([]string{"run", "sample", "--config", roots.config, "--input", roots.input}, &stdout, &stderr, opts)
 			if code != 1 || stdout.Len() != 0 || stderr.Len() == 0 {
@@ -207,16 +196,20 @@ func TestRunAutoReusesPlanWhenRunInputsChange(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	configText := replaceRequiredOnce(t, string(data), "    chunk: test/chunk\n", `    chunk:
+	configText := replaceRequiredOnce(t, string(data), "    chunk:\n      module: test/chunk\n", `    chunk:
       module: test/chunk
       options:
         strategy: first
 `)
-	configText = replaceRequiredOnce(t, configText, "    output: test/output\n", `      other:
-        extract: test/extract
-        merge: test/merge
-        normalize: test/normalize
-    output: test/output
+	configText = replaceRequiredOnce(t, configText, "    output:\n      module: test/output\n", `      other:
+        extract:
+          module: test/extract
+        merge:
+          module: test/merge
+        normalize:
+          module: test/normalize
+    output:
+      module: test/output
 `)
 	if err := os.WriteFile(roots.config, []byte(configText), 0o600); err != nil {
 		t.Fatal(err)

internal/cli/command_contract_test.go

@@ -8,6 +8,8 @@ import (
 	"path/filepath"
 	"strings"
 	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/buildinfo"
 )
 
 func TestCommandHelpSpellingsWriteUsageToStdout(t *testing.T) {
@@ -38,6 +40,7 @@ func TestCommandSyntaxErrorsUseStderrAndExitTwo(t *testing.T) {
 		{name: "unknown pipelines subcommand", args: []string{"pipelines", "unknown"}, want: "unknown pipelines subcommand"},
 		{name: "malformed run flag", args: []string{"run", "demo", "--chunk_cache", "invalid"}, want: "not supported"},
 		{name: "unknown flag", args: []string{"config", "validate", "--unknown"}, want: "flag provided but not defined"},
+		{name: "version arguments", args: []string{"--version", "extra"}, want: "--version does not accept arguments"},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
@@ -50,6 +53,33 @@ func TestCommandSyntaxErrorsUseStderrAndExitTwo(t *testing.T) {
 	}
 }
 
+func TestCommandVersionOutput(t *testing.T) {
+	previous := buildinfo.Override
+	t.Cleanup(func() { buildinfo.Override = previous })
+
+	tests := []struct {
+		name       string
+		override   string
+		wantCode   int
+		wantStdout string
+		wantStderr string
+	}{
+		{name: "development", wantStdout: "notarius development\n"},
+		{name: "release override", override: "v1.2.3", wantStdout: "notarius v1.2.3\n"},
+		{name: "invalid override", override: "release", wantCode: 1, wantStderr: "notarius: build version override is not a stable release tag\n"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			buildinfo.Override = tt.override
+			var stdout, stderr bytes.Buffer
+			code := RunWithOptions([]string{"--version"}, &stdout, &stderr, Options{})
+			if code != tt.wantCode || stdout.String() != tt.wantStdout || stderr.String() != tt.wantStderr {
+				t.Fatalf("code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
+			}
+		})
+	}
+}
+
 func TestConfigDiscoveryPrefersExplicitPathThenEnvironment(t *testing.T) {
 	explicit := writeCommandConfig(t, "explicit", "alpha")
 	environment := writeCommandConfig(t, "environment", "beta")
@@ -110,7 +140,7 @@ func TestConfigLoadingFailuresReturnOneWithPathContext(t *testing.T) {
 	stdout.Reset()
 	stderr.Reset()
 	code = RunWithOptions([]string{"config", "validate", "--config", malformed}, &stdout, &stderr, commandContractOptions(t))
-	if code != 1 || !strings.Contains(stderr.String(), malformed) || !strings.Contains(stderr.String(), "parse config file") || stdout.Len() != 0 {
+	if code != 1 || !strings.Contains(stderr.String(), "configuration contains invalid YAML") || stdout.Len() != 0 {
 		t.Fatalf("malformed config: code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
 	}
 }
@@ -155,14 +185,16 @@ func TestConfigValidateResolvesPipelineAndChecksSelection(t *testing.T) {
 }
 
 func TestConfigValidatePipelineDefaultProfileIsOffline(t *testing.T) {
-	configPath := writeCommandConfigContent(t, `version: 4
+	configPath := writeCommandConfigContent(t, `version: 5
 pipelines:
   demo:
     llm_profile: dnd-extraction
-    input: seriatim
+    input:
+      module: seriatim
     artifacts:
       spells:
-        extract: dnd/spells
+        extract:
+          module: dnd/spells
 `)
 	var stdout, stderr bytes.Buffer
 	code := RunWithOptions([]string{"config", "validate", "--config", configPath, "--pipeline", "demo"}, &stdout, &stderr, Options{})
@@ -231,19 +263,21 @@ func commandContractOptionsWithLookup(t *testing.T, lookup func(string) (string,
 
 func writeCommandConfig(t *testing.T, firstID, secondID string) string {
 	t.Helper()
-	content := fmt.Sprintf("version: 4\npipelines:\n  %q:\n    input: seriatim\n  %q:\n    input: seriatim\n", firstID, secondID)
+	content := fmt.Sprintf("version: 5\npipelines:\n  %q:\n    input:\n      module: seriatim\n  %q:\n    input:\n      module: seriatim\n", firstID, secondID)
 	return writeCommandConfigContent(t, content)
 }
 
 func writeResolvableCommandConfig(t *testing.T) string {
 	t.Helper()
-	return writeCommandConfigContent(t, `version: 4
+	return writeCommandConfigContent(t, `version: 5
 pipelines:
   demo:
-    input: seriatim
+    input:
+      module: seriatim
     artifacts:
       spells:
-        extract: dnd/spells
+        extract:
+          module: dnd/spells
 `)
 }
 

internal/cli/configkit_commands.go

@@ -0,0 +1,157 @@
+package cli
+
+import (
+	"context"
+	"encoding/json"
+	"flag"
+	"fmt"
+	"io"
+	"strings"
+
+	configkit "gitea.maximumdirect.net/eric/ConfigKit"
+	"gitea.maximumdirect.net/eric/notarius/internal/core/config"
+)
+
+func runConfigEffective(args []string, stdout, stderr io.Writer, opts Options) int {
+	fs := flag.NewFlagSet("config effective", flag.ContinueOnError)
+	fs.SetOutput(io.Discard)
+	configPath := fs.String("config", "", "config file path")
+	profile := fs.String("config-profile", "", "configuration profile")
+	if err := fs.Parse(args); err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 2
+	}
+	if fs.NArg() != 0 {
+		fmt.Fprintf(stderr, "notarius: unexpected argument %q\n", fs.Arg(0))
+		return 2
+	}
+	loaded, _, err := loadConfigKit(*configPath, *profile, nil, opts)
+	if err != nil {
+		writeConfigLoadError(stderr, err)
+		return 1
+	}
+	if err := loaded.WriteEffective(stdout); err != nil {
+		fmt.Fprintf(stderr, "notarius: write effective configuration: %v\n", err)
+		return 1
+	}
+	return 0
+}
+
+func runConfigExplain(args []string, stdout, stderr io.Writer, opts Options) int {
+	fs := flag.NewFlagSet("config explain", flag.ContinueOnError)
+	fs.SetOutput(io.Discard)
+	configPath := fs.String("config", "", "config file path")
+	profile := fs.String("config-profile", "", "configuration profile")
+	if err := fs.Parse(args); err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 2
+	}
+	if fs.NArg() != 1 {
+		fmt.Fprintln(stderr, "notarius: config explain requires exactly one JSON Pointer")
+		return 2
+	}
+	pointer := fs.Arg(0)
+	parsed, err := configkit.ParsePath(pointer)
+	if err != nil || parsed.String() != pointer {
+		fmt.Fprintf(stderr, "notarius: %q is not a canonical JSON Pointer\n", pointer)
+		return 2
+	}
+	loaded, _, err := loadConfigKit(*configPath, *profile, nil, opts)
+	if err != nil {
+		writeConfigLoadError(stderr, err)
+		return 1
+	}
+	if err := loaded.WriteExplanation(stdout, pointer); err != nil {
+		fmt.Fprintf(stderr, "notarius: write configuration explanation: %v\n", err)
+		return 1
+	}
+	return 0
+}
+
+func runConfigValidate(args []string, stdout, stderr io.Writer, opts Options) int {
+	fs := flag.NewFlagSet("config validate", flag.ContinueOnError)
+	fs.SetOutput(io.Discard)
+	configPath := fs.String("config", "", "config file path")
+	profile := fs.String("config-profile", "", "configuration profile")
+	pipelineID := fs.String("pipeline", "", "pipeline ID")
+	onlyRaw := fs.String("only", "", "comma-separated artifact lanes")
+	if err := fs.Parse(args); err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 2
+	}
+	if fs.NArg() != 0 {
+		fmt.Fprintf(stderr, "notarius: unexpected argument %q\n", fs.Arg(0))
+		return 2
+	}
+	if strings.TrimSpace(*onlyRaw) != "" && strings.TrimSpace(*pipelineID) == "" {
+		fmt.Fprintln(stderr, "notarius: --only requires --pipeline")
+		return 2
+	}
+	only, err := parseOnly(*onlyRaw)
+	if err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 2
+	}
+	loaded, path, err := loadConfigKit(*configPath, *profile, nil, opts)
+	if err != nil {
+		writeConfigLoadError(stderr, err)
+		return 1
+	}
+	if strings.TrimSpace(*pipelineID) == "" {
+		fmt.Fprintf(stdout, "config %q is valid\n", path)
+		return 0
+	}
+	catalog, err := effectiveCatalog(opts)
+	if err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 1
+	}
+	effective, err := loaded.Config.Resolve(config.ResolveInput{PipelineID: *pipelineID, Only: only, Catalog: catalog})
+	if err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 1
+	}
+	if err := validateExplicitPromptKitProfiles(context.Background(), effective.Config, effectiveLLMProfileIDs(effective.ResolvedPipeline), opts.promptKitAssets); err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 1
+	}
+	fmt.Fprintf(stdout, "config %q is valid for pipeline %q\n", path, strings.TrimSpace(*pipelineID))
+	return 0
+}
+
+func runPipelinesList(args []string, stdout, stderr io.Writer, opts Options) int {
+	fs := flag.NewFlagSet("pipelines list", flag.ContinueOnError)
+	fs.SetOutput(io.Discard)
+	configPath := fs.String("config", "", "config file path")
+	profile := fs.String("config-profile", "", "configuration profile")
+	jsonOutput := fs.Bool("json", false, "write JSON output")
+	if err := fs.Parse(args); err != nil {
+		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		return 2
+	}
+	if fs.NArg() != 0 {
+		fmt.Fprintf(stderr, "notarius: unexpected argument %q\n", fs.Arg(0))
+		return 2
+	}
+	loaded, _, err := loadConfigKit(*configPath, *profile, nil, opts)
+	if err != nil {
+		writeConfigLoadError(stderr, err)
+		return 1
+	}
+	ids := sortedPipelineIDs(loaded.Config)
+	if *jsonOutput {
+		payload, err := json.Marshal(struct {
+			Pipelines []string `json:"pipelines"`
+		}{Pipelines: ids})
+		if err != nil {
+			fmt.Fprintf(stderr, "notarius: marshal pipeline list: %v\n", err)
+			return 1
+		}
+		fmt.Fprintf(stdout, "%s\n", payload)
+		return 0
+	}
+	for _, id := range ids {
+		fmt.Fprintln(stdout, id)
+	}
+	return 0
+}

internal/cli/configkit_commands_test.go

@@ -0,0 +1,132 @@
+package cli
+
+import (
+	"bytes"
+	"context"
+	"errors"
+	"io"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/artifacts"
+	"gitea.maximumdirect.net/eric/notarius/internal/core/config"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+)
+
+func TestConfigKitReadOnlyCommandsLoadImportsAndProfiles(t *testing.T) {
+	t.Parallel()
+	root := t.TempDir()
+	path := filepath.Join(root, "notarius.yml")
+	writeConfigKitFile(t, filepath.Join(root, "fragment.yml"), "pipelines:\n  imported:\n    input: {module: input}\n")
+	writeConfigKitFile(t, filepath.Join(root, "profiles", "development.yml"), "pipelines:\n  profiled:\n    input: {module: input}\n")
+	writeConfigKitFile(t, path, "version: 5\nconfigkit:\n  imports:\n    - path: fragment.yml\npipelines:\n  base:\n    input: {module: input}\n")
+	var stdout, stderr bytes.Buffer
+	if code := RunWithOptions([]string{"pipelines", "list", "--config", path, "--config-profile", "development"}, &stdout, &stderr, commandContractOptions(t)); code != 0 || stdout.String() != "base\nimported\nprofiled\n" {
+		t.Fatalf("list = %d, %q, %q", code, stdout.String(), stderr.String())
+	}
+	stdout.Reset()
+	stderr.Reset()
+	if code := RunWithOptions([]string{"config", "validate", "--config", path, "--config-profile", "development"}, &stdout, &stderr, commandContractOptions(t)); code != 0 || !strings.Contains(stdout.String(), "is valid") {
+		t.Fatalf("validate = %d, %q, %q", code, stdout.String(), stderr.String())
+	}
+}
+
+func TestConfigKitReadOnlyCommandsRejectVersionFour(t *testing.T) {
+	t.Parallel()
+	path := filepath.Join(t.TempDir(), "notarius.yml")
+	writeConfigKitFile(t, path, "version: 4\n")
+	var stdout, stderr bytes.Buffer
+	if code := runPipelinesList([]string{"--config", path}, &stdout, &stderr, commandContractOptions(t)); code != 1 || !strings.Contains(stderr.String(), "version") {
+		t.Fatalf("list = %d, %q", code, stderr.String())
+	}
+}
+
+func TestConfigurationInspectionCommandsRenderSafeConfigKitViews(t *testing.T) {
+	t.Parallel()
+	root := t.TempDir()
+	path := filepath.Join(root, "notarius.yml")
+	writeConfigKitFile(t, path, `version: 5
+concurrency:
+  total_llm: 4
+pipelines:
+  "0":
+    input: {module: input}
+`)
+	writeConfigKitFile(t, filepath.Join(root, "profiles", "development.yml"), "concurrency:\n  total_llm: 3\n")
+	writeConfigKitFile(t, filepath.Join(root, "profiles", "retired.yml"), "pipelines: null\n")
+
+	providerCalled := false
+	opts := commandContractOptions(t)
+	opts.LLMClientFactory = func(context.Context, config.Config, string, LLMRuntimeOverrides) (contracts.StructuredLLMClient, []artifacts.LLMProfileManifest, error) {
+		providerCalled = true
+		return nil, nil, errors.New("LLM client must not be constructed")
+	}
+
+	var stdout, stderr bytes.Buffer
+	if code := RunWithOptions([]string{"config", "effective", "--config", path, "--config-profile", "development"}, &stdout, &stderr, opts); code != 0 || !strings.Contains(stdout.String(), "total_llm: 3") || stderr.Len() != 0 {
+		t.Fatalf("effective = %d, %q, %q", code, stdout.String(), stderr.String())
+	}
+	if providerCalled {
+		t.Fatal("config effective constructed an LLM client")
+	}
+
+	stdout.Reset()
+	stderr.Reset()
+	if code := RunWithOptions([]string{"config", "explain", "--config", path, "--config-profile", "development", "/concurrency/total_llm"}, &stdout, &stderr, opts); code != 0 || !strings.Contains(stdout.String(), `path: "/concurrency/total_llm"`) || stderr.Len() != 0 {
+		t.Fatalf("explain effective = %d, %q, %q", code, stdout.String(), stderr.String())
+	}
+
+	stdout.Reset()
+	stderr.Reset()
+	if code := RunWithOptions([]string{"config", "explain", "--config", path, ""}, &stdout, &stderr, opts); code != 0 || !strings.Contains(stdout.String(), `path: ""`) || stderr.Len() != 0 {
+		t.Fatalf("explain root = %d, %q, %q", code, stdout.String(), stderr.String())
+	}
+
+	stdout.Reset()
+	stderr.Reset()
+	if code := RunWithOptions([]string{"config", "explain", "--config", path, "--config-profile", "retired", "/pipelines/0"}, &stdout, &stderr, opts); code != 0 || !strings.Contains(stdout.String(), `path: "/pipelines/0"`) || stderr.Len() != 0 {
+		t.Fatalf("explain removed mapping key = %d, %q, %q", code, stdout.String(), stderr.String())
+	}
+}
+
+func TestConfigExplainRejectsInvalidAndUnknownPointers(t *testing.T) {
+	t.Parallel()
+	path := writeCommandConfig(t, "session", "other")
+	for _, test := range []struct {
+		name     string
+		pointer  string
+		wantCode int
+	}{
+		{name: "missing", wantCode: 2},
+		{name: "malformed escape", pointer: "/pipelines~2session", wantCode: 2},
+		{name: "unknown", pointer: "/unknown", wantCode: 1},
+		{name: "invalid sequence index", pointer: "/pipelines/session/steps/+0", wantCode: 1},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			var stdout, stderr bytes.Buffer
+			args := []string{"config", "explain", "--config", path}
+			if test.name != "missing" {
+				args = append(args, test.pointer)
+			}
+			if code := RunWithOptions(args, &stdout, &stderr, commandContractOptions(t)); code != test.wantCode || stdout.Len() != 0 || stderr.Len() == 0 {
+				t.Fatalf("code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
+			}
+		})
+	}
+}
+
+func TestConfigInspectionCommandsReportWriterFailures(t *testing.T) {
+	t.Parallel()
+	path := writeCommandConfig(t, "session", "other")
+	var stderr bytes.Buffer
+	if code := RunWithOptions([]string{"config", "effective", "--config", path}, failingInspectionWriter{}, &stderr, commandContractOptions(t)); code != 1 || !strings.Contains(stderr.String(), "write effective configuration") {
+		t.Fatalf("effective writer failure = %d, %q", code, stderr.String())
+	}
+}
+
+type failingInspectionWriter struct{}
+
+func (failingInspectionWriter) Write([]byte) (int, error) {
+	return 0, io.ErrClosedPipe
+}

internal/cli/configkit_load.go

@@ -0,0 +1,49 @@
+package cli
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"os"
+
+	configkit "gitea.maximumdirect.net/eric/ConfigKit"
+	"gitea.maximumdirect.net/eric/notarius/internal/core/config"
+)
+
+func loadConfigKit(configPath, profile string, overrides []configkit.Override, opts Options) (*config.LoadedConfig, string, error) {
+	path, err := discoverConfigPath(configPath, opts)
+	if err != nil {
+		return nil, "", err
+	}
+	workingDirectory, err := os.Getwd()
+	if err != nil {
+		return nil, "", fmt.Errorf("determine invocation directory: %w", err)
+	}
+	loaded, err := config.Load(path, config.LoadOptions{Profile: profile, Overrides: overrides, InvocationDirectory: workingDirectory})
+	if err != nil {
+		return nil, "", err
+	}
+	return loaded, path, nil
+}
+
+func writeConfigLoadError(output io.Writer, err error) {
+	var loadError *configkit.LoadError
+	if errors.As(err, &loadError) {
+		if writeErr := configkit.WriteDiagnostics(output, loadError.Diagnostics()); writeErr == nil {
+			// ConfigKit's locations contain source IDs. Keep its renderer and
+			// supply the acquired-source legend so imported/profile errors can
+			// be traced to a file. Quoting also escapes terminal controls.
+			for _, source := range loadError.Sources() {
+				label := source.Path
+				if label == "" {
+					label = source.Label
+				}
+				if _, err := fmt.Fprintf(output, "source %q: %q\n", source.ID, label); err != nil {
+					return
+				}
+			}
+			return
+		}
+	}
+	fmt.Fprintf(output, "notarius: %v\n", err)
+}

internal/cli/configkit_load_test.go

@@ -0,0 +1,85 @@
+package cli
+
+import (
+	"bytes"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+	"testing"
+
+	configkit "gitea.maximumdirect.net/eric/ConfigKit"
+)
+
+func TestConfigLoadDiagnosticsIdentifyComposedSources(t *testing.T) {
+	t.Parallel()
+	for _, profile := range []bool{false, true} {
+		t.Run(strconv.FormatBool(profile), func(t *testing.T) {
+			root := t.TempDir()
+			main := filepath.Join(root, "config.yml")
+			fragment := filepath.Join(root, "pipeline\n\x1b.yml")
+			args := []string{"config", "validate", "--config", main}
+			if profile {
+				fragment = filepath.Join(root, "profiles", "development.yml")
+				writeConfigKitFile(t, main, "version: 5\n")
+				args = append(args, "--config-profile", "development")
+			} else {
+				writeConfigKitFile(t, main, "version: 5\nconfigkit:\n  imports:\n    - path: "+strconv.Quote(filepath.Base(fragment))+"\n")
+			}
+			writeConfigKitFile(t, fragment, "concurrency: {total_llm: invalid-private-value}\n")
+			var stdout, stderr bytes.Buffer
+			if code := RunWithOptions(args, &stdout, &stderr, commandContractOptions(t)); code != 1 {
+				t.Fatalf("status = %d; %s", code, stderr.String())
+			}
+			got := stderr.String()
+			if stdout.Len() != 0 || !strings.Contains(got, "decode.type") ||
+				!strings.Contains(got, "source-2:1:") ||
+				!strings.Contains(got, strconv.Quote(fragment)) ||
+				strings.Contains(got, "invalid-private-value") ||
+				strings.ContainsRune(got, '\x1b') {
+				t.Fatalf("stdout = %q; stderr = %q", stdout.String(), got)
+			}
+		})
+	}
+}
+
+func TestLoadConfigKitUsesDiscoveryProfileAndOverrides(t *testing.T) {
+	t.Parallel()
+	root := t.TempDir()
+	path := filepath.Join(root, "notarius.yml")
+	writeConfigKitFile(t, path, "version: 5\npipelines:\n  base:\n    input: {module: input}\n")
+	profiles := filepath.Join(root, "profiles")
+	writeConfigKitFile(t, filepath.Join(profiles, "development.yml"), "pipelines:\n  profiled:\n    input: {module: input}\n")
+	loaded, gotPath, err := loadConfigKit(path, "development", []configkit.Override{{Path: configkit.Keys("output", "directory"), Value: "generated", Flag: "--output"}}, commandContractOptions(t))
+	if err != nil {
+		t.Fatalf("loadConfigKit() error = %v", err)
+	}
+	if gotPath != path || loaded.Config.Pipelines["profiled"].Input.Module != "input" || loaded.Config.Output.Directory == "generated" {
+		t.Fatalf("loaded = %#v, path=%q", loaded.Config, gotPath)
+	}
+}
+
+func TestWriteConfigLoadErrorUsesSafeDiagnostics(t *testing.T) {
+	t.Parallel()
+	path := filepath.Join(t.TempDir(), "invalid.yml")
+	writeConfigKitFile(t, path, "version: 4\n")
+	_, _, err := loadConfigKit(path, "", nil, commandContractOptions(t))
+	if err == nil {
+		t.Fatal("loadConfigKit() error = nil")
+	}
+	var output bytes.Buffer
+	writeConfigLoadError(&output, err)
+	if !strings.Contains(output.String(), "version") || strings.Contains(output.String(), "notarius: configkit") {
+		t.Fatalf("diagnostics = %q", output.String())
+	}
+}
+
+func writeConfigKitFile(t *testing.T, path, contents string) {
+	t.Helper()
+	if err := os.MkdirAll(filepath.Dir(path), 0o700); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(path, []byte(contents), 0o600); err != nil {
+		t.Fatal(err)
+	}
+}

internal/cli/dnd_enemy_events_contract_test.go

@@ -264,13 +264,29 @@ func completeExampleConfigWithTemporaryCache(t *testing.T) string {
 	updated := strings.Replace(string(content), "directory: ./notarius-cache/chunk-plans", fmt.Sprintf("directory: %q", filepath.Join(cacheRoot, "chunk-plans")), 1)
 	updated = strings.Replace(updated, "directory: ./notarius-cache/checkpoints", fmt.Sprintf("directory: %q", filepath.Join(cacheRoot, "checkpoints")), 1)
 	for relative, absolute := range map[string]string{
-		"./dnd-party.txt":          repositoryPath("examples", "dnd-party.txt"),
-		"./dnd-glossary.txt":       repositoryPath("examples", "dnd-glossary.txt"),
-		"./dnd-spell-catalog.json": repositoryPath("examples", "dnd-spell-catalog.json"),
+		"./llm-profiles/dnd-extraction.yml": repositoryPath("examples", "llm-profiles", "dnd-extraction.yml"),
 	} {
 		updated = strings.ReplaceAll(updated, relative, fmt.Sprintf("%q", absolute))
 	}
-	path := filepath.Join(t.TempDir(), "dnd-complete.config.yml")
+	root := t.TempDir()
+	fragment, err := os.ReadFile(repositoryPath("examples", "pipelines", "dnd-session.yml"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	for relative, absolute := range map[string]string{
+		"../dnd-party.txt":          repositoryPath("examples", "dnd-party.txt"),
+		"../dnd-glossary.txt":       repositoryPath("examples", "dnd-glossary.txt"),
+		"../dnd-spell-catalog.json": repositoryPath("examples", "dnd-spell-catalog.json"),
+	} {
+		fragment = []byte(strings.ReplaceAll(string(fragment), relative, fmt.Sprintf("%q", absolute)))
+	}
+	if err := os.MkdirAll(filepath.Join(root, "pipelines"), 0o700); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(filepath.Join(root, "pipelines", "dnd-session.yml"), fragment, 0o600); err != nil {
+		t.Fatal(err)
+	}
+	path := filepath.Join(root, "dnd-complete.config.yml")
 	if err := os.WriteFile(path, []byte(updated), 0o600); err != nil {
 		t.Fatal(err)
 	}
@@ -396,8 +412,12 @@ func (client *enemyEventLLMClient) CompleteStructured(ctx context.Context, reque
 	if err := json.Unmarshal(content, output); err != nil {
 		return contracts.StructuredCompletionResponse{}, fmt.Errorf("populate fake structured target: %w", err)
 	}
+	snapshot, err := contracts.CloneStructuredCompletionRequest(request)
+	if err != nil {
+		return contracts.StructuredCompletionResponse{}, fmt.Errorf("clone fake request: %w", err)
+	}
 	client.mu.Lock()
-	client.requests = append(client.requests, request)
+	client.requests = append(client.requests, snapshot)
 	client.mu.Unlock()
 	return contracts.StructuredCompletionResponse{Content: content, Provider: "test", Model: "deterministic", ProfileID: request.ProfileID}, nil
 }
@@ -408,7 +428,11 @@ func (client *enemyEventLLMClient) requestsFor(promptID string) []contracts.Stru
 	var requests []contracts.StructuredCompletionRequest
 	for _, request := range client.requests {
 		if request.PromptID == promptID {
-			requests = append(requests, request)
+			snapshot, err := contracts.CloneStructuredCompletionRequest(request)
+			if err != nil {
+				panic(err)
+			}
+			requests = append(requests, snapshot)
 		}
 	}
 	return requests

internal/cli/dnd_scene_descriptions_contract_test.go

@@ -91,8 +91,8 @@ func TestProductionSceneDescriptionWorkflow(t *testing.T) {
 	if !reflect.DeepEqual(durable, want) {
 		t.Fatalf("durable output payload = %#v, want %#v", durable, want)
 	}
-	if len(output.Warnings) != 0 {
-		t.Fatalf("warnings = %#v, want grounded descriptions without warnings", output.Warnings)
+	if len(output.Diagnostics.Groups) != 0 {
+		t.Fatalf("diagnostics = %#v, want grounded descriptions without diagnostics", output.Diagnostics)
 	}
 }
 

internal/cli/example_contract_test.go

@@ -147,8 +147,15 @@ func TestMaintainedConfigurationExampleSet(t *testing.T) {
 		}
 	}
 	sort.Strings(names)
-	if got := strings.Join(names, ","); got != "dnd-extraction.yml" {
-		t.Fatalf("maintained operator profiles = %q, want dnd-extraction.yml", got)
+	if got := strings.Join(names, ","); got != "development.yml" {
+		t.Fatalf("maintained configuration profiles = %q, want development.yml", got)
+	}
+	llmProfiles, err := os.ReadDir(repositoryPath("examples", "llm-profiles"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(llmProfiles) != 1 || llmProfiles[0].Name() != "dnd-extraction.yml" {
+		t.Fatalf("maintained LLM profiles = %#v, want dnd-extraction.yml", llmProfiles)
 	}
 }
 
@@ -239,10 +246,10 @@ func TestMaintainedMinimalInvocationProducesJSONBundle(t *testing.T) {
 		t.Fatalf("rejected = %#v, want empty rejection list", rejected.Rejected)
 	}
 	warnings := readProductionJSON[struct {
-		Warnings []json.RawMessage `json:"warnings"`
+		Groups []json.RawMessage `json:"groups"`
 	}](t, filepath.Join(runRoot, "warnings.json"))
-	if len(warnings.Warnings) != 0 {
-		t.Fatalf("warnings = %#v, want empty warning list", warnings.Warnings)
+	if len(warnings.Groups) != 0 {
+		t.Fatalf("warnings = %#v, want empty warning list", warnings.Groups)
 	}
 }
 

internal/cli/npc_registry_contract_test.go

@@ -23,7 +23,7 @@ func TestOversizedNPCRegistryFailsBeforeRuntimeAndCheckpointConstruction(t *test
 		t.Fatal(err)
 	}
 	checkpointRoot := filepath.Join(t.TempDir(), "checkpoints")
-	content := fmt.Sprintf(`version: 4
+	content := fmt.Sprintf(`version: 5
 cache:
   chunk_plans:
     mode: bypass
@@ -32,14 +32,17 @@ cache:
     directory: %q
 pipelines:
   dnd-session:
-    input: seriatim
+    input:
+      module: seriatim
     artifacts:
       spells:
         extract:
           module: dnd/spells
           references:
-            npc_registry: %q
-        normalize: dnd/spells
+            npc_registry:
+              path: %q
+        normalize:
+          module: dnd/spells
 `, checkpointRoot, npcPath)
 	configPath := filepath.Join(t.TempDir(), "config.yml")
 	if err := os.WriteFile(configPath, []byte(content), 0o600); err != nil {

internal/cli/production_contract_test.go

@@ -331,27 +331,8 @@ func TestProductionPromptAssetsPrepareWithoutProviderCredentials(t *testing.T) {
 	if err != nil {
 		t.Fatalf("production prompt engine: %v", err)
 	}
-	inputs := map[string]promptkit.ArtifactRef{
-		"candidates": promptkit.Inline(`{"candidates":[{"candidate_id":1,"label":"Alias","source_refs":[{"start_unit_id":1,"end_unit_id":1}]}]}`),
-		"transcript": promptkit.Inline(`{"windows":[{"units":[]}]}`),
-	}
-	for _, prompt := range []struct {
-		id      string
-		version string
-	}{
-		{id: npcnormalize.PromptID, version: npcnormalize.PromptVersion},
-		{id: itemregistrynormalize.PromptID, version: itemregistrynormalize.PromptVersion},
-		{id: locationnormalize.PromptID, version: locationnormalize.PromptVersion},
-	} {
-		prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{
-			PromptID: prompt.id, PromptVersion: prompt.version, ProfileID: "assembled-prompt-test", Inputs: inputs,
-		})
-		if err != nil {
-			t.Fatalf("prepare production prompt %q: %v", prompt.id, err)
-		}
-		if prepared.OutputContract.SchemaPath != filepath.Base(semanticreconcile.SchemaAssetPath) {
-			t.Fatalf("prompt %q schema = %q, want generic reconciliation schema", prompt.id, prepared.OutputContract.SchemaPath)
-		}
+	if engine == nil {
+		t.Fatal("production PromptKit engine is nil")
 	}
 }
 
@@ -701,13 +682,13 @@ func TestProductionConfigValidationCoversModuleAndVariantFailures(t *testing.T)
 	}{
 		{
 			name:      "unknown module",
-			content:   replaceRequiredOnce(t, base, "    input: seriatim\n", "    input: missing/input\n"),
+			content:   replaceRequiredOnce(t, base, "    input:\n      module: seriatim\n", "    input:\n      module: missing/input\n"),
 			options:   productionCLIOptions(t),
 			fragments: []string{"pipeline \"dnd-session\"", "input", "missing/input"},
 		},
 		{
 			name:      "unknown validator",
-			content:   replaceRequiredOnce(t, base, "        extract: dnd/spells\n", "        extract:\n          module: dnd/spells\n          validators:\n            - module: missing/validator\n"),
+			content:   replaceRequiredOnce(t, base, "        extract:\n          module: dnd/spells\n", "        extract:\n          module: dnd/spells\n          validators:\n            - module: missing/validator\n"),
 			options:   productionCLIOptions(t),
 			fragments: []string{"validator", "missing/validator"},
 		},
@@ -719,7 +700,7 @@ func TestProductionConfigValidationCoversModuleAndVariantFailures(t *testing.T)
 		},
 		{
 			name:      "deterministic validator with profile",
-			content:   replaceRequiredOnce(t, base, "        extract: dnd/spells\n", "        extract:\n          module: dnd/spells\n          validators:\n            - module: generic/valid_json\n              llm_profile: forbidden-profile\n"),
+			content:   replaceRequiredOnce(t, base, "        extract:\n          module: dnd/spells\n", "        extract:\n          module: dnd/spells\n          validators:\n            - module: generic/valid_json\n              llm_profile: forbidden-profile\n"),
 			options:   productionCLIOptions(t),
 			fragments: []string{"deterministic validator", "llm_profile"},
 		},
@@ -743,7 +724,7 @@ func TestProductionConfigValidationCoversModuleAndVariantFailures(t *testing.T)
 
 func TestProductionNormalizeValidatorOverrideRemainsAuthoritative(t *testing.T) {
 	base := string(readRepositoryFile(t, "examples", "dnd-minimal.config.yml"))
-	content := replaceRequiredOnce(t, base, "        normalize: dnd/spells\n", "        normalize:\n          module: dnd/spells\n          validators:\n            - module: generic/always_accept\n            - module: generic/valid_json\n")
+	content := replaceRequiredOnce(t, base, "        normalize:\n          module: dnd/spells\n", "        normalize:\n          module: dnd/spells\n          validators:\n            - module: generic/always_accept\n            - module: generic/valid_json\n")
 	path := writeProductionContractConfig(t, content)
 	components := productionTestComponents(t)
 	effective, err := loadMaintainedExample(t, path).Resolve(resolveInputForMaintainedExample(components, "dnd-session"))
@@ -816,10 +797,10 @@ func TestProductionSceneRunRecordsAnnotationFreeChunkPlanAndProvenance(t *testin
 		t.Fatalf("chunk map range annotations = %#v, want none", chunkMap.Chunks[0].Annotations)
 	}
 	warnings := readProductionJSON[struct {
-		Warnings []contracts.Warning `json:"warnings"`
+		Groups []contracts.DiagnosticGroup `json:"groups"`
 	}](t, filepath.Join(outputRoot, productionRunID, "warnings.json"))
-	if len(warnings.Warnings) != 0 {
-		t.Fatalf("warnings = %#v, want none", warnings.Warnings)
+	if len(warnings.Groups) != 0 {
+		t.Fatalf("warnings = %#v, want none", warnings.Groups)
 	}
 	if len(fake.requestsFor(scenes.PromptID)) != 1 || len(fake.requestsFor(spells.PromptID)) != 1 || len(fake.requestsFor(itemoccurrenceextract.PromptID)) != 1 {
 		t.Fatalf("fake prompt requests = %#v, want one scene, spell, and item-occurrence request", fake.requestPrompts())
@@ -843,7 +824,7 @@ func maintainedExampleFiles(t *testing.T) []maintainedExample {
 
 func productionSpellCatalogContractConfig(t *testing.T) string {
 	t.Helper()
-	return fmt.Sprintf(`version: 4
+	return fmt.Sprintf(`version: 5
 cache:
   chunk_plans:
     mode: bypass
@@ -852,21 +833,26 @@ cache:
     directory: ""
 pipelines:
   dnd-session:
-    input: seriatim
+    input:
+      module: seriatim
     references:
-      party: %q
-      glossary: %q
+      party:
+        path: %q
+      glossary:
+        path: %q
     artifacts:
       spells:
         extract:
           module: dnd/spells
           retries: 2
           references:
-            spell_catalog: %q
+            spell_catalog:
+              path: %q
         normalize:
           module: dnd/spells
           references:
-            spell_catalog: %q
+            spell_catalog:
+              path: %q
 `, repositoryPath("examples", "dnd-party.txt"), repositoryPath("examples", "dnd-glossary.txt"), repositoryPath("examples", "dnd-spell-catalog.json"), repositoryPath("examples", "dnd-spell-catalog.json"))
 }
 
@@ -877,14 +863,11 @@ func writeProductionSpellCatalogContractConfig(t *testing.T) string {
 
 func loadMaintainedExample(t *testing.T, path string) config.Config {
 	t.Helper()
-	fileConfig, err := config.LoadFileConfig(path)
+	loaded, err := config.Load(path, config.LoadOptions{})
 	if err != nil {
 		t.Fatalf("load maintained config %q: %v", path, err)
 	}
-	cfg := config.Default()
-	if err := cfg.ApplyFileConfig(fileConfig); err != nil {
-		t.Fatalf("apply maintained config %q: %v", path, err)
-	}
+	cfg := loaded.RuntimeConfig()
 	if err := cfg.Validate(); err != nil {
 		t.Fatalf("validate maintained config %q: %v", path, err)
 	}
@@ -941,7 +924,7 @@ func productionRunOptions(t *testing.T, fake *productionFakeLLMClient) Options {
 }
 
 func productionRunConfig(outputRoot, chunkModule string) string {
-	return fmt.Sprintf(`version: 4
+	return fmt.Sprintf(`version: 5
 output:
   directory: %q
 cache:
@@ -952,8 +935,10 @@ debug:
   directory: %q
 pipelines:
   dnd-session:
-    input: seriatim
-    chunk: %s
+    input:
+      module: seriatim
+    chunk:
+      module: %s
     output:
       module: json
       options:
@@ -962,8 +947,10 @@ pipelines:
       - id: identify-items
         artifacts:
           item-registry:
-            extract: dnd/item-registry
-            normalize: dnd/item-registry
+            extract:
+              module: dnd/item-registry
+            normalize:
+              module: dnd/item-registry
       - id: extract-occurrences
         references:
           item_registry:
@@ -972,9 +959,11 @@ pipelines:
               lane: item-registry
         artifacts:
           spells:
-            extract: dnd/spells
+            extract:
+              module: dnd/spells
           item-occurrences:
-            extract: dnd/item-occurrences
+            extract:
+              module: dnd/item-occurrences
 `, outputRoot, filepath.Join(filepath.Dir(outputRoot), "debug"), chunkModule)
 }
 
@@ -1063,8 +1052,12 @@ func (client *productionFakeLLMClient) CompleteStructured(ctx context.Context, r
 	if err := json.Unmarshal(content, out); err != nil {
 		return contracts.StructuredCompletionResponse{}, fmt.Errorf("populate fake structured target: %w", err)
 	}
+	snapshot, err := contracts.CloneStructuredCompletionRequest(req)
+	if err != nil {
+		return contracts.StructuredCompletionResponse{}, fmt.Errorf("clone fake request: %w", err)
+	}
 	client.mu.Lock()
-	client.requests = append(client.requests, req)
+	client.requests = append(client.requests, snapshot)
 	client.mu.Unlock()
 	return contracts.StructuredCompletionResponse{Content: content, Provider: "test", Model: "deterministic", ProfileID: req.ProfileID}, nil
 }
@@ -1075,7 +1068,11 @@ func (client *productionFakeLLMClient) requestsFor(promptID string) []contracts.
 	var requests []contracts.StructuredCompletionRequest
 	for _, req := range client.requests {
 		if req.PromptID == promptID {
-			requests = append(requests, req)
+			snapshot, err := contracts.CloneStructuredCompletionRequest(req)
+			if err != nil {
+				panic(err)
+			}
+			requests = append(requests, snapshot)
 		}
 	}
 	return requests

internal/cli/promptkit_profiles_test.go

@@ -71,10 +71,10 @@ api_key_env: NOTARIUS_PROMPTKIT_PROFILE_INSPECTION_TEST_KEY
 		},
 		{
 			name:        "malformed profile",
-			profilePath: writeProfile(t, "malformed-profile", "id: malformed-profile\nbackend: [\n"),
+			profilePath: writeProfile(t, "malformed-profile", "id: malformed-profile\nendpoint: https://provider.example/v1?credential=forbidden\nmodel: malformed-model\n"),
 			profileID:   "malformed-profile",
 			wantErr:     []string{`PromptKit profile "malformed-profile" is invalid or unreadable`},
-			rejectErr:   []string{"malformed-profile.yaml", "backend: ["},
+			rejectErr:   []string{"malformed-profile.yaml", "credential=forbidden"},
 		},
 		{
 			name:        "invalid profile source",
@@ -157,3 +157,25 @@ func TestExplicitPromptKitProfileValidationUsesFallbackAssets(t *testing.T) {
 		t.Fatalf("validateExplicitPromptKitProfiles() error = %v, want nil", err)
 	}
 }
+
+func TestExplicitPromptKitProfileValidationRejectsInvalidInheritanceBeforeGeneration(t *testing.T) {
+	for _, profiles := range []string{
+		"id: child\nbase_profile: missing\n",
+		"id: first\nbase_profile: second\n\n---\nid: second\nbase_profile: first\n",
+	} {
+		t.Run("invalid inheritance", func(t *testing.T) {
+			profilePath := filepath.Join(t.TempDir(), "profiles.yaml")
+			if err := os.WriteFile(profilePath, []byte(profiles), 0o600); err != nil {
+				t.Fatal(err)
+			}
+			profileID := "child"
+			if strings.Contains(profiles, "id: first") {
+				profileID = "first"
+			}
+			err := validateExplicitPromptKitProfiles(context.Background(), config.Config{PromptKit: config.PromptKitConfig{ProfileFile: profilePath}}, []string{profileID}, nil)
+			if err == nil || !strings.Contains(err.Error(), "invalid or unreadable") || strings.Contains(err.Error(), profilePath) {
+				t.Fatalf("profile preflight error = %v", err)
+			}
+		})
+	}
+}

internal/cli/recompute_execution_contract_test.go

@@ -194,7 +194,7 @@ func (recomputeTestOutput) Encode(_ context.Context, req contracts.OutputRequest
 func newRecomputeTestRoots(t *testing.T) stateTestRoots {
 	t.Helper()
 	roots := newStateTestRoots(t)
-	config := fmt.Sprintf(`version: 4
+	config := fmt.Sprintf(`version: 5
 output:
   directory: %q
 cache:
@@ -208,19 +208,27 @@ debug:
   directory: %q
 pipelines:
   sample:
-    input: test/input
-    chunk: test/chunk
+    input:
+      module: test/input
+    chunk:
+      module: test/chunk
     steps:
       - id: first
         artifacts:
           producer:
-            extract: test/extract/producer
-            merge: test/merge
-            normalize: test/normalize
+            extract:
+              module: test/extract/producer
+            merge:
+              module: test/merge
+            normalize:
+              module: test/normalize
           unrelated:
-            extract: test/extract/unrelated
-            merge: test/merge
-            normalize: test/normalize
+            extract:
+              module: test/extract/unrelated
+            merge:
+              module: test/merge
+            normalize:
+              module: test/normalize
       - id: second
         references:
           upstream:
@@ -229,9 +237,12 @@ pipelines:
               lane: producer
         artifacts:
           middle:
-            extract: test/extract/middle
-            merge: test/merge
-            normalize: test/normalize
+            extract:
+              module: test/extract/middle
+            merge:
+              module: test/merge
+            normalize:
+              module: test/normalize
       - id: third
         references:
           upstream:
@@ -240,10 +251,14 @@ pipelines:
               lane: middle
         artifacts:
           dependent:
-            extract: test/extract/dependent
-            merge: test/merge
-            normalize: test/normalize
-    output: test/recompute-output
+            extract:
+              module: test/extract/dependent
+            merge:
+              module: test/merge
+            normalize:
+              module: test/normalize
+    output:
+      module: test/recompute-output
 `, roots.output, roots.plans, roots.checkpoints, roots.debug)
 	if err := os.WriteFile(roots.config, []byte(config), 0o600); err != nil {
 		t.Fatal(err)

internal/cli/reference_contract_test.go

@@ -4,6 +4,7 @@ import (
 	"bytes"
 	"os"
 	"path/filepath"
+	"slices"
 	"strings"
 	"testing"
 
@@ -14,20 +15,16 @@ import (
 
 func TestReferenceSelectorsParseAndApplyAllDocumentedForms(t *testing.T) {
 	tests := []struct {
-		name      string
-		selector  string
-		only      []string
-		wantStage pipeline.ModuleStage
-		wantLane  string
-		wantSlot  string
+		name     string
+		selector string
+		want     []string
 	}{
-		{name: "flat", selector: "alpha-slot", wantStage: pipeline.StageExtract, wantLane: "alpha", wantSlot: "alpha-slot"},
-		{name: "chunk", selector: "chunk.chunk-slot", wantStage: pipeline.StageChunk, wantSlot: "chunk-slot"},
-		{name: "merge", selector: "merge.alpha-merge", only: []string{"alpha"}, wantStage: pipeline.StageMerge, wantLane: "alpha", wantSlot: "alpha-merge"},
-		{name: "lane", selector: "alpha.alpha-slot", wantStage: pipeline.StageExtract, wantLane: "alpha", wantSlot: "alpha-slot"},
-		{name: "lane extract", selector: "alpha.extract.alpha-slot", wantStage: pipeline.StageExtract, wantLane: "alpha", wantSlot: "alpha-slot"},
-		{name: "lane merge", selector: "alpha.merge.alpha-merge", wantStage: pipeline.StageMerge, wantLane: "alpha", wantSlot: "alpha-merge"},
-		{name: "lane normalize", selector: "alpha.normalize.alpha-normalize", wantStage: pipeline.StageNormalize, wantLane: "alpha", wantSlot: "alpha-normalize"},
+		{name: "pipeline", selector: "shared", want: []string{"alpha.extract.shared", "alpha.merge.shared", "alpha.normalize.shared", "beta.extract.shared", "beta.merge.shared", "beta.normalize.shared"}},
+		{name: "chunk", selector: "chunk.chunk-slot", want: []string{"chunk.chunk-slot"}},
+		{name: "lane", selector: "alpha.shared", want: []string{"alpha.extract.shared", "alpha.merge.shared", "alpha.normalize.shared"}},
+		{name: "lane extract", selector: "alpha.extract.alpha-slot", want: []string{"alpha.extract.alpha-slot"}},
+		{name: "lane merge", selector: "alpha.merge.alpha-merge", want: []string{"alpha.merge.alpha-merge"}},
+		{name: "lane normalize", selector: "alpha.normalize.alpha-normalize", want: []string{"alpha.normalize.alpha-normalize"}},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
@@ -37,70 +34,135 @@ func TestReferenceSelectorsParseAndApplyAllDocumentedForms(t *testing.T) {
 			if err != nil {
 				t.Fatal(err)
 			}
-			overrides, _, err := resolveCLIReferenceRequests(cfg, "demo", tt.only, catalog, []cliReferenceRequest{{Selector: selector, Source: "reference.txt"}}, nil)
+			overrides, _, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog, []cliReferenceRequest{{Selector: selector, Source: "reference.txt"}}, nil)
 			if err != nil {
 				t.Fatalf("resolve selector: %v", err)
 			}
-			if len(overrides) != 1 {
-				t.Fatalf("overrides = %#v, want one binding", overrides)
+			if got := referenceContractBindingLabels(overrides); !slices.Equal(got, tt.want) {
+				t.Fatalf("binding targets = %#v, want %#v", got, tt.want)
 			}
-			got := overrides[0]
-			if got.Stage != tt.wantStage || got.LaneID != tt.wantLane || got.SlotName != tt.wantSlot || got.BindingSource != contracts.ReferenceBindingSourceCLI {
-				t.Fatalf("binding = %#v, want %s/%s/%s from CLI", got, tt.wantStage, tt.wantLane, tt.wantSlot)
-			}
-		})
-	}
-}
-
-func TestReferenceSelectorsRejectAmbiguityWithSpecificSuggestions(t *testing.T) {
-	cfg := referenceContractConfig()
-	catalog := referenceContractCatalog(t, true, true)
-	for _, tt := range []struct {
-		name     string
-		selector string
-		want     []string
-	}{
-		{name: "flat shared slot", selector: "shared", want: []string{"alpha.extract.shared", "beta.extract.shared"}},
-		{name: "lane shared slot", selector: "alpha.shared", want: []string{"alpha.extract.shared", "alpha.merge.shared", "alpha.normalize.shared"}},
-		{name: "all mergers", selector: "merge.shared", want: []string{"alpha.merge.shared", "beta.merge.shared"}},
-	} {
-		t.Run(tt.name, func(t *testing.T) {
-			selector, err := parseReferenceSelector(tt.selector, "--reference")
-			if err != nil {
-				t.Fatal(err)
-			}
-			_, _, err = resolveCLIReferenceRequests(cfg, "demo", nil, catalog, []cliReferenceRequest{{Selector: selector, Source: "reference.txt"}}, nil)
-			if err == nil {
-				t.Fatal("resolve selector succeeded, want ambiguity error")
-			}
-			for _, fragment := range tt.want {
-				if !strings.Contains(err.Error(), fragment) {
-					t.Fatalf("error = %q, want suggestion %q", err, fragment)
+			for _, binding := range overrides {
+				if binding.Source != "reference.txt" || binding.BindingSource != contracts.ReferenceBindingSourceCLI {
+					t.Fatalf("binding = %#v, want CLI source", binding)
 				}
 			}
 		})
 	}
 }
 
-func TestReferenceSelectorsRespectSelectedLanesBeforeMaterialization(t *testing.T) {
+func TestReferenceSelectorSpecificityAndFinalOccurrenceChooseConcreteBindings(t *testing.T) {
+	cfg := referenceContractConfig()
+	catalog := referenceContractCatalog(t, true, true)
+	requests := []cliReferenceRequest{
+		{Selector: mustParseReferenceSelector(t, "shared", "--reference"), Source: "pipeline-first.txt"},
+		{Selector: mustParseReferenceSelector(t, "shared", "--reference"), Source: "pipeline-final.txt"},
+		{Selector: mustParseReferenceSelector(t, "alpha.shared", "--reference"), Source: "lane.txt"},
+		{Selector: mustParseReferenceSelector(t, "alpha.extract.shared", "--reference"), Source: "binding.txt"},
+	}
+	overrides, unbinds, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog, requests, nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(unbinds) != 0 {
+		t.Fatalf("unbinds = %#v, want none", unbinds)
+	}
+	want := map[string]string{
+		"alpha.extract.shared":   "binding.txt",
+		"alpha.merge.shared":     "lane.txt",
+		"alpha.normalize.shared": "lane.txt",
+		"beta.extract.shared":    "pipeline-final.txt",
+		"beta.merge.shared":      "pipeline-final.txt",
+		"beta.normalize.shared":  "pipeline-final.txt",
+	}
+	for _, binding := range overrides {
+		label := referenceContractBindingLabel(binding)
+		if binding.Source != want[label] {
+			t.Fatalf("binding %s source = %q, want %q", label, binding.Source, want[label])
+		}
+		delete(want, label)
+	}
+	if len(want) != 0 {
+		t.Fatalf("missing bindings: %#v", want)
+	}
+}
+
+func TestCompleteDNDSharedCLIReferencesExpandAcrossCompatibleTargets(t *testing.T) {
+	cfg := loadMaintainedExample(t, repositoryPath("examples", "dnd-complete.config.yml"))
+	catalog := catalogFromRegistries(productionTestComponents(t).registries)
+	sources := map[string]string{
+		"party":         "/references/party.txt",
+		"players":       "/references/players.txt",
+		"glossary":      "/references/glossary.txt",
+		"spell_catalog": "/references/spells.json",
+	}
+	requests := make([]cliReferenceRequest, 0, len(sources))
+	for _, slot := range []string{"party", "players", "glossary", "spell_catalog"} {
+		requests = append(requests, cliReferenceRequest{
+			Selector: mustParseReferenceSelector(t, slot, "--reference"),
+			Source:   sources[slot],
+		})
+	}
+	overrides, unbinds, err := resolveCLIReferenceRequests(cfg, "dnd-session", nil, catalog, requests, nil)
+	if err != nil {
+		t.Fatalf("expand complete D&D references: %v", err)
+	}
+	if len(unbinds) != 0 {
+		t.Fatalf("unbinds = %#v, want none", unbinds)
+	}
+
+	actual := make(map[string]pipeline.ReferenceBinding, len(overrides))
+	for _, binding := range overrides {
+		actual[referenceContractBindingLabel(binding)] = binding
+	}
+	targets, err := selectedReferenceTargets(cfg, "dnd-session", nil, catalog)
+	if err != nil {
+		t.Fatal(err)
+	}
+	matched := make(map[string]int, len(sources))
+	for _, target := range targets {
+		for slot, sourcePath := range sources {
+			if _, ok := target.slots[slot]; !ok {
+				continue
+			}
+			matched[slot]++
+			label := targetLabel(target) + "." + slot
+			binding, ok := actual[label]
+			if !ok || binding.Source != sourcePath || binding.BindingSource != contracts.ReferenceBindingSourceCLI {
+				t.Fatalf("binding %q = %#v, want CLI source %q", label, binding, sourcePath)
+			}
+		}
+	}
+	for slot := range sources {
+		if matched[slot] < 2 {
+			t.Fatalf("reference %q matched %d target(s), want a shared D&D reference", slot, matched[slot])
+		}
+	}
+	if _, err := cfg.Resolve(config.ResolveInput{PipelineID: "dnd-session", Catalog: catalog, ReferenceOverrides: overrides}); err != nil {
+		t.Fatalf("resolve complete D&D CLI references: %v", err)
+	}
+}
+
+func TestReferenceSelectorsRejectInvalidOrUnselectedScopesBeforeMaterialization(t *testing.T) {
 	cfg := referenceContractConfig()
 	catalog := referenceContractCatalog(t, true, true)
 	for _, tt := range []struct {
 		name     string
 		selector string
+		only     []string
 		want     string
 	}{
-		{name: "unselected lane", selector: "beta.extract.beta-slot", want: `reference lane "beta" is not selected`},
+		{name: "pipeline slot", selector: "missing", want: `reference slot "missing" is not declared by any selected target`},
+		{name: "lane slot", selector: "alpha.missing", want: `reference slot "missing" is not declared by selected lane "alpha"`},
+		{name: "binding slot", selector: "alpha.extract.missing", want: `reference slot "missing" is not declared`},
+		{name: "former merge shorthand", selector: "merge.shared", want: `reference lane "merge" is not selected`},
+		{name: "unselected lane", selector: "beta.extract.beta-slot", only: []string{"alpha"}, want: `reference lane "beta" is not selected`},
 		{name: "unknown lane", selector: "missing.extract.beta-slot", want: `reference lane "missing" is not selected`},
 	} {
 		t.Run(tt.name, func(t *testing.T) {
-			selector, err := parseReferenceSelector(tt.selector, "--reference")
-			if err != nil {
-				t.Fatal(err)
-			}
-			_, _, err = resolveCLIReferenceRequests(cfg, "demo", []string{"alpha"}, catalog, []cliReferenceRequest{{Selector: selector, Source: filepath.Join(t.TempDir(), "missing.txt")}}, nil)
+			selector := mustParseReferenceSelector(t, tt.selector, "--reference")
+			_, _, err := resolveCLIReferenceRequests(cfg, "demo", tt.only, catalog, []cliReferenceRequest{{Selector: selector, Source: filepath.Join(t.TempDir(), "missing.txt")}}, nil)
 			if err == nil || !strings.Contains(err.Error(), tt.want) || strings.Contains(err.Error(), "missing.txt") {
-				t.Fatalf("error = %v, want selection failure before file access", err)
+				t.Fatalf("error = %v, want selection failure containing %q before file access", err, tt.want)
 			}
 		})
 	}
@@ -131,40 +193,50 @@ func TestReferenceSyntaxErrorsReturnTwo(t *testing.T) {
 	}
 }
 
-func TestReferenceOverridesUseFinalExactTargetBinding(t *testing.T) {
+func TestReferenceBindAndUnbindSpecificity(t *testing.T) {
 	cfg := referenceContractConfig()
 	catalog := referenceContractCatalog(t, true, true)
-	alphaShared, err := parseReferenceSelector("alpha.extract.shared", "--reference")
-	if err != nil {
-		t.Fatal(err)
-	}
-	betaShared, err := parseReferenceSelector("beta.extract.shared", "--reference")
-	if err != nil {
-		t.Fatal(err)
-	}
-	overrides, unbinds, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog, []cliReferenceRequest{
-		{Selector: alphaShared, Source: "alpha-first.txt"},
-		{Selector: alphaShared, Source: "alpha-final.txt"},
-		{Selector: betaShared, Source: "beta-only.txt"},
-	}, nil)
-	if err != nil {
-		t.Fatal(err)
-	}
-	if len(unbinds) != 0 {
-		t.Fatalf("unbinds = %#v, want none", unbinds)
-	}
-	effective, err := cfg.Resolve(config.ResolveInput{PipelineID: "demo", Catalog: catalog, ReferenceOverrides: overrides})
-	if err != nil {
-		t.Fatalf("resolve pipeline: %v", err)
-	}
-	alpha := referenceContractLane(t, effective.ResolvedPipeline, "alpha")
-	beta := referenceContractLane(t, effective.ResolvedPipeline, "beta")
-	if source := referenceContractBindingSource(alpha.ExtractReferences.Bindings, "shared"); source != "alpha-final.txt" {
-		t.Fatalf("alpha shared source = %q, want final exact-target override", source)
-	}
-	if source := referenceContractBindingSource(beta.ExtractReferences.Bindings, "shared"); source != "beta-only.txt" {
-		t.Fatalf("beta shared source = %q, want target-specific override", source)
-	}
+	t.Run("specific unbind carves out broad binding", func(t *testing.T) {
+		overrides, unbinds, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog,
+			[]cliReferenceRequest{{Selector: mustParseReferenceSelector(t, "shared", "--reference"), Source: "shared.txt"}},
+			[]cliReferenceUnbindRequest{{Selector: mustParseReferenceSelector(t, "alpha.extract.shared", "--without-reference")}},
+		)
+		if err != nil {
+			t.Fatal(err)
+		}
+		if got := referenceContractBindingLabels(overrides); slices.Contains(got, "alpha.extract.shared") || len(got) != 5 {
+			t.Fatalf("overrides = %#v, want all shared targets except alpha extract", got)
+		}
+		if got := referenceContractUnbindLabels(unbinds); !slices.Equal(got, []string{"alpha.extract.shared"}) {
+			t.Fatalf("unbinds = %#v, want alpha extract", got)
+		}
+	})
+
+	t.Run("specific binding restores broad unbind", func(t *testing.T) {
+		overrides, unbinds, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog,
+			[]cliReferenceRequest{{Selector: mustParseReferenceSelector(t, "alpha.extract.shared", "--reference"), Source: "alpha.txt"}},
+			[]cliReferenceUnbindRequest{{Selector: mustParseReferenceSelector(t, "shared", "--without-reference")}},
+		)
+		if err != nil {
+			t.Fatal(err)
+		}
+		if got := referenceContractBindingLabels(overrides); !slices.Equal(got, []string{"alpha.extract.shared"}) {
+			t.Fatalf("overrides = %#v, want alpha extract", got)
+		}
+		if got := referenceContractUnbindLabels(unbinds); slices.Contains(got, "alpha.extract.shared") || len(got) != 5 {
+			t.Fatalf("unbinds = %#v, want all shared targets except alpha extract", got)
+		}
+	})
+
+	t.Run("same specificity conflicts", func(t *testing.T) {
+		_, _, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog,
+			[]cliReferenceRequest{{Selector: mustParseReferenceSelector(t, "alpha.shared", "--reference"), Source: "alpha.txt"}},
+			[]cliReferenceUnbindRequest{{Selector: mustParseReferenceSelector(t, "alpha.shared", "--without-reference")}},
+		)
+		if err == nil || !strings.Contains(err.Error(), "same specificity") {
+			t.Fatalf("error = %v, want same-specificity conflict", err)
+		}
+	})
 }
 
 func TestReferenceUnbindsRemoveOptionalAndProtectRequiredSlots(t *testing.T) {
@@ -217,7 +289,7 @@ func TestReferenceMaterializationSeparatesCLIAndConfigPathOrigins(t *testing.T)
 	workingDir := t.TempDir()
 	cfg := referenceContractConfig()
 	configPath := filepath.Join(configDir, "config.yml")
-	if err := os.WriteFile(configPath, []byte("version: 4\n"), 0o600); err != nil {
+	if err := os.WriteFile(configPath, []byte("version: 5\n"), 0o600); err != nil {
 		t.Fatal(err)
 	}
 	if err := os.WriteFile(filepath.Join(configDir, "required.txt"), []byte("config reference"), 0o600); err != nil {
@@ -257,6 +329,52 @@ func TestReferenceMaterializationSeparatesCLIAndConfigPathOrigins(t *testing.T)
 	}
 }
 
+func TestPipelineScopedCLIReferenceProtectsGeneratedHandoff(t *testing.T) {
+	cfg := referenceContractConfig()
+	profile := cfg.Pipelines["demo"]
+	alpha := profile.Artifacts["alpha"]
+	beta := profile.Artifacts["beta"]
+	alpha.Extract.References["shared"] = pipeline.GeneratedReference("produce", "beta")
+	profile.Artifacts = nil
+	profile.Steps = []pipeline.PipelineStepProfile{
+		{ID: "produce", Artifacts: map[string]pipeline.ArtifactLaneProfile{"beta": beta}},
+		{ID: "consume", Artifacts: map[string]pipeline.ArtifactLaneProfile{"alpha": alpha}},
+	}
+	cfg.Pipelines["demo"] = profile
+	catalog := referenceContractCatalog(t, true, true)
+
+	t.Run("binding conflicts before file access", func(t *testing.T) {
+		overrides, _, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog, []cliReferenceRequest{{
+			Selector: mustParseReferenceSelector(t, "shared", "--reference"),
+			Source:   filepath.Join(t.TempDir(), "never-read.json"),
+		}}, nil)
+		if err != nil {
+			t.Fatalf("expand CLI reference: %v", err)
+		}
+		_, err = cfg.Resolve(config.ResolveInput{PipelineID: "demo", Catalog: catalog, ReferenceOverrides: overrides})
+		if err == nil || !strings.Contains(err.Error(), "conflicting generated and external bindings") || strings.Contains(err.Error(), "never-read.json") {
+			t.Fatalf("resolve error = %v, want generated/external conflict before file access", err)
+		}
+	})
+
+	t.Run("unbind leaves generated source intact", func(t *testing.T) {
+		_, unbinds, err := resolveCLIReferenceRequests(cfg, "demo", nil, catalog, nil, []cliReferenceUnbindRequest{{
+			Selector: mustParseReferenceSelector(t, "shared", "--without-reference"),
+		}})
+		if err != nil {
+			t.Fatalf("expand CLI unbind: %v", err)
+		}
+		effective, err := cfg.Resolve(config.ResolveInput{PipelineID: "demo", Catalog: catalog, ReferenceUnbinds: unbinds})
+		if err != nil {
+			t.Fatalf("resolve generated reference with CLI unbind: %v", err)
+		}
+		binding := referenceContractFindBinding(referenceContractLane(t, effective.ResolvedPipeline, "alpha").ExtractReferences.Bindings, "shared")
+		if binding == nil || binding.Artifact == nil || binding.Artifact.Step != "produce" || binding.Artifact.Lane != "beta" {
+			t.Fatalf("generated binding = %#v, want preserved produce/beta handoff", binding)
+		}
+	})
+}
+
 func TestReferenceTargetLookupUsesArtifactVariantsAndReportsMissingContext(t *testing.T) {
 	cfg := referenceContractConfig()
 	full := referenceContractCatalog(t, true, true)
@@ -368,7 +486,7 @@ func referenceContractCatalog(t *testing.T, includeBetaMerger, includeBetaNormal
 	register(registries.Chunkers.RegisterWithSpec(pipeline.ModuleSpec{Key: "reference/chunk", Stage: pipeline.StageChunk, ExecutionClass: contracts.ExecutionClassDeterministic, Requires: []string{"source"}, Provides: []string{"chunks"}, ReferenceSlots: []contracts.ReferenceSlot{{Name: "chunk-slot"}, {Name: "required-chunk", Required: true}}}, func() (contracts.Chunker, error) { return stateTestChunker{}, nil }))
 	register(pipeline.RegisterArtifactCodec(registries.ArtifactCodecs, referenceContractCodecA{}))
 	register(pipeline.RegisterArtifactCodec(registries.ArtifactCodecs, referenceContractCodecB{}))
-	register(pipeline.RegisterExtractor(registries.Extractors, pipeline.ModuleSpec{Key: "reference/extract-alpha", Stage: pipeline.StageExtract, ExecutionClass: contracts.ExecutionClassDeterministic, Requires: []string{"chunks"}, Provides: []string{"artifact"}, ArtifactKind: referenceContractKindAlpha, ReferenceSlots: []contracts.ReferenceSlot{{Name: "shared"}, {Name: "alpha-slot"}, {Name: "required-extract", Required: true}}}, func() (contracts.Extractor[stateTestArtifact], error) { return stateTestExtractor{}, nil }))
+	register(pipeline.RegisterExtractor(registries.Extractors, pipeline.ModuleSpec{Key: "reference/extract-alpha", Stage: pipeline.StageExtract, ExecutionClass: contracts.ExecutionClassDeterministic, Requires: []string{"chunks"}, Provides: []string{"artifact"}, ArtifactKind: referenceContractKindAlpha, ReferenceSlots: []contracts.ReferenceSlot{{Name: "shared", AcceptedArtifactKinds: []contracts.ArtifactKind{referenceContractKindBeta}, AcceptedMediaTypes: []string{"application/json"}}, {Name: "alpha-slot"}, {Name: "required-extract", Required: true}}}, func() (contracts.Extractor[stateTestArtifact], error) { return stateTestExtractor{}, nil }))
 	register(pipeline.RegisterExtractor(registries.Extractors, pipeline.ModuleSpec{Key: "reference/extract-beta", Stage: pipeline.StageExtract, ExecutionClass: contracts.ExecutionClassDeterministic, Requires: []string{"chunks"}, Provides: []string{"artifact"}, ArtifactKind: referenceContractKindBeta, ReferenceSlots: []contracts.ReferenceSlot{{Name: "shared"}, {Name: "beta-slot"}, {Name: "required-extract", Required: true}}}, func() (contracts.Extractor[stateTestArtifact], error) { return stateTestExtractor{}, nil }))
 	register(pipeline.RegisterMerger(registries.Mergers, pipeline.ModuleSpec{Key: "reference/shared-merge", Stage: pipeline.StageMerge, ExecutionClass: contracts.ExecutionClassDeterministic, Requires: []string{"artifact"}, Provides: []string{"merged"}, ArtifactKind: referenceContractKindAlpha, ReferenceSlots: []contracts.ReferenceSlot{{Name: "shared"}, {Name: "alpha-merge"}, {Name: "required-merge", Required: true}}}, func() (contracts.Merger[stateTestArtifact], error) { return stateTestMerger{}, nil }))
 	if includeBetaMerger {
@@ -444,6 +562,42 @@ func referenceContractBindingSource(bindings []pipeline.ReferenceBinding, slot s
 	return ""
 }
 
+func mustParseReferenceSelector(t *testing.T, value, flagName string) cliReferenceSelector {
+	t.Helper()
+	selector, err := parseReferenceSelector(value, flagName)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return selector
+}
+
+func referenceContractBindingLabels(bindings []pipeline.ReferenceBinding) []string {
+	labels := make([]string, 0, len(bindings))
+	for _, binding := range bindings {
+		labels = append(labels, referenceContractBindingLabel(binding))
+	}
+	return labels
+}
+
+func referenceContractBindingLabel(binding pipeline.ReferenceBinding) string {
+	if binding.Stage == pipeline.StageChunk {
+		return "chunk." + binding.SlotName
+	}
+	return binding.LaneID + "." + string(binding.Stage) + "." + binding.SlotName
+}
+
+func referenceContractUnbindLabels(unbinds []pipeline.ReferenceUnbind) []string {
+	labels := make([]string, 0, len(unbinds))
+	for _, unbind := range unbinds {
+		labels = append(labels, referenceContractBindingLabel(pipeline.ReferenceBinding{
+			Stage:    unbind.Stage,
+			LaneID:   unbind.LaneID,
+			SlotName: unbind.SlotName,
+		}))
+	}
+	return labels
+}
+
 func referenceContractFindBinding(bindings []pipeline.ReferenceBinding, slot string) *pipeline.ReferenceBinding {
 	for i := range bindings {
 		if bindings[i].SlotName == slot {

internal/cli/run.go

@@ -4,7 +4,6 @@ import (
 	"context"
 	"crypto/sha256"
 	"encoding/hex"
-	"encoding/json"
 	"errors"
 	"flag"
 	"fmt"
@@ -16,6 +15,8 @@ import (
 	"strings"
 	"time"
 
+	configkit "gitea.maximumdirect.net/eric/ConfigKit"
+	"gitea.maximumdirect.net/eric/notarius/internal/buildinfo"
 	"gitea.maximumdirect.net/eric/notarius/internal/core/artifacts"
 	"gitea.maximumdirect.net/eric/notarius/internal/core/config"
 	"gitea.maximumdirect.net/eric/notarius/internal/core/debugbundle"
@@ -31,9 +32,12 @@ import (
 const defaultConfigPath = "/usr/local/etc/notarius/config.yml"
 const usage = `Usage:
   notarius help
+  notarius --version
   notarius run <pipeline-id> --input path/to/source.json [--json] [flags]
-  notarius config validate --config path/to/config.yml [--pipeline pipeline-id] [--only lane-a,lane-b]
-  notarius pipelines list --config path/to/config.yml [--json]
+  notarius config validate --config path/to/config.yml [--config-profile name] [--pipeline pipeline-id] [--only lane-a,lane-b]
+  notarius config effective --config path/to/config.yml [--config-profile name]
+  notarius config explain --config path/to/config.yml [--config-profile name] <json-pointer>
+  notarius pipelines list --config path/to/config.yml [--config-profile name] [--json]
 `
 
 type Options struct {
@@ -62,6 +66,20 @@ func Run(args []string, stdout, stderr io.Writer) int {
 }
 
 func RunWithOptions(args []string, stdout, stderr io.Writer, opts Options) int {
+	if len(args) > 0 && args[0] == "--version" {
+		if len(args) != 1 {
+			fmt.Fprintln(stderr, "notarius: --version does not accept arguments")
+			return 2
+		}
+		version, err := buildinfo.Version()
+		if err != nil {
+			fmt.Fprintf(stderr, "notarius: %v\n", err)
+			return 1
+		}
+		fmt.Fprintf(stdout, "notarius %s\n", version)
+		return 0
+	}
+
 	var err error
 	opts, err = normalizeOptions(opts)
 	if err != nil {
@@ -142,6 +160,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 	fs := flag.NewFlagSet("run", flag.ContinueOnError)
 	fs.SetOutput(io.Discard)
 	configPath := fs.String("config", "", "config file path")
+	configProfile := fs.String("config-profile", "", "configuration profile")
 	inputPath := fs.String("input", "", "source input file path")
 	onlyRaw := fs.String("only", "", "comma-separated artifact lanes")
 	outputDir := fs.String("output-dir", "", "output directory")
@@ -161,7 +180,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 	fs.Var(&llmProfile, "llm-profile", "LLM profile override")
 	fs.Var(&reasoningEffort, "reasoning-effort", "reasoning effort override")
 	fs.Var(&chunkCache, "chunk_cache", "chunk plan cache mode: auto, bypass, or refresh")
-	fs.Var(&referenceFlags, "reference", "reference binding, as slot=path, chunk.slot=path, merge.slot=path, lane.slot=path, lane.extract.slot=path, lane.merge.slot=path, or lane.normalize.slot=path")
+	fs.Var(&referenceFlags, "reference", "reference binding, as slot=path, chunk.slot=path, lane.slot=path, lane.extract.slot=path, lane.merge.slot=path, or lane.normalize.slot=path")
 	fs.Var(&withoutReferenceFlags, "without-reference", "unbind a reference, using the same selector forms as --reference")
 	fs.Var(&recomputeStep, "recompute-step", "recompute one ordered pipeline step and dependent lanes")
 	if err := validateRunFlagValues(args); err != nil {
@@ -249,20 +268,12 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 		return 2
 	}
 
-	cfg, loadedConfigPath, err := loadConfig(*configPath, opts)
+	loaded, loadedConfigPath, err := loadConfigKit(*configPath, *configProfile, runConfigurationOverrides(chunkCache, *outputDir, *debugDir), opts)
 	if err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
+		writeConfigLoadError(stderr, err)
 		return 1
 	}
-	if chunkCache.set {
-		cfg.Cache.ChunkPlans.Mode = chunkCache.value
-	}
-	if dir := strings.TrimSpace(*outputDir); dir != "" {
-		cfg.Output.Directory = dir
-	}
-	if dir := strings.TrimSpace(*debugDir); dir != "" {
-		cfg.Debug.Directory = dir
-	}
+	cfg := loaded.RuntimeConfig()
 	if err := cfg.Validate(); err != nil {
 		fmt.Fprintf(stderr, "notarius: %v\n", err)
 		return 1
@@ -320,6 +331,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 		InputPath:               strings.TrimSpace(*inputPath),
 		ConfigPath:              loadedConfigPath,
 		ConfigSource:            configSource(*configPath),
+		ConfigProfile:           effectiveConfigProfile(*configProfile),
 		OnlyLanes:               append([]string(nil), only...),
 		ChunkCacheOverride:      chunkCache.explicitValue(),
 		ReasoningEffortOverride: runtimeOverrides.ReasoningEffort,
@@ -359,7 +371,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 	if err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("resolve working directory: %w", err))
 	}
-	materialized, referenceWarnings, err := pipeline.MaterializeReferences(effective.ResolvedPipeline, catalog, pipeline.ReferenceMaterializationOptions{
+	materialized, referenceDiagnostics, err := pipeline.MaterializeReferences(effective.ResolvedPipeline, catalog, pipeline.ReferenceMaterializationOptions{
 		ConfigPath: loadedConfigPath,
 		WorkingDir: workingDir,
 	})
@@ -378,7 +390,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 	if err := writeSummary(summary, func() error { return summary.WriteInvocation(invocation) }); err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("write debug invocation metadata: %w", err))
 	}
-	if err := writeSummary(summary, func() error { return summary.WriteRedactedEffectiveConfig(effective) }); err != nil {
+	if err := writeSummary(summary, func() error { return summary.WriteEffectiveConfig(loaded) }); err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("write debug effective config: %w", err))
 	}
 	if err := writeSummary(summary, func() error { return summary.WriteResolvedPipeline(effective) }); err != nil {
@@ -446,7 +458,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 		StartedAt:        startedAt,
 		LLMProfiles:      llmProfiles,
 		Metadata:         runMetadata(effective.Config.Output.Directory, debugPath),
-		Warnings:         referenceWarnings,
+		Diagnostics:      referenceDiagnostics,
 		ChunkCacheMode:   effective.Config.Cache.ChunkPlans.Mode,
 		ChunkPlans:       chunkPlans,
 		Checkpoints:      checkpointRecorder,
@@ -455,7 +467,10 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 		Debug:            debugRecorder,
 		ExtractWorkers:   cfg.Concurrency.StageWorkers["extract"],
 	})
-	commandState.observeOutput(output)
+	diagnosticProjection, diagnosticErr := contracts.ProjectDiagnosticCollection(output.Diagnostics)
+	if diagnosticErr == nil {
+		commandState.observeOutput(output, diagnosticProjection)
+	}
 	if err != nil {
 		primaryErr := fmt.Errorf("run pipeline %q: %w", pipelineID, err)
 		if output.Manifest.PipelineID != "" {
@@ -463,15 +478,21 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 				return failPipelineCommand(stderr, commandState, terminalWriter, primaryErr, fmt.Errorf("write debug summary: %w", summaryErr))
 			}
 		}
+		if diagnosticErr != nil {
+			return failPipelineCommand(stderr, commandState, terminalWriter, primaryErr, fmt.Errorf("summarize run diagnostics: %w", diagnosticErr))
+		}
 		return failPipelineCommand(stderr, commandState, terminalWriter, primaryErr)
 	}
+	if diagnosticErr != nil {
+		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("summarize run diagnostics: %w", diagnosticErr))
+	}
 
 	if err := writePartialSummary(summary, output); err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("write debug summary: %w", err))
 	}
 	var encodedResult []byte
 	if *machineOutput {
-		result, err := newRunResult(effective.ResolvedPipeline, output, runOutputDir, debugPath)
+		result, err := newRunResultWithDiagnostics(effective.ResolvedPipeline, output, runOutputDir, debugPath, diagnosticProjection)
 		if err != nil {
 			return failPipelineCommand(stderr, commandState, terminalWriter, err)
 		}
@@ -497,12 +518,25 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 			fmt.Fprintf(stdout, "debug=%s\n", debugPath)
 		}
 	}
-	if len(output.Warnings) > 0 {
-		fmt.Fprintf(stderr, "notarius: run completed with %d warning(s)\n", len(output.Warnings))
+	if warningGroups := len(diagnosticProjection.Warnings); warningGroups > 0 {
+		fmt.Fprintf(stderr, "notarius: run completed with %d warning group(s), %d occurrence(s)", warningGroups, diagnosticProjection.WarningOccurrenceCount)
+		if warningFile, ok := logicalOutputFile(output.OutputFiles, "warnings.json"); ok {
+			fmt.Fprintf(stderr, "; details=%s", filepath.Join(runOutputDir, warningFile))
+		}
+		fmt.Fprintln(stderr)
 	}
 	return 0
 }
 
+func logicalOutputFile(files []contracts.OutputFile, name string) (string, bool) {
+	for _, file := range files {
+		if file.Name == name {
+			return file.Name, true
+		}
+	}
+	return "", false
+}
+
 func writeSummary(summary *debugbundle.SummaryWriter, write func() error) error {
 	if summary == nil {
 		return nil
@@ -522,7 +556,7 @@ func writePartialSummary(summary *debugbundle.SummaryWriter, output pipeline.Run
 			return err
 		}
 	}
-	if err := summary.WriteWarnings(output.Warnings); err != nil {
+	if err := summary.WriteDiagnostics(output.Diagnostics); err != nil {
 		return err
 	}
 	return summary.WriteCheckpointEvents(output.CheckpointEvents)
@@ -843,13 +877,39 @@ func reorderRunArgs(args []string) []string {
 
 func runFlagTakesValue(arg string) bool {
 	switch arg {
-	case "--config", "--input", "--only", "--output-dir", "--debug-dir", "--llm-profile", "--session-id", "--reasoning-effort", "--chunk_cache", "--reference", "--without-reference", "--recompute-step":
+	case "--config", "--config-profile", "--input", "--only", "--output-dir", "--debug-dir", "--llm-profile", "--session-id", "--reasoning-effort", "--chunk_cache", "--reference", "--without-reference", "--recompute-step":
 		return true
 	default:
 		return false
 	}
 }
 
+func runConfigurationOverrides(chunkCache chunkCacheFlag, outputDir, debugDir string) []configkit.Override {
+	overrides := make([]configkit.Override, 0, 3)
+	if chunkCache.set {
+		overrides = append(overrides, configkit.Override{
+			Path:  configkit.Keys("cache", "chunk_plans", "mode"),
+			Value: string(chunkCache.value),
+			Flag:  "--chunk_cache",
+		})
+	}
+	if outputDir = strings.TrimSpace(outputDir); outputDir != "" {
+		overrides = append(overrides, configkit.Override{
+			Path:  configkit.Keys("output", "directory"),
+			Value: outputDir,
+			Flag:  "--output-dir",
+		})
+	}
+	if debugDir = strings.TrimSpace(debugDir); debugDir != "" {
+		overrides = append(overrides, configkit.Override{
+			Path:  configkit.Keys("debug", "directory"),
+			Value: debugDir,
+			Flag:  "--debug-dir",
+		})
+	}
+	return overrides
+}
+
 type chunkCacheFlag struct {
 	value pipeline.ChunkCacheMode
 	set   bool
@@ -903,7 +963,7 @@ func chunkPlanStoreForRun(cfg config.ChunkPlanCacheConfig, opts Options) (pipeli
 
 func validateRunFlagValues(args []string) error {
 	for i, arg := range args {
-		if arg != "--session-id" && arg != "--reasoning-effort" {
+		if arg != "--config-profile" && arg != "--session-id" && arg != "--reasoning-effort" {
 			continue
 		}
 		if i+1 >= len(args) || strings.HasPrefix(args[i+1], "-") {
@@ -913,6 +973,13 @@ func validateRunFlagValues(args []string) error {
 	return nil
 }
 
+func effectiveConfigProfile(profile string) string {
+	if profile == "" || profile == "default" {
+		return ""
+	}
+	return profile
+}
+
 func flagWasProvided(args []string, name string) bool {
 	for _, arg := range args {
 		if arg == name || strings.HasPrefix(arg, name+"=") {
@@ -988,6 +1055,10 @@ func runConfig(args []string, stdout, stderr io.Writer, opts Options) int {
 	switch args[0] {
 	case "validate":
 		return runConfigValidate(args[1:], stdout, stderr, opts)
+	case "effective":
+		return runConfigEffective(args[1:], stdout, stderr, opts)
+	case "explain":
+		return runConfigExplain(args[1:], stdout, stderr, opts)
 	default:
 		fmt.Fprintf(stderr, "notarius: unknown config subcommand %q\n", args[0])
 		writeUsage(stderr)
@@ -995,67 +1066,6 @@ func runConfig(args []string, stdout, stderr io.Writer, opts Options) int {
 	}
 }
 
-func runConfigValidate(args []string, stdout, stderr io.Writer, opts Options) int {
-	fs := flag.NewFlagSet("config validate", flag.ContinueOnError)
-	fs.SetOutput(io.Discard)
-	configPath := fs.String("config", "", "config file path")
-	pipelineID := fs.String("pipeline", "", "pipeline ID")
-	onlyRaw := fs.String("only", "", "comma-separated artifact lanes")
-	if err := fs.Parse(args); err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 2
-	}
-	if fs.NArg() != 0 {
-		fmt.Fprintf(stderr, "notarius: unexpected argument %q\n", fs.Arg(0))
-		return 2
-	}
-	if strings.TrimSpace(*onlyRaw) != "" && strings.TrimSpace(*pipelineID) == "" {
-		fmt.Fprintln(stderr, "notarius: --only requires --pipeline")
-		return 2
-	}
-	only, err := parseOnly(*onlyRaw)
-	if err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 2
-	}
-
-	cfg, path, err := loadConfig(*configPath, opts)
-	if err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 1
-	}
-
-	if strings.TrimSpace(*pipelineID) != "" {
-		catalog, err := effectiveCatalog(opts)
-		if err != nil {
-			fmt.Fprintf(stderr, "notarius: %v\n", err)
-			return 1
-		}
-		effective, err := cfg.Resolve(config.ResolveInput{
-			PipelineID: *pipelineID,
-			Only:       only,
-			Catalog:    catalog,
-		})
-		if err != nil {
-			fmt.Fprintf(stderr, "notarius: %v\n", err)
-			return 1
-		}
-		if err := validateExplicitPromptKitProfiles(context.Background(), effective.Config, effectiveLLMProfileIDs(effective.ResolvedPipeline), opts.promptKitAssets); err != nil {
-			fmt.Fprintf(stderr, "notarius: %v\n", err)
-			return 1
-		}
-		fmt.Fprintf(stdout, "config %q is valid for pipeline %q\n", path, strings.TrimSpace(*pipelineID))
-		return 0
-	}
-
-	if err := cfg.Validate(); err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 1
-	}
-	fmt.Fprintf(stdout, "config %q is valid\n", path)
-	return 0
-}
-
 func runPipelines(args []string, stdout, stderr io.Writer, opts Options) int {
 	if len(args) == 0 {
 		fmt.Fprintln(stderr, "notarius: pipelines requires a subcommand")
@@ -1072,70 +1082,6 @@ func runPipelines(args []string, stdout, stderr io.Writer, opts Options) int {
 	}
 }
 
-func runPipelinesList(args []string, stdout, stderr io.Writer, opts Options) int {
-	fs := flag.NewFlagSet("pipelines list", flag.ContinueOnError)
-	fs.SetOutput(io.Discard)
-	configPath := fs.String("config", "", "config file path")
-	jsonOutput := fs.Bool("json", false, "write JSON output")
-	if err := fs.Parse(args); err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 2
-	}
-	if fs.NArg() != 0 {
-		fmt.Fprintf(stderr, "notarius: unexpected argument %q\n", fs.Arg(0))
-		return 2
-	}
-
-	cfg, _, err := loadConfig(*configPath, opts)
-	if err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 1
-	}
-	if err := cfg.Validate(); err != nil {
-		fmt.Fprintf(stderr, "notarius: %v\n", err)
-		return 1
-	}
-
-	ids := sortedPipelineIDs(cfg)
-	if *jsonOutput {
-		payload := struct {
-			Pipelines []string `json:"pipelines"`
-		}{Pipelines: ids}
-		encoded, err := json.Marshal(payload)
-		if err != nil {
-			fmt.Fprintf(stderr, "notarius: marshal pipeline list: %v\n", err)
-			return 1
-		}
-		fmt.Fprintf(stdout, "%s\n", encoded)
-		return 0
-	}
-
-	for _, id := range ids {
-		fmt.Fprintln(stdout, id)
-	}
-	return 0
-}
-
-func loadConfig(configPath string, opts Options) (config.Config, string, error) {
-	path, err := discoverConfigPath(configPath, opts)
-	if err != nil {
-		return config.Config{}, "", err
-	}
-
-	fileCfg, err := config.LoadFileConfig(path)
-	if err != nil {
-		return config.Config{}, "", err
-	}
-	cfg := config.Default()
-	if err := cfg.ApplyFileConfigWithLookup(fileCfg, opts.LookupEnv); err != nil {
-		return config.Config{}, "", err
-	}
-	if err := cfg.ApplyEnvOverridesWithLookup(opts.LookupEnv); err != nil {
-		return config.Config{}, "", err
-	}
-	return cfg, path, nil
-}
-
 func discoverConfigPath(configPath string, opts Options) (string, error) {
 	if path := strings.TrimSpace(configPath); path != "" {
 		if err := requireConfigFile(path); err != nil {
@@ -1249,11 +1195,21 @@ type cliReferenceUnbindRequest struct {
 }
 
 type cliReferenceSelector struct {
+	Scope    cliReferenceSelectorScope
 	LaneID   string
 	Stage    pipeline.ModuleStage
 	SlotName string
 }
 
+type cliReferenceSelectorScope uint8
+
+const (
+	cliReferenceScopePipeline cliReferenceSelectorScope = iota
+	cliReferenceScopeLane
+	cliReferenceScopeChunk
+	cliReferenceScopeBinding
+)
+
 func parseReferenceFlags(values []string) ([]cliReferenceRequest, error) {
 	if len(values) == 0 {
 		return nil, nil
@@ -1262,7 +1218,7 @@ func parseReferenceFlags(values []string) ([]cliReferenceRequest, error) {
 	for _, raw := range values {
 		name, source, ok := strings.Cut(raw, "=")
 		if !ok {
-			return nil, fmt.Errorf("--reference must use slot=path or lane.slot=path")
+			return nil, fmt.Errorf("--reference must use slot=path, lane.slot=path, or lane.stage.slot=path")
 		}
 		if strings.TrimSpace(source) == "" {
 			return nil, fmt.Errorf("--reference path must not be empty; use --without-reference to unbind")
@@ -1312,17 +1268,14 @@ func parseReferenceSelector(raw string, flagName string) (cliReferenceSelector,
 	}
 	switch len(parts) {
 	case 1:
-		return cliReferenceSelector{SlotName: strings.TrimSpace(parts[0])}, nil
+		return cliReferenceSelector{Scope: cliReferenceScopePipeline, SlotName: strings.TrimSpace(parts[0])}, nil
 	case 2:
 		first := strings.TrimSpace(parts[0])
 		slotName := strings.TrimSpace(parts[1])
 		if first == string(pipeline.StageChunk) {
-			return cliReferenceSelector{Stage: pipeline.StageChunk, SlotName: slotName}, nil
+			return cliReferenceSelector{Scope: cliReferenceScopeChunk, Stage: pipeline.StageChunk, SlotName: slotName}, nil
 		}
-		if first == string(pipeline.StageMerge) {
-			return cliReferenceSelector{Stage: pipeline.StageMerge, SlotName: slotName}, nil
-		}
-		return cliReferenceSelector{LaneID: first, SlotName: slotName}, nil
+		return cliReferenceSelector{Scope: cliReferenceScopeLane, LaneID: first, SlotName: slotName}, nil
 	case 3:
 		laneID := strings.TrimSpace(parts[0])
 		stage := pipeline.ModuleStage(strings.TrimSpace(parts[1]))
@@ -1330,9 +1283,9 @@ func parseReferenceSelector(raw string, flagName string) (cliReferenceSelector,
 		if stage != pipeline.StageExtract && stage != pipeline.StageMerge && stage != pipeline.StageNormalize {
 			return cliReferenceSelector{}, fmt.Errorf("%s lane-qualified selector must use lane.extract.slot, lane.merge.slot, or lane.normalize.slot", flagName)
 		}
-		return cliReferenceSelector{LaneID: laneID, Stage: stage, SlotName: slotName}, nil
+		return cliReferenceSelector{Scope: cliReferenceScopeBinding, LaneID: laneID, Stage: stage, SlotName: slotName}, nil
 	default:
-		return cliReferenceSelector{}, fmt.Errorf("%s must use slot, chunk.slot, merge.slot, lane.slot, lane.extract.slot, lane.merge.slot, or lane.normalize.slot", flagName)
+		return cliReferenceSelector{}, fmt.Errorf("%s must use slot, chunk.slot, lane.slot, lane.extract.slot, lane.merge.slot, or lane.normalize.slot", flagName)
 	}
 }
 
@@ -1353,37 +1306,153 @@ func resolveCLIReferenceRequests(
 		return nil, nil, err
 	}
 
-	overrides := make([]pipeline.ReferenceBinding, 0, len(referenceRequests))
+	// Broad CLI selectors are only presentation syntax. Collapse them into one
+	// highest-specificity action per concrete framework target before pipeline
+	// resolution so the generic reference contract stays stage-and-lane exact.
+	actions := make(map[cliReferenceTargetKey]resolvedCLIReferenceAction)
 	for _, request := range referenceRequests {
-		target, err := resolveCLIReferenceTarget(targets, request.Selector)
+		matches, err := resolveCLIReferenceTargets(targets, request.Selector)
 		if err != nil {
 			return nil, nil, err
 		}
-		overrides = append(overrides, pipeline.ReferenceBinding{
-			Stage:         target.stage,
-			LaneID:        target.laneID,
-			SlotName:      request.Selector.SlotName,
-			Source:        request.Source,
-			BindingSource: contracts.ReferenceBindingSourceCLI,
-		})
+		for _, target := range matches {
+			candidate := resolvedCLIReferenceAction{
+				kind:        cliReferenceActionBind,
+				selector:    request.Selector,
+				target:      target,
+				slotName:    request.Selector.SlotName,
+				source:      request.Source,
+				specificity: request.Selector.specificity(),
+			}
+			if err := mergeCLIReferenceAction(actions, candidate); err != nil {
+				return nil, nil, err
+			}
+		}
 	}
-
-	unbinds := make([]pipeline.ReferenceUnbind, 0, len(unbindRequests))
 	for _, request := range unbindRequests {
-		target, err := resolveCLIReferenceTarget(targets, request.Selector)
+		matches, err := resolveCLIReferenceTargets(targets, request.Selector)
 		if err != nil {
 			return nil, nil, err
 		}
-		unbinds = append(unbinds, pipeline.ReferenceUnbind{
-			Stage:    target.stage,
-			LaneID:   target.laneID,
-			SlotName: request.Selector.SlotName,
-		})
+		for _, target := range matches {
+			candidate := resolvedCLIReferenceAction{
+				kind:        cliReferenceActionUnbind,
+				selector:    request.Selector,
+				target:      target,
+				slotName:    request.Selector.SlotName,
+				specificity: request.Selector.specificity(),
+			}
+			if err := mergeCLIReferenceAction(actions, candidate); err != nil {
+				return nil, nil, err
+			}
+		}
 	}
 
+	resolved := make([]resolvedCLIReferenceAction, 0, len(actions))
+	for _, action := range actions {
+		resolved = append(resolved, action)
+	}
+	sort.Slice(resolved, func(i, j int) bool {
+		left, right := resolved[i], resolved[j]
+		if left.target.laneID != right.target.laneID {
+			return left.target.laneID < right.target.laneID
+		}
+		if left.target.stage != right.target.stage {
+			return referenceStageOrder(left.target.stage) < referenceStageOrder(right.target.stage)
+		}
+		return left.slotName < right.slotName
+	})
+
+	overrides := make([]pipeline.ReferenceBinding, 0, len(resolved))
+	unbinds := make([]pipeline.ReferenceUnbind, 0, len(resolved))
+	for _, action := range resolved {
+		switch action.kind {
+		case cliReferenceActionBind:
+			overrides = append(overrides, pipeline.ReferenceBinding{
+				Stage:         action.target.stage,
+				LaneID:        action.target.laneID,
+				SlotName:      action.slotName,
+				Source:        action.source,
+				BindingSource: contracts.ReferenceBindingSourceCLI,
+			})
+		case cliReferenceActionUnbind:
+			unbinds = append(unbinds, pipeline.ReferenceUnbind{
+				Stage:    action.target.stage,
+				LaneID:   action.target.laneID,
+				SlotName: action.slotName,
+			})
+		}
+	}
 	return overrides, unbinds, nil
 }
 
+type cliReferenceActionKind uint8
+
+const (
+	cliReferenceActionBind cliReferenceActionKind = iota
+	cliReferenceActionUnbind
+)
+
+type cliReferenceTargetKey struct {
+	stage    pipeline.ModuleStage
+	laneID   string
+	slotName string
+}
+
+type resolvedCLIReferenceAction struct {
+	kind        cliReferenceActionKind
+	selector    cliReferenceSelector
+	target      selectedReferenceTarget
+	slotName    string
+	source      string
+	specificity int
+}
+
+func mergeCLIReferenceAction(actions map[cliReferenceTargetKey]resolvedCLIReferenceAction, candidate resolvedCLIReferenceAction) error {
+	key := cliReferenceTargetKey{stage: candidate.target.stage, laneID: candidate.target.laneID, slotName: candidate.slotName}
+	current, ok := actions[key]
+	if !ok || candidate.specificity > current.specificity {
+		actions[key] = candidate
+		return nil
+	}
+	if candidate.specificity < current.specificity {
+		return nil
+	}
+	if candidate.kind != current.kind {
+		return fmt.Errorf("reference target %q slot %q is both bound by %q and unbound by %q at the same specificity", targetLabel(candidate.target), candidate.slotName, current.selector.String(), candidate.selector.String())
+	}
+	actions[key] = candidate
+	return nil
+}
+
+func (selector cliReferenceSelector) specificity() int {
+	switch selector.Scope {
+	case cliReferenceScopePipeline:
+		return 0
+	case cliReferenceScopeLane:
+		return 1
+	case cliReferenceScopeChunk, cliReferenceScopeBinding:
+		return 2
+	default:
+		return -1
+	}
+}
+
+func (selector cliReferenceSelector) String() string {
+	switch selector.Scope {
+	case cliReferenceScopePipeline:
+		return selector.SlotName
+	case cliReferenceScopeLane:
+		return selector.LaneID + "." + selector.SlotName
+	case cliReferenceScopeChunk:
+		return "chunk." + selector.SlotName
+	case cliReferenceScopeBinding:
+		return selector.LaneID + "." + string(selector.Stage) + "." + selector.SlotName
+	default:
+		return selector.SlotName
+	}
+}
+
 type selectedReferenceTarget struct {
 	laneID string
 	stage  pipeline.ModuleStage
@@ -1600,114 +1669,77 @@ func referenceSlotSet(slots []contracts.ReferenceSlot) map[string]struct{} {
 	return slotSet
 }
 
-func resolveCLIReferenceTarget(targets []selectedReferenceTarget, selector cliReferenceSelector) (selectedReferenceTarget, error) {
+func resolveCLIReferenceTargets(targets []selectedReferenceTarget, selector cliReferenceSelector) ([]selectedReferenceTarget, error) {
 	slotName := strings.TrimSpace(selector.SlotName)
 	if slotName == "" {
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot must not be empty")
+		return nil, fmt.Errorf("reference slot must not be empty")
 	}
-	if selector.Stage == pipeline.StageChunk {
+	switch selector.Scope {
+	case cliReferenceScopePipeline:
+		matches := make([]selectedReferenceTarget, 0, len(targets))
+		for _, target := range targets {
+			if _, ok := target.slots[slotName]; ok {
+				matches = append(matches, target)
+			}
+		}
+		if len(matches) == 0 {
+			return nil, fmt.Errorf("reference slot %q is not declared by any selected target", slotName)
+		}
+		return matches, nil
+	case cliReferenceScopeChunk:
 		for _, target := range targets {
 			if target.stage != pipeline.StageChunk {
 				continue
 			}
 			if _, ok := target.slots[slotName]; !ok {
-				return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is not declared by chunk module %q", slotName, target.module)
+				return nil, fmt.Errorf("reference slot %q is not declared by chunk module %q", slotName, target.module)
 			}
-			return target, nil
-		}
-		return selectedReferenceTarget{}, fmt.Errorf("reference chunk target is not selected")
-	}
-	if selector.Stage == pipeline.StageExtract || selector.Stage == pipeline.StageMerge || selector.Stage == pipeline.StageNormalize {
-		if selector.LaneID == "" && selector.Stage == pipeline.StageMerge {
-			return resolveCLIReferenceStageTarget(targets, selector.Stage, slotName)
+			return []selectedReferenceTarget{target}, nil
 		}
+		return nil, fmt.Errorf("reference chunk target is not selected")
+	case cliReferenceScopeLane:
+		laneSelected := false
+		matches := make([]selectedReferenceTarget, 0, 3)
 		for _, target := range targets {
-			if target.laneID == selector.LaneID && target.stage == selector.Stage {
-				if _, ok := target.slots[slotName]; !ok {
-					return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is not declared by selected %s target %q", slotName, selector.Stage, targetLabel(target))
-				}
-				return target, nil
+			if target.laneID != selector.LaneID {
+				continue
+			}
+			laneSelected = true
+			if _, ok := target.slots[slotName]; ok {
+				matches = append(matches, target)
 			}
 		}
-		return selectedReferenceTarget{}, fmt.Errorf("reference lane %q is not selected", selector.LaneID)
-	}
-	if strings.TrimSpace(selector.LaneID) != "" {
-		return resolveCLIReferenceLaneTarget(targets, strings.TrimSpace(selector.LaneID), slotName)
-	}
-	return resolveCLIReferenceFlatTarget(targets, slotName)
-}
-
-func resolveCLIReferenceStageTarget(targets []selectedReferenceTarget, stage pipeline.ModuleStage, slotName string) (selectedReferenceTarget, error) {
-	matches := make([]selectedReferenceTarget, 0, 2)
-	for _, target := range targets {
-		if target.stage != stage {
-			continue
+		if !laneSelected {
+			return nil, fmt.Errorf("reference lane %q is not selected", selector.LaneID)
 		}
-		if _, ok := target.slots[slotName]; ok {
-			matches = append(matches, target)
+		if len(matches) == 0 {
+			return nil, fmt.Errorf("reference slot %q is not declared by selected lane %q", slotName, selector.LaneID)
 		}
-	}
-	switch len(matches) {
-	case 0:
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is not declared by any selected %s target", slotName, stage)
-	case 1:
-		return matches[0], nil
+		return matches, nil
+	case cliReferenceScopeBinding:
+		laneSelected := false
+		for _, target := range targets {
+			if target.laneID != selector.LaneID {
+				continue
+			}
+			laneSelected = true
+			if target.stage != selector.Stage {
+				continue
+			}
+			if _, ok := target.slots[slotName]; !ok {
+				return nil, fmt.Errorf("reference slot %q is not declared by selected %s target %q", slotName, selector.Stage, targetLabel(target))
+			}
+			return []selectedReferenceTarget{target}, nil
+		}
+		if !laneSelected {
+			return nil, fmt.Errorf("reference lane %q is not selected", selector.LaneID)
+		}
+		return nil, fmt.Errorf("reference %s target is not selected for lane %q", selector.Stage, selector.LaneID)
 	default:
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is declared by multiple selected %s targets (%s); use a more specific selector such as %s", slotName, stage, targetList(matches), selectorSuggestions(matches, slotName))
+		return nil, fmt.Errorf("reference selector has unknown scope")
 	}
 }
 
-func resolveCLIReferenceLaneTarget(targets []selectedReferenceTarget, laneID string, slotName string) (selectedReferenceTarget, error) {
-	laneSelected := false
-	matches := make([]selectedReferenceTarget, 0, 2)
-	for _, target := range targets {
-		if target.laneID != laneID {
-			continue
-		}
-		laneSelected = true
-		if _, ok := target.slots[slotName]; ok {
-			matches = append(matches, target)
-		}
-	}
-	if !laneSelected {
-		return selectedReferenceTarget{}, fmt.Errorf("reference lane %q is not selected", laneID)
-	}
-	switch len(matches) {
-	case 0:
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is not declared by selected lane %q", slotName, laneID)
-	case 1:
-		return matches[0], nil
-	default:
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is declared by multiple selected targets in lane %q (%s); use a more specific selector such as %s", slotName, laneID, targetList(matches), selectorSuggestions(matches, slotName))
-	}
-}
-
-func resolveCLIReferenceFlatTarget(targets []selectedReferenceTarget, slotName string) (selectedReferenceTarget, error) {
-	matches := make([]selectedReferenceTarget, 0, 2)
-	for _, target := range targets {
-		if _, ok := target.slots[slotName]; ok {
-			matches = append(matches, target)
-		}
-	}
-	switch len(matches) {
-	case 0:
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is not declared by any selected reference target", slotName)
-	case 1:
-		return matches[0], nil
-	default:
-		return selectedReferenceTarget{}, fmt.Errorf("reference slot %q is declared by multiple selected targets (%s); use a more specific selector such as %s", slotName, targetList(matches), selectorSuggestions(matches, slotName))
-	}
-}
-
-func targetList(targets []selectedReferenceTarget) string {
-	labels := make([]string, 0, len(targets))
-	for _, target := range targets {
-		labels = append(labels, targetLabel(target))
-	}
-	sort.Strings(labels)
-	return strings.Join(labels, ", ")
-}
-
 func targetLabel(target selectedReferenceTarget) string {
 	if target.stage == pipeline.StageChunk {
 		return "chunk"
@@ -1715,17 +1747,19 @@ func targetLabel(target selectedReferenceTarget) string {
 	return target.laneID + "." + string(target.stage)
 }
 
-func selectorSuggestions(targets []selectedReferenceTarget, slotName string) string {
-	suggestions := make([]string, 0, len(targets))
-	for _, target := range targets {
-		if target.stage == pipeline.StageChunk {
-			suggestions = append(suggestions, "chunk."+slotName)
-			continue
-		}
-		suggestions = append(suggestions, target.laneID+"."+string(target.stage)+"."+slotName)
+func referenceStageOrder(stage pipeline.ModuleStage) int {
+	switch stage {
+	case pipeline.StageChunk:
+		return 0
+	case pipeline.StageExtract:
+		return 1
+	case pipeline.StageMerge:
+		return 2
+	case pipeline.StageNormalize:
+		return 3
+	default:
+		return 4
 	}
-	sort.Strings(suggestions)
-	return strings.Join(suggestions, " or ")
 }
 
 func sortedPipelineIDs(cfg config.Config) []string {

internal/cli/run_contract_test.go

@@ -12,6 +12,7 @@ import (
 
 	"gitea.maximumdirect.net/eric/notarius/internal/core/artifacts"
 	"gitea.maximumdirect.net/eric/notarius/internal/core/config"
+	"gitea.maximumdirect.net/eric/notarius/internal/core/debugbundle"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/checkpoint"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
@@ -37,6 +38,9 @@ func TestRunControlsRejectSyntaxWithoutAllocatingState(t *testing.T) {
 		{name: "blank debug directory", args: func(roots stateTestRoots) []string {
 			return []string{"run", "sample", "--config", roots.config, "--input", roots.input, "--debug-dir", ""}
 		}},
+		{name: "missing configuration profile", args: func(roots stateTestRoots) []string {
+			return []string{"run", "sample", "--config", roots.config, "--input", roots.input, "--config-profile"}
+		}},
 		{name: "debug directory without debug", args: func(roots stateTestRoots) []string {
 			return []string{"run", "sample", "--config", roots.config, "--input", roots.input, "--debug-dir", filepath.Join(filepath.Dir(roots.debug), "requested-debug")}
 		}},
@@ -79,7 +83,7 @@ func TestRecomputeStepCLIContract(t *testing.T) {
 		{
 			name: "explicit step",
 			configure: func(t *testing.T, roots stateTestRoots) {
-				replaceStateTestConfigLine(t, roots.config, "    artifacts:\n      items:\n        extract: test/extract\n        merge: test/merge\n        normalize: test/normalize\n", "    steps:\n      - id: chosen\n        artifacts:\n          items:\n            extract: test/extract\n            merge: test/merge\n            normalize: test/normalize\n")
+				replaceStateTestConfigLine(t, roots.config, "    artifacts:\n      items:\n        extract:\n          module: test/extract\n        merge:\n          module: test/merge\n        normalize:\n          module: test/normalize\n", "    steps:\n      - id: chosen\n        artifacts:\n          items:\n            extract:\n              module: test/extract\n            merge:\n              module: test/merge\n            normalize:\n              module: test/normalize\n")
 			},
 			flags:      []string{"--resume", "--recompute-step", "chosen"},
 			wantCode:   0,
@@ -161,7 +165,7 @@ func TestRunOnlyExecutesSelectedLanes(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	data = []byte(replaceRequiredOnce(t, string(data), "    output: test/output\n", "      other:\n        extract: test/extract\n    output: test/output\n"))
+	data = []byte(replaceRequiredOnce(t, string(data), "    output:\n      module: test/output\n", "      other:\n        extract:\n          module: test/extract\n    output:\n      module: test/output\n"))
 	if err := os.WriteFile(roots.config, data, 0o600); err != nil {
 		t.Fatal(err)
 	}
@@ -180,7 +184,7 @@ func TestRunOnlyExecutesSelectedLanes(t *testing.T) {
 }
 
 func TestRunStateRootsHonorEnvironmentFlagsAndDefaults(t *testing.T) {
-	t.Run("environment roots", func(t *testing.T) {
+	t.Run("retired environment roots have no effect", func(t *testing.T) {
 		roots := newStateTestRoots(t)
 		environmentOutput := filepath.Join(t.TempDir(), "environment-output")
 		environmentDebug := filepath.Join(t.TempDir(), "environment-debug")
@@ -193,10 +197,10 @@ func TestRunStateRootsHonorEnvironmentFlagsAndDefaults(t *testing.T) {
 		if result.code != 0 {
 			t.Fatalf("code=%d stdout=%q stderr=%q", result.code, result.stdout, result.stderr)
 		}
-		assertFile(t, filepath.Join(environmentOutput, filepath.Base(onlyChildDir(t, environmentOutput)), "result.json"))
-		onlyChildDir(t, environmentDebug)
-		assertAbsent(t, roots.output)
-		assertAbsent(t, roots.debug)
+		assertFile(t, filepath.Join(roots.output, filepath.Base(onlyChildDir(t, roots.output)), "result.json"))
+		onlyChildDir(t, roots.debug)
+		assertAbsent(t, environmentOutput)
+		assertAbsent(t, environmentDebug)
 	})
 
 	t.Run("command flags override environment", func(t *testing.T) {
@@ -227,23 +231,57 @@ func TestRunStateRootsHonorEnvironmentFlagsAndDefaults(t *testing.T) {
 			t.Fatal(err)
 		}
 		text := string(data)
-		text = replaceRequiredOnce(t, text, fmt.Sprintf("  directory: %q\n", roots.output), "")
-		text = replaceRequiredOnce(t, text, fmt.Sprintf("  directory: %q\n", roots.debug), "")
+		text = replaceRequiredOnce(t, text, fmt.Sprintf("output:\n  directory: %q\n", roots.output), "")
+		text = replaceRequiredOnce(t, text, fmt.Sprintf("debug:\n  directory: %q\n", roots.debug), "")
 		if err := os.WriteFile(roots.config, []byte(text), 0o600); err != nil {
 			t.Fatal(err)
 		}
-		workDir := t.TempDir()
-		t.Chdir(workDir)
 		opts := newStateTestHarness().options()
 		result := runWithStateRoots(t, roots, opts, nil)
 		if result.code != 0 {
 			t.Fatalf("code=%d stdout=%q stderr=%q", result.code, result.stdout, result.stderr)
 		}
-		assertFile(t, filepath.Join(workDir, "notarius-output", filepath.Base(onlyChildDir(t, filepath.Join(workDir, "notarius-output"))), "result.json"))
-		onlyChildDir(t, filepath.Join(workDir, "notarius-debug"))
+		configDirectory := filepath.Dir(roots.config)
+		assertFile(t, filepath.Join(configDirectory, "notarius-output", filepath.Base(onlyChildDir(t, filepath.Join(configDirectory, "notarius-output"))), "result.json"))
+		onlyChildDir(t, filepath.Join(configDirectory, "notarius-debug"))
 	})
 }
 
+func TestRunAppliesConfigurationProfileAndRecordsSelection(t *testing.T) {
+	roots := newStateTestRoots(t)
+	profileDirectory := filepath.Join(filepath.Dir(roots.config), "profiles")
+	profileOutput := filepath.Join(profileDirectory, "profile-output")
+	profileDebug := filepath.Join(profileDirectory, "profile-debug")
+	writeConfigKitFile(t, filepath.Join(profileDirectory, "development.yml"), "output:\n  directory: ./profile-output\ndebug:\n  directory: ./profile-debug\n")
+
+	var stdout, stderr bytes.Buffer
+	code := RunWithOptions([]string{
+		"run", "sample", "--config", roots.config, "--config-profile", "development",
+		"--input", roots.input, "--chunk_cache", "bypass", "--debug",
+	}, &stdout, &stderr, newStateTestHarness().options())
+	if code != 0 || stderr.Len() != 0 {
+		t.Fatalf("code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
+	}
+	assertFile(t, filepath.Join(profileOutput, filepath.Base(onlyChildDir(t, profileOutput)), "result.json"))
+	bundle := onlyChildDir(t, profileDebug)
+	var invocation debugbundle.Invocation
+	readStateTestSummaryJSON(t, bundle, "invocation.json", &invocation)
+	if invocation.ConfigProfile != "development" {
+		t.Fatalf("config profile = %q, want development", invocation.ConfigProfile)
+	}
+}
+
+func TestEffectiveConfigProfileOmitsBaseSelection(t *testing.T) {
+	for _, profile := range []string{"", "default"} {
+		if got := effectiveConfigProfile(profile); got != "" {
+			t.Fatalf("effectiveConfigProfile(%q) = %q, want empty", profile, got)
+		}
+	}
+	if got := effectiveConfigProfile("development"); got != "development" {
+		t.Fatalf("effectiveConfigProfile(development) = %q", got)
+	}
+}
+
 func TestRunLLMProfileOverrideAndValidationUseInjectedBoundaries(t *testing.T) {
 	t.Run("one effective profile reaches the factory and modules", func(t *testing.T) {
 		roots := newStateTestRoots(t)
@@ -567,7 +605,7 @@ func TestRunFactoryAndPreparationFailuresAreProcessFailures(t *testing.T) {
 		if err != nil {
 			t.Fatal(err)
 		}
-		data = []byte(replaceRequiredOnce(t, string(data), "extract: test/extract", "extract: test/failing-extract"))
+		data = []byte(replaceRequiredOnce(t, string(data), "module: test/extract", "module: test/failing-extract"))
 		if err := os.WriteFile(roots.config, data, 0o600); err != nil {
 			t.Fatal(err)
 		}
@@ -590,10 +628,11 @@ func TestRunWarningsRemainSuccessfulAndReachDurableSurfaces(t *testing.T) {
 	roots := newStateTestRoots(t)
 	harness := newStateTestHarness()
 	harness.includeWarnings = true
-	harness.chunkWarnings = []contracts.Warning{{Scope: "chunk", ReasonCode: "contract-warning", Message: "warning retained"}}
+	harness.includeWarningFile = true
+	harness.chunkDiagnostics = []contracts.ProducerDiagnostic{stateTestDiagnostic("chunk", "contract-warning", "warning retained")}
 	var stdout, stderr bytes.Buffer
 	code := RunWithOptions([]string{"run", "sample", "--config", roots.config, "--input", roots.input, "--chunk_cache", "bypass", "--debug"}, &stdout, &stderr, harness.options())
-	if code != 0 || !strings.Contains(stdout.String(), "outputs=1") || !strings.Contains(stderr.String(), "1 warning(s)") {
+	if code != 0 || !strings.Contains(stdout.String(), "outputs=1") || !strings.Contains(stderr.String(), "1 warning group(s), 1 occurrence(s)") || !strings.Contains(stderr.String(), "warnings.json") {
 		t.Fatalf("code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
 	}
 	outputPath := filepath.Join(onlyChildDir(t, roots.output), "result.json")
@@ -601,11 +640,12 @@ func TestRunWarningsRemainSuccessfulAndReachDurableSurfaces(t *testing.T) {
 	if err != nil || !strings.Contains(string(output), "contract-warning") {
 		t.Fatalf("durable output = %q, %v", output, err)
 	}
+	assertFile(t, filepath.Join(filepath.Dir(outputPath), "warnings.json"))
 	bundle := onlyChildDir(t, roots.debug)
-	var warnings []contracts.Warning
-	readStateTestSummaryJSON(t, bundle, "warnings.json", &warnings)
-	if len(warnings) != 1 || warnings[0].ReasonCode != "contract-warning" {
-		t.Fatalf("debug warnings = %#v", warnings)
+	var diagnostics contracts.DiagnosticCollection
+	readStateTestSummaryJSON(t, bundle, "final-diagnostics.json", &diagnostics)
+	if len(diagnostics.Groups) != 1 || diagnostics.Groups[0].ReasonCode != "contract-warning" {
+		t.Fatalf("debug diagnostics = %#v", diagnostics)
 	}
 }
 

internal/cli/run_id_test.go

@@ -48,8 +48,15 @@ func TestWriteOutputFilesSupportsNestedLogicalPaths(t *testing.T) {
 		if err != nil {
 			t.Fatal(err)
 		}
-		if info.Mode().Perm() != want {
-			t.Fatalf("%s mode = %#o, want %#o", path, info.Mode().Perm(), want)
+		mode := info.Mode().Perm()
+		if mode&^want != 0 {
+			t.Fatalf("%s mode = %#o, must not be broader than %#o", path, mode, want)
+		}
+		if info.IsDir() && mode&0o700 != 0o700 {
+			t.Fatalf("%s mode = %#o, want owner access", path, mode)
+		}
+		if !info.IsDir() && mode != want {
+			t.Fatalf("%s mode = %#o, want %#o", path, mode, want)
 		}
 	}
 	entries, err := os.ReadDir(filepath.Join(runPath, "nested"))

internal/cli/run_result.go

@@ -7,25 +7,40 @@ import (
 	"path/filepath"
 	"strings"
 
+	"gitea.maximumdirect.net/eric/notarius/internal/core/artifacts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
 )
 
-const runResultSchemaVersion = "notarius.run-result.v1"
+const runResultSchemaVersion = "notarius.run-result.v2"
 
 type runResult struct {
-	SchemaVersion         string `json:"schema_version"`
-	RunID                 string `json:"run_id"`
-	PipelineID            string `json:"pipeline_id"`
-	OutputDirectory       string `json:"output_directory"`
-	IndexFile             string `json:"index_file,omitempty"`
-	NormalizedOutputCount int    `json:"normalized_output_count"`
-	RejectedOutputCount   int    `json:"rejected_output_count"`
-	WarningCount          int    `json:"warning_count"`
-	ValidationStatus      string `json:"validation_status"`
-	DebugDirectory        string `json:"debug_directory,omitempty"`
+	SchemaVersion             string                        `json:"schema_version"`
+	RunID                     string                        `json:"run_id"`
+	PipelineID                string                        `json:"pipeline_id"`
+	OutputDirectory           string                        `json:"output_directory"`
+	IndexFile                 string                        `json:"index_file,omitempty"`
+	NormalizedOutputCount     int                           `json:"normalized_output_count"`
+	RejectedOutputCount       int                           `json:"rejected_output_count"`
+	WarningGroupCount         int                           `json:"warning_group_count"`
+	WarningOccurrenceCount    int                           `json:"warning_occurrence_count"`
+	DiagnosticGroupCount      int                           `json:"diagnostic_group_count"`
+	DiagnosticOccurrenceCount int                           `json:"diagnostic_occurrence_count"`
+	DiagnosticsTruncated      bool                          `json:"diagnostics_truncated"`
+	ValidationStatus          string                        `json:"validation_status"`
+	ValidationSummaries       []artifacts.ValidationSummary `json:"validation_summaries,omitempty"`
+	DebugDirectory            string                        `json:"debug_directory,omitempty"`
 }
 
 func newRunResult(resolved pipeline.ResolvedPipeline, output pipeline.RunOutput, outputDirectory, debugDirectory string) (runResult, error) {
+	diagnosticProjection, err := contracts.ProjectDiagnosticCollection(output.Diagnostics)
+	if err != nil {
+		return runResult{}, fmt.Errorf("summarize run diagnostics: %w", err)
+	}
+	return newRunResultWithDiagnostics(resolved, output, outputDirectory, debugDirectory, diagnosticProjection)
+}
+
+func newRunResultWithDiagnostics(resolved pipeline.ResolvedPipeline, output pipeline.RunOutput, outputDirectory, debugDirectory string, diagnosticProjection contracts.DiagnosticProjection) (runResult, error) {
 	if strings.TrimSpace(output.Manifest.RunID) == "" {
 		return runResult{}, fmt.Errorf("run result requires a run ID")
 	}
@@ -51,14 +66,19 @@ func newRunResult(resolved pipeline.ResolvedPipeline, output pipeline.RunOutput,
 	}
 
 	result := runResult{
-		SchemaVersion:         runResultSchemaVersion,
-		RunID:                 output.Manifest.RunID,
-		PipelineID:            output.Manifest.PipelineID,
-		OutputDirectory:       absOutputDirectory,
-		NormalizedOutputCount: len(output.NormalizeOutputs),
-		RejectedOutputCount:   len(output.Rejected),
-		WarningCount:          len(output.Warnings),
-		ValidationStatus:      output.Manifest.ValidationStatus,
+		SchemaVersion:             runResultSchemaVersion,
+		RunID:                     output.Manifest.RunID,
+		PipelineID:                output.Manifest.PipelineID,
+		OutputDirectory:           absOutputDirectory,
+		NormalizedOutputCount:     len(output.NormalizeOutputs),
+		RejectedOutputCount:       len(output.Rejected),
+		WarningGroupCount:         len(diagnosticProjection.Warnings),
+		WarningOccurrenceCount:    diagnosticProjection.WarningOccurrenceCount,
+		DiagnosticGroupCount:      len(diagnosticProjection.Diagnostics),
+		DiagnosticOccurrenceCount: diagnosticProjection.DiagnosticOccurrenceCount,
+		DiagnosticsTruncated:      output.Diagnostics.Truncated,
+		ValidationStatus:          output.Manifest.ValidationStatus,
+		ValidationSummaries:       cloneValidationSummaries(output.Manifest.ValidationSummaries),
 	}
 
 	if strings.TrimSpace(debugDirectory) != "" {
@@ -85,6 +105,17 @@ func newRunResult(resolved pipeline.ResolvedPipeline, output pipeline.RunOutput,
 	return result, nil
 }
 
+func cloneValidationSummaries(summaries []artifacts.ValidationSummary) []artifacts.ValidationSummary {
+	if len(summaries) == 0 {
+		return nil
+	}
+	cloned := make([]artifacts.ValidationSummary, len(summaries))
+	for index, summary := range summaries {
+		cloned[index] = artifacts.CloneValidationSummary(summary)
+	}
+	return cloned
+}
+
 func encodeRunResult(result runResult) ([]byte, error) {
 	encoded, err := json.Marshal(result)
 	if err != nil {

internal/cli/run_result_command_test.go

@@ -27,7 +27,7 @@ func TestMaintainedMinimalInvocationEmitsRunResult(t *testing.T) {
 	}
 
 	receipt := decodeRunResultDocument(t, stdout.String())
-	if got := receipt["schema_version"]; got != "notarius.run-result.v1" {
+	if got := receipt["schema_version"]; got != "notarius.run-result.v2" {
 		t.Fatalf("schema_version = %q", got)
 	}
 	if got := receipt["run_id"]; got != productionRunID {
@@ -45,8 +45,8 @@ func TestMaintainedMinimalInvocationEmitsRunResult(t *testing.T) {
 	if got := receipt["rejected_output_count"]; got != float64(0) {
 		t.Fatalf("rejected_output_count = %v", got)
 	}
-	if got := receipt["warning_count"]; got != float64(0) {
-		t.Fatalf("warning_count = %v", got)
+	if got := receipt["warning_group_count"]; got != float64(0) || receipt["warning_occurrence_count"] != float64(0) || receipt["diagnostic_group_count"] != float64(0) || receipt["diagnostic_occurrence_count"] != float64(0) || receipt["diagnostics_truncated"] != false {
+		t.Fatalf("diagnostic counts = %#v", receipt)
 	}
 	if got := receipt["validation_status"]; got != "approved" {
 		t.Fatalf("validation_status = %q", got)
@@ -63,19 +63,19 @@ func TestMaintainedMinimalInvocationEmitsRunResult(t *testing.T) {
 func TestRunResultReportsWarningsAndDebugBundle(t *testing.T) {
 	roots := newStateTestRoots(t)
 	harness := newStateTestHarness()
-	harness.chunkWarnings = []contracts.Warning{{Scope: "chunk", ReasonCode: "contract-warning", Message: "warning retained"}}
+	harness.chunkDiagnostics = []contracts.ProducerDiagnostic{stateTestDiagnostic("chunk", "contract-warning", "warning retained")}
 	var stdout, stderr bytes.Buffer
 	code := RunWithOptions([]string{
 		"run", "sample", "--config", roots.config, "--input", roots.input,
 		"--chunk_cache", "bypass", "--debug", "--json",
 	}, &stdout, &stderr, harness.options())
-	if code != 0 || !strings.Contains(stderr.String(), "1 warning(s)") {
+	if code != 0 || !strings.Contains(stderr.String(), "1 warning group(s), 1 occurrence(s)") || strings.Contains(stderr.String(), "warnings.json") {
 		t.Fatalf("code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
 	}
 
 	receipt := decodeRunResultDocument(t, stdout.String())
-	if got := receipt["warning_count"]; got != float64(1) {
-		t.Fatalf("warning_count = %v", got)
+	if got := receipt["warning_group_count"]; got != float64(1) || receipt["warning_occurrence_count"] != float64(1) || receipt["diagnostic_group_count"] != float64(0) || receipt["diagnostic_occurrence_count"] != float64(0) || receipt["diagnostics_truncated"] != false {
+		t.Fatalf("diagnostic counts = %#v", receipt)
 	}
 	debugDirectory, ok := receipt["debug_directory"].(string)
 	if !ok || !filepath.IsAbs(debugDirectory) || debugDirectory != onlyChildDir(t, roots.debug) {
@@ -92,7 +92,7 @@ func TestRunResultReportsSuccessfulRejection(t *testing.T) {
 	if err != nil {
 		t.Fatal(err)
 	}
-	configBytes = []byte(replaceRequiredOnce(t, string(configBytes), "        normalize: test/normalize\n", "        normalize:\n          module: test/normalize\n          validators:\n            - generic/always_reject\n"))
+	configBytes = []byte(replaceRequiredOnce(t, string(configBytes), "        normalize:\n          module: test/normalize\n", "        normalize:\n          module: test/normalize\n          validators:\n            - module: generic/always_reject\n          validation_policy:\n            semantic_rejection: reject_output\n"))
 	if err := os.WriteFile(roots.config, configBytes, 0o600); err != nil {
 		t.Fatal(err)
 	}