Tighten evidence context source excerpts

.gitignore

@@ -2,6 +2,7 @@
 notarius
 notarius-output
 workspace/
+.codebase-memory/
 
 # ---> Go
 # If you prefer the allow list template instead of the deny list, see community template:
@@ -73,4 +74,3 @@ Icon
 Network Trash Folder
 Temporary Items
 .apdisk
-

README.md

@@ -1,34 +1,46 @@
 # Notarius
 
-Notarius is a Go CLI for extracting structured artifacts from source material
+Notarius is a Go CLI for turning source material into structured artifacts with
-with explicit, configurable pipeline modules.
+configured extraction pipelines. The implemented D&D workflow reads Seriatim
+transcript JSON and can produce NPC, location, and item registries; their
+source-grounded occurrences; scene descriptions, combat turns, enemy events,
+and spell casts.
 
-The current implementation reads Seriatim transcript JSON, chunks the source
+## Quickstart
-units, extracts D&D spell-cast artifacts with a Scriptorium-backed LLM runtime,
-and writes JSON output. Add `--debug` when a per-run inspection bundle is
-needed.
 
-```sh
+Provide an OpenRouter API key through the environment, then run the maintained
-OPENROUTER_API_KEY=... \
+minimal example:
-go run ./cmd/notarius run dnd-session \
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json
-```
 
-This invocation uses the maintained example configuration and input. See the
+~~~
-configuration and operations references for profile selection, credentials, and
+OPENROUTER_API_KEY=your-api-key \
-run artifacts.
+  go run ./cmd/notarius run dnd-session \
+    --config examples/dnd-minimal.config.yml \
+    --input examples/seriatim-minimal-transcript.json
+~~~
 
-Useful references:
+The command publishes a JSON output bundle. Its command syntax and exit
+behavior are documented in the [CLI reference](docs/cli.md); configuration,
+credentials, and module selection are owned by the
+[configuration reference](docs/config.md).
 
-- [CLI reference](docs/cli.md)
+For the complete ordered D&D workflow, use
-- [Configuration reference](docs/config.md)
+[the complete configuration](examples/dnd-complete.config.yml) with
-- [Operations](docs/operations.md)
+[its synthetic transcript](examples/dnd-complete-transcript.json). It
-- [Seriatim input contract](docs/integrations/seriatim.md)
+demonstrates all implemented D&D lanes and the supporting campaign references.
-- [JSON output contract](docs/integrations/json-output.md)
+
-- [D&D spell artifact contract](docs/integrations/dnd-spell-artifacts.md)
+## Documentation
-- [Developer guide](docs/development.md)
+
-- [Internal implementation docs](docs/internal/overview.md)
+- [CLI reference](docs/cli.md) — commands, flags, output streams, and exits.
-- [Maintained example config](examples/dnd-spells.config.yml)
+- [Configuration reference](docs/config.md) — configuration files, profiles,
-- [NPC-grounded example config](examples/dnd-npc-grounded.config.yml)
+  validation, and module selection.
-- [Maintained example input](examples/seriatim-minimal-transcript.json)
+- [Operations](docs/operations.md) — output, state, recovery, and debug
+  handling.
+- [Integration contracts](docs/integrations/) — Seriatim input and published
+  artifact formats.
+- [Subprocess consumer guide](docs/consumers/subprocess.md) — invoke Notarius
+  from an orchestrator and consume a published result.
+- [Internal overview](docs/internal/overview.md) — implemented component map
+  for maintainers.
+- [Developer guide](docs/development.md) — contributor orientation and
+  validation guidance.
+- [Future work](docs/roadmap/future.md) — unimplemented ideas and priorities.

assets/dnd/combat-turns/prompts/instructions.md

@@ -0,0 +1,18 @@
+Extract Dungeons & Dragons combat-turn artifacts from the supplied transcript.
+Include a record only when the transcript establishes that an in-world
+participant takes a combat turn or performs a discrete interrupting combat
+event. Keep events in transcript chronology; place an interrupting event where
+it occurs.
+
+Exclude initiative setup without a turn or combat event, tactical planning,
+table talk, rules lookup, hypothetical events, abandoned intentions, recaps
+outside the current passage, and downstream consequences. Do not infer combat
+events from Dungeons & Dragons rules knowledge. Preserve the session as played
+and attribute relevant nonstandard rulings to the GM or table. Unmatched actors
+remain permitted.
+
+Treat each record as one turn-level event and keep its supporting transcript
+evidence together. Use `turn` for a regular combat turn, `reaction` for an
+off-turn reaction, `legendary_action` for a legendary action,
+`lair_action` for a lair action, and `other` for another discrete combat
+event that does not fit those categories.

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

@@ -1,6 +1,6 @@
 id: dnd.combat_turns
 version: "v1"
-default_profile: gemini-2-flash
+default_profile: dnd-extraction
 inputs:
   - name: transcript
     required: true
@@ -14,34 +14,30 @@ inputs:
   - name: glossary
     required: false
     content_type: text/plain
-  - name: npcs
+  - name: npc_registry
     required: false
     content_type: application/json
 messages:
   - role: system
     content_file: ./sharedassets/common-dnd-system.md
-  - role: user
-    content_file: ./sharedassets/common-dnd-extraction-evidence.md
   - role: user
     content_file: ./sharedassets/common-dnd-identity.md
-    cache_control:
-      type: ephemeral
   - role: user
     content_file: ./sharedassets/common-dnd-references.md
     cache_control:
       type: ephemeral
   - role: user
-    content_file: ./sharedassets/common-dnd-npcs.md
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
     cache_control:
       type: ephemeral
   - role: user
-    content_file: ./task.md
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-npc-registry.md
   - role: user
     content_file: ./instructions.md
     cache_control:
       type: ephemeral
-  - role: user
-    content_file: ./sharedassets/common-dnd-transcript.md
 output:
   format: json
   validation_mode: json_schema

assets/dnd/enemy-events/prompts/combat-grounding.md

@@ -0,0 +1,12 @@
+Compact combat grounding is supplied below. It can guide attention and
+disambiguation, but it is not evidence. Do not derive an event, subject,
+outcome, or source range from either list. The current transcript alone must
+directly establish every returned event.
+
+Combat-turn grounding:
+
+{{ input "combat_turns" }}
+
+Named combat-opponent grounding:
+
+{{ input "npc_occurrences" }}

assets/dnd/enemy-events/prompts/instructions.md

@@ -0,0 +1,25 @@
+Extract Dungeons & Dragons enemy events from the supplied combat transcript.
+An `engaged` event requires direct establishment that a subject is actively
+opposing the party in combat. A `killed`, `fled`, `captured`, or
+`incapacitated` event requires explicit establishment of that outcome. An
+outcome may share evidence with an engagement, and a later engagement or
+outcome for the same subject remains a separate observation. Emit at most one
+`engaged` observation for the same subject in this combat scene.
+
+For `killed`, direct death or killing is required. For `fled`, the subject
+must explicitly escape, retreat, or leave combat to avoid continued engagement.
+For `captured`, the subject must be explicitly taken prisoner or secured
+under the party's control. For `incapacitated`, the subject must be explicitly
+unable to continue acting without being established as killed or captured.
+
+When the transcript identifies a named NPC, use its normalized registry
+spelling. A hostile creature without a registry entry is allowed. For unnamed
+individuals or groups, use only the narrowest transcript-grounded label, such
+as `Orcs`, `One orc`, or `Remaining orcs`; never invent member names, IDs,
+or quantities.
+
+Exclude party members, allies, neutral observers, mentioned-but-absent enemies,
+hazards, traps, environmental effects, uncertain allegiance, table talk,
+planning, hypotheses, recaps outside this passage, and downstream inference.
+Do not infer an engagement or outcome from initiative, turn absence, damage,
+defeat, movement, or a scene ending.

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

@@ -0,0 +1,53 @@
+id: dnd.enemy_events
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+  - name: npc_registry
+    required: true
+    content_type: application/json
+  - name: combat_turns
+    required: true
+    content_type: application/json
+  - name: npc_occurrences
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-npc-registry.md
+  - role: user
+    content_file: ./combat-grounding.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_enemy_events_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,33 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.enemy_events.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["events"],
+  "properties": {
+    "events": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "kind", "source_refs"],
+        "properties": {
+          "name": {"type": "string"},
+          "kind": {"type": "string"},
+          "source_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {"type": "integer"},
+                "end_unit_id": {"type": "integer"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,30 @@
+Extract meaningful Dungeons & Dragons item and currency occurrences: discoveries and changes
+in party possession established by the transcript. This is an occurrence history,
+not an inventory or ledger: do not calculate balances, resolve item identity
+across records, or infer ownership that the transcript does not establish.
+
+For every occurrence, use the supplied canonical item `name`. Record a stated
+quantity as an integer and leave it null when the transcript does not state
+one. Preserve the stated currency denomination through the selected canonical
+registry name.
+
+Use `discovered` when the party learns of or encounters an item without
+establishing possession. Use `acquired` when the party or a party member gains
+possession. Use `lost` when party possession ends through a gift, sale, payment,
+theft, abandonment, or destruction not caused by intended use. Use `consumed`
+when intended use depletes an expendable item. Monetary spending, purchases, and
+payments are always `lost`, not `consumed`. Classify currency as `consumed` only
+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.
+
+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
+independently established.

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

@@ -0,0 +1,6 @@
+Use the supplied item registry only to ground each occurrence. Every record
+must use one registry item's canonical `name`; do not invent, rename, merge,
+or infer registry items. The registry is not transcript evidence: cite only the
+current transcript chunk in `source_refs`.
+
+{{ input "item_registry" }}

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

@@ -0,0 +1,45 @@
+id: dnd.item_occurrences
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+  - name: item_registry
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./item-registry.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_item_occurrences_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,36 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.item_occurrences.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["occurrences"],
+  "properties": {
+    "occurrences": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "kind", "quantity", "from", "to", "source_refs"],
+        "properties": {
+          "name": {"type": "string"},
+          "kind": {"type": "string"},
+          "quantity": {"type": ["integer", "null"]},
+          "from": {"type": ["string", "null"]},
+          "to": {"type": ["string", "null"]},
+          "source_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {"type": "integer"},
+                "end_unit_id": {"type": "integer"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,12 @@
+Extract only items established by the provided Dungeons & Dragons transcript.
+
+Include named unique items, concrete reusable item types, and stable unique
+designations. Record each currency denomination separately when it is
+established, such as copper pieces, silver pieces, gold pieces, or platinum
+pieces. Do not use capitalization as an eligibility test. Keep distinct names
+and designations as separate candidates; do not merge aliases or invent
+qualifiers.
+
+Do not record vague categories such as "loot", "treasure", or "some gear";
+generic weapons; inferred properties; quantities; or inferred uniqueness. Omit
+uncertain or unsupported items.

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

@@ -1,6 +1,6 @@
-id: dnd.npcs
+id: dnd.item_registry
 version: "v1"
-default_profile: gemini-2-flash
+default_profile: dnd-extraction
 inputs:
   - name: transcript
     required: true
@@ -17,26 +17,24 @@ inputs:
 messages:
   - role: system
     content_file: ./sharedassets/common-dnd-system.md
-  - role: user
-    content_file: ./sharedassets/common-dnd-extraction-evidence.md
   - role: user
     content_file: ./sharedassets/common-dnd-identity.md
-    cache_control:
-      type: ephemeral
   - role: user
     content_file: ./sharedassets/common-dnd-references.md
     cache_control:
       type: ephemeral
   - role: user
-    content_file: ./task.md
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
   - role: user
     content_file: ./instructions.md
     cache_control:
       type: ephemeral
-  - role: user
-    content_file: ./sharedassets/common-dnd-transcript.md
 output:
   format: json
   validation_mode: json_schema
-  schema_path: dnd_npcs_llm.v1.json
+  schema_path: dnd_item_registry_llm.v1.json
   repair_attempts: 0

assets/dnd/item-registry/extract/schemas/dnd_item_registry_llm.v1.json

@@ -0,0 +1,32 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.item_registry.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["items"],
+  "properties": {
+    "items": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "source_refs"],
+        "properties": {
+          "name": {"type": "string"},
+          "source_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {"type": "integer"},
+                "end_unit_id": {"type": "integer"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,9 @@
+Determine whether candidates identify the same item type or unique designation
+using their contextual labels and cited transcript windows. Do not treat nearby
+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.

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

@@ -0,0 +1,30 @@
+id: dnd.item_registry.normalize
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: candidates
+    required: true
+    content_type: application/json
+  - name: transcript
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/protocol.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/candidates.md
+  - role: user
+    content_file: ./sharedassets/transcript-windows.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: semantic_reconciliation_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,34 @@
+Extract Dungeons & Dragons location occurrences from the supplied transcript.
+Include an occurrence only when the transcript establishes one supplied
+location, one occurrence kind, and a coherent passage supporting both.
+
+Use exactly one kind per occurrence:
+
+- visited: party members are physically present, arrive, remain, or depart;
+- planned: the party explicitly proposes, intends, or agrees to future travel;
+- recalled: the transcript explicitly recounts an earlier party visit; or
+- mentioned: the location is explicitly referenced without stronger support,
+  including non-actionable speculation or a mere hypothetical reference.
+
+A mere hypothetical or speculative reference is not planned unless the
+transcript also establishes an actual proposal, intention, or agreement to
+travel. When the hypothetical explicitly names a supplied location, it may be
+mentioned.
+
+A generic phrase in the current chunk may refer to a supplied named registry
+location only when the chunk's context supports that coreference. It must not
+create a registry location, and registry content or provenance must never
+replace current-chunk evidence.
+
+For every occurrence, return the exact selector from the location registry:
+the canonical `name`, plus an empty `registry_refs` array for a unique name or
+the complete ordered `registry_refs` array for a repeated name. Registry ranges
+and context identify the location only; they are not occurrence evidence.
+
+For overlapping support, visited outranks planned, recalled, and mentioned;
+planned outranks recalled and mentioned; recalled outranks mentioned. A passage
+may produce multiple records when it independently establishes separate facts,
+such as recalling an earlier visit while planning a return. Omit inferred,
+unstated, uncertain, or unsupported places and occurrences. Do not infer a
+location or occurrence from surrounding events when the transcript does not
+state it. Do not summarize location descriptions.

assets/dnd/location-occurrences/prompts/location-registry.md

@@ -0,0 +1,11 @@
+A contextual location registry is provided below for identity grounding. It may
+be empty. Every record supplies a canonical display name. A name that appears
+once is selected with that name and an empty `registry_refs` array. A repeated
+name is selected only by copying both its name and its complete, ordered
+`registry_refs` array exactly as supplied.
+
+Registry content is context, not occurrence evidence. Do not derive an
+occurrence or `source_refs` range from the registry. Do not invent a location
+or selector that is absent from it.
+
+{{ input "location_registry" }}

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

@@ -0,0 +1,45 @@
+id: dnd.location_occurrences
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+  - name: location_registry
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./location-registry.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_location_occurrences_llm.v1.json
+  repair_attempts: 0

assets/dnd/location-occurrences/schemas/dnd_location_occurrences_llm.v1.json

@@ -0,0 +1,45 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.location_occurrences.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["occurrences"],
+  "properties": {
+    "occurrences": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "registry_refs", "kind", "source_refs"],
+        "properties": {
+          "name": {"type": "string"},
+          "registry_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {"type": "integer", "minimum": 1},
+                "end_unit_id": {"type": "integer", "minimum": 1}
+              }
+            }
+          },
+          "kind": {"enum": ["visited", "planned", "recalled", "mentioned"]},
+          "source_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {"type": "integer"},
+                "end_unit_id": {"type": "integer"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,13 @@
+Extract only physical places established by the provided Dungeons & Dragons
+transcript that have a stable proper name or unique in-world designation. This
+includes named planes, regions, settlements, districts, buildings, rooms,
+landmarks, routes, and geographic features.
+
+Do not create a registry location for generic, temporary, relative, or merely
+descriptive phrases, including "the room", "the bar", "the hallway",
+"outside", and "upstairs". Do not use capitalization as an eligibility test.
+Keep aliases and nested places when the transcript identifies them; do not merge
+or invent qualifiers for similarly named places.
+
+Exclude people, creatures, objects, organizations, abstract concepts, and
+places merely inferred from an event. Omit uncertain or unsupported places.

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

@@ -0,0 +1,40 @@
+id: dnd.location_registry
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_location_registry_llm.v1.json
+  repair_attempts: 0

assets/dnd/location-registry/extract/schemas/dnd_location_registry_llm.v1.json

@@ -0,0 +1,32 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.location_registry.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["locations"],
+  "properties": {
+    "locations": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "source_refs"],
+        "properties": {
+          "name": {"type": "string"},
+          "source_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {"type": "integer"},
+                "end_unit_id": {"type": "integer"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,8 @@
+Determine whether candidates identify the same physical place using their
+contextual labels and cited transcript windows. Do not treat matching names,
+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.

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

@@ -0,0 +1,30 @@
+id: dnd.location_registry.normalize
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: candidates
+    required: true
+    content_type: application/json
+  - name: transcript
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/protocol.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/candidates.md
+  - role: user
+    content_file: ./sharedassets/transcript-windows.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: semantic_reconciliation_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,26 @@
+Extract Dungeons & Dragons NPC occurrences from the supplied
+transcript. Include an occurrence only when the transcript establishes one
+supplied NPC, one occurrence kind, and a coherent passage supporting both.
+Use the supplied canonical NPC `name`; never invent or substitute a similar
+name. Cite current-transcript evidence for every occurrence.
+
+Do not summarize, infer relationships, sentiment, factions, motives, aliases,
+or persistent state. Do not identify player characters, anonymous groups, or
+invented NPCs. Split records when an NPC's occurrence kind changes, when
+combat alignment changes, or when an NPC is first mentioned and later becomes
+present.
+
+Use exactly one kind per occurrence:
+
+- mentioned: the NPC is referred to but is not established as present or communicating;
+- noncombat_presence: the NPC is present and relevant but does not meaningfully participate in dialogue or combat;
+- dialogue: the NPC speaks, responds, or is directly engaged in a meaningful non-combat exchange;
+- combat_ally: the NPC actively participates in combat on the party's side;
+- combat_opponent: the NPC actively participates in combat against the party; or
+- other: the transcript clearly establishes a direct NPC occurrence that fits none of the preceding kinds.
+
+When activities overlap, active combat participation outranks dialogue,
+presence, and mention; dialogue outranks noncombat presence and mention; and
+noncombat presence outranks mention. Other is only for directly evidenced
+activity outside those categories. Split an occurrence rather than assigning
+both combat alignments.

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

@@ -0,0 +1,45 @@
+id: dnd.npc_occurrences
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+  - name: npc_registry
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-npc-registry.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_npc_occurrences_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,41 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.npc_occurrences.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["occurrences"],
+  "properties": {
+    "occurrences": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "kind", "source_refs"],
+        "properties": {
+          "name": {
+            "type": "string"
+          },
+          "kind": {
+            "type": "string"
+          },
+          "source_refs": {
+            "type": "array",
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["start_unit_id", "end_unit_id"],
+              "properties": {
+                "start_unit_id": {
+                  "type": "integer"
+                },
+                "end_unit_id": {
+                  "type": "integer"
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,19 @@
+Extract the individually identifiable Dungeons & Dragons non-player characters
+established by the provided transcript.
+
+Include an in-world non-PC only when the transcript factually establishes a
+proper name or a stable, individually distinguishing title or alias. A factual
+third-party mention establishes that identity even when the NPC is not
+physically present, does not speak, and takes no direct action in this chunk.
+Record only the NPC identity and the transcript evidence that establishes it;
+do not infer or classify a separate occurrence.
+
+Exclude human players, transcript speakers, and the GM as out-of-world people;
+player characters identified by the player or party references; names used only
+in hypothetical, speculative, or imagined examples; corrected transcription
+mistakes; anonymous or generic roles; indistinguishable crowds or groups;
+invented descriptive labels; and temporary summoned creatures or spell effects
+without a persistent individual identity.
+
+Preserve observed display spelling. Do not invent a label for an anonymous
+creature, crowd, or generic role.

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

@@ -0,0 +1,40 @@
+id: dnd.npc_registry
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_npc_registry_llm.v1.json
+  repair_attempts: 0

assets/dnd/npc-registry/extract/schemas/dnd_npc_registry_llm.v1.json

@@ -1,6 +1,6 @@
 {
   "$schema": "https://json-schema.org/draft/2020-12/schema",
-  "$id": "notarius.dnd.npcs.llm",
+  "$id": "notarius.dnd.npc_registry.llm",
   "type": "object",
   "additionalProperties": false,
   "required": ["npcs"],

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

@@ -0,0 +1,11 @@
+Determine whether candidates refer to the same individual using their
+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
+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
+is not inherently more canonical; for example, do not prefer `Captain Aria`
+over `Aria` solely because it includes the contextual title `Captain`.

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

@@ -0,0 +1,30 @@
+id: dnd.npc_registry.normalize
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: candidates
+    required: true
+    content_type: application/json
+  - name: transcript
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/protocol.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/candidates.md
+  - role: user
+    content_file: ./sharedassets/transcript-windows.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: semantic_reconciliation_llm.v1.json
+  repair_attempts: 0

assets/dnd/profiles/dnd-extraction.yaml

@@ -0,0 +1,5 @@
+id: dnd-extraction
+backend: openrouter
+model: openai/gpt-5.6-luna
+timeout_seconds: 240
+service_tier: flex

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

@@ -0,0 +1,45 @@
+Describe exactly one accepted Dungeons & Dragons scene from the supplied
+transcript chunk. The complete chunk is the evidence boundary: do not split it
+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.
+- 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,
+  in-character planning, and aftermath.
+- recap: the scene's organizing purpose is to recount events from a previous
+  session for the table. An in-world character recounting history during
+  current play remains narrative.
+- meta: the scene's organizing purpose is sustained out-of-character
+  discussion about the game or session rather than advancing current in-world
+  play.
+
+Narrative is the default for actual current-session gameplay that does not meet
+another definition. When the accepted chunk is mixed:
+
+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
+   primary table purpose;
+3. 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.
+
+Brief table talk, dice resolution, rules clarification, jokes, or
+administrative comments do not make a gameplay scene meta. A short recollection
+used to orient current action does not make a scene recap.
+
+The title must be a short, distinguishing phrase rather than a sentence,
+chapter number, or generic label such as "Scene." It may use names and places
+established by the transcript or disambiguated by campaign references, but it
+must not invent a proper noun.
+
+The summary must briefly state the main activity and material transition or
+outcome established within the accepted chunk. Do not add analysis, inferred
+motives, hidden state, future consequences, relationship claims, or facts from
+outside the chunk.

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

@@ -0,0 +1,38 @@
+id: dnd.scene_descriptions
+version: "v1"
+default_profile: dnd-extraction
+inputs:
+  - name: transcript
+    required: true
+    content_type: application/json
+  - name: players
+    required: false
+    content_type: text/plain
+  - name: party
+    required: false
+    content_type: text/plain
+  - name: glossary
+    required: false
+    content_type: text/plain
+messages:
+  - role: system
+    content_file: ./sharedassets/common-dnd-system.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-identity.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-references.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_scene_descriptions_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,18 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.scene_descriptions.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["kind", "title", "summary"],
+  "properties": {
+    "kind": {
+      "type": "string"
+    },
+    "title": {
+      "type": "string"
+    },
+    "summary": {
+      "type": "string"
+    }
+  }
+}

assets/dnd/scenes/prompts/instructions.md

@@ -0,0 +1,19 @@
+Divide the complete provided transcript into coherent Dungeons & Dragons scenes
+for the `dnd/scenes` chunk module.
+
+A scene is a coherent unit of play. Start a new scene when the transcript
+establishes a meaningful change in location, objective, threat, activity,
+encounter, or mode of play. Good reasons include a material move, beginning or
+ending combat, a substantially different encounter phase, a shift between
+combat, exploration, social interaction, planning, travel, rest, or downtime,
+a change in the central NPC, faction, threat, or objective, or a sustained
+table-level interruption that materially changes the activity.
+
+Do not split a scene merely because a speaker or combat round changes, a
+routine turn occurs, or the table briefly digresses. Prefer fewer coherent
+scenes over speculative or fine-grained boundaries.
+
+Cover the complete transcript from its first source unit to its last. Return
+scenes in source-unit order with no gaps or overlaps. Use only positive integer
+source-unit IDs from the transcript, and give every scene one inclusive
+`start_unit_id` and one inclusive `end_unit_id`.

assets/dnd/scenes/prompts/prompt.yaml

@@ -1,6 +1,6 @@
 id: dnd.scenes
 version: "v1"
-default_profile: gemini-2-flash
+default_profile: dnd-extraction
 inputs:
   - name: transcript
     required: true
@@ -17,20 +17,18 @@ inputs:
 messages:
   - role: system
     content_file: ./sharedassets/common-dnd-system.md
-  - role: user
-    content_file: ./sharedassets/common-dnd-transcript.md
-    cache_control:
-      type: ephemeral
   - role: user
     content_file: ./sharedassets/common-dnd-references.md
     cache_control:
       type: ephemeral
-  - role: user
-    content_file: ./task.md
   - role: user
     content_file: ./instructions.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-transcript-full.md
+    cache_control:
+      type: ephemeral
 output:
   format: json
   validation_mode: json_schema
-  schema_path: dnd_scenes.v1.json
+  schema_path: dnd_scenes_llm.v1.json
   repair_attempts: 0

assets/dnd/scenes/schemas/dnd_scenes_llm.v1.json

@@ -0,0 +1,28 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.scenes.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["scenes"],
+  "properties": {
+    "scenes": {
+      "type": "array",
+      "minItems": 1,
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["start_unit_id", "end_unit_id"],
+        "properties": {
+          "start_unit_id": {
+            "type": "integer",
+            "minimum": 1
+          },
+          "end_unit_id": {
+            "type": "integer",
+            "minimum": 1
+          }
+        }
+      }
+    }
+  }
+}

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

@@ -0,0 +1,6 @@
+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.
+
+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-npc-registry.md

@@ -1,10 +1,10 @@
-An optional normalized Dungeons & Dragons NPC registry is provided below as
+A normalized Dungeons & Dragons NPC registry is provided below as grounding
-grounding material. Use it only to prefer exact canonical participant names
+material. It may be empty. Use it only to prefer exact canonical participant
-and recognize their aliases when the transcript identifies a participant.
+names when the transcript identifies a participant.
 
 Registry content is context, not event evidence. Do not extract events,
 participants, effects, or source references from the registry. Registry source
 references describe registry provenance and may belong to another session; they
 are never evidence for the current transcript.
 
-{{ input "npcs" }}
+{{ input "npc_registry" }}

assets/dnd/shared/prompts/common-dnd-system.md

@@ -0,0 +1,5 @@
+You process Dungeons & Dragons gameplay transcripts.
+
+As input, you will receive one or more portions of a transcript. The transcript may contain transcription errors, repeated lines, incomplete sentences, and misheard proper nouns.
+
+Return exactly one JSON object that conforms to the configured response schema, with no explanatory prose.

assets/dnd/shared/prompts/common-dnd-transcript-chunk.md

@@ -0,0 +1,3 @@
+One extraction chunk from a Dungeons & Dragons gameplay transcript is provided below. Report and infer only what is within this chunk. Its unit IDs retain their source-wide meaning.
+
+{{ input "transcript" }}

assets/dnd/shared/prompts/common-dnd-transcript-full.md

@@ -0,0 +1,3 @@
+The complete ordered transcript of this Dungeons & Dragons gameplay session is provided below.
+
+{{ input "transcript" }}

assets/dnd/spells/prompts/instructions.md

@@ -0,0 +1,14 @@
+Extract Dungeons & Dragons spell-cast artifacts from the provided transcript.
+Include an actual casting event or an unambiguous declared casting attempt.
+Exclude spell mentions, hypothetical plans, rules discussion, and catalog
+matches that do not establish a casting event in the transcript.
+
+For every extracted cast, the transcript evidence must collectively support the
+in-world caster, the spell, and the fact that the cast or declared attempt
+occurred.
+
+Attribute every cast to its in-world caster. Map first-person player speech to
+the associated player character, and attribute a spell narrated by the GM to
+the in-world creature that casts it. If the caster cannot be resolved, use only
+the most specific in-world identity supported by the transcript; do not invent
+a name.

assets/dnd/spells/prompts/prompt.yaml

@@ -1,6 +1,6 @@
 id: dnd.spells
 version: "v1"
-default_profile: gemini-2-flash
+default_profile: dnd-extraction
 inputs:
   - name: transcript
     required: true
@@ -17,36 +17,32 @@ inputs:
   - name: glossary
     required: false
     content_type: text/plain
-  - name: npcs
+  - name: npc_registry
     required: false
     content_type: application/json
 messages:
   - role: system
     content_file: ./sharedassets/common-dnd-system.md
-  - role: user
-    content_file: ./sharedassets/common-dnd-extraction-evidence.md
   - role: user
     content_file: ./sharedassets/common-dnd-identity.md
-    cache_control:
-      type: ephemeral
   - role: user
     content_file: ./sharedassets/common-dnd-references.md
     cache_control:
       type: ephemeral
   - role: user
-    content_file: ./sharedassets/common-dnd-npcs.md
+    content_file: ./sharedassets/common-dnd-transcript-chunk.md
     cache_control:
       type: ephemeral
   - role: user
-    content_file: ./catalog.md
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
   - role: user
-    content_file: ./task.md
+    content_file: ./sharedassets/common-dnd-npc-registry.md
+  - role: user
+    content_file: ./spell-catalog.md
   - role: user
     content_file: ./instructions.md
     cache_control:
       type: ephemeral
-  - role: user
-    content_file: ./sharedassets/common-dnd-transcript.md
 output:
   format: json
   validation_mode: json_schema

assets/dnd/spells/prompts/spell-catalog.md

@@ -0,0 +1,6 @@
+The spell catalog for this extraction is provided below as JSON. Each entry
+lists a `canonical_name` and its recognized `aliases`. If the transcript uses
+an alias, select that entry's `canonical_name`. Return spell names using the
+canonical spelling exactly; never return an alias as a spell name.
+
+{{ input "spell_catalog" }}

assets/dnd/spells/schemas/dnd_spells_llm.v1.json

@@ -1,6 +1,6 @@
 {
   "$schema": "https://json-schema.org/draft/2020-12/schema",
-  "$id": "notarius.dnd.spells",
+  "$id": "notarius.dnd.spells.llm",
   "type": "object",
   "additionalProperties": false,
   "required": ["spell_casts"],

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

@@ -0,0 +1,3 @@
+Candidate material:
+
+{{ input "candidates" }}

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

@@ -0,0 +1,5 @@
+Identify only high-confidence duplicate entities among the supplied candidates.
+
+Preserve distinct entities even when their names are similar. Treat contextual descriptions and transcript evidence as supporting material, not as permission to merge ambiguous records.
+
+When several records are duplicates, choose as canonical the candidate with the clearest stable identity. Prefer a complete proper name over an abbreviation, and prefer an unadorned proper name over one with incidental descriptors unless the evidence establishes those descriptors as part of the name. A longer name is not inherently more canonical.

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

@@ -0,0 +1,27 @@
+id: generic.semantic_reconciliation
+version: "v1"
+inputs:
+  - name: candidates
+    required: true
+    content_type: application/json
+  - name: transcript
+    required: true
+    content_type: application/json
+messages:
+  - role: system
+    content_file: ./system.md
+  - role: user
+    content_file: ./protocol.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./candidates.md
+  - role: user
+    content_file: ./transcript-windows.md
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: semantic_reconciliation_llm.v1.json
+  repair_attempts: 0

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

@@ -0,0 +1,7 @@
+Use only the positive integer `candidate_id` values supplied in the candidate material.
+
+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.
+
+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.
+
+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.

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

@@ -0,0 +1,2 @@
+You reconcile structured records that may describe the same underlying entity.
+Follow the supplied protocol and return only the requested structured result.

assets/generic/normalize/deduplication/prompts/transcript-windows.md

@@ -0,0 +1,3 @@
+Transcript evidence windows:
+
+{{ input "transcript" }}

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

@@ -0,0 +1,32 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.generic.semantic_reconciliation.llm",
+  "title": "notarius_semantic_reconciliation_llm_v1",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["duplicate_groups"],
+  "properties": {
+    "duplicate_groups": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["candidate_ids", "canonical_candidate_id"],
+        "properties": {
+          "candidate_ids": {
+            "type": "array",
+            "minItems": 2,
+            "items": {
+              "type": "integer",
+              "minimum": 1
+            }
+          },
+          "canonical_candidate_id": {
+            "type": "integer",
+            "minimum": 1
+          }
+        }
+      }
+    }
+  }
+}

assets/package.go

@@ -0,0 +1,15 @@
+// Package assets exposes embedded LLM-facing content.
+package assets
+
+import (
+	"embed"
+	"io/fs"
+)
+
+//go:embed dnd generic
+var embedded embed.FS
+
+// FS returns the embedded read-only asset filesystem.
+func FS() fs.FS {
+	return embedded
+}

docs/adr/0004-package-modules-by-domain.md

@@ -1,6 +1,6 @@
 # ADR-0004: Package modules by domain, not by stage
 
-**Status:** Accepted
+**Status:** Accepted — its asset-co-location rule is superseded by [ADR-0011](0011-centralize-llm-assets.md); its domain-first module packaging decision remains accepted.
 **Date:** 2026-07-13
 
 ## Context

docs/adr/0010-workload-oriented-llm-profile-defaults.md

@@ -0,0 +1,49 @@
+# ADR-0010: Use workload-oriented LLM profile defaults
+
+**Status:** Accepted
+**Date:** 2026-08-03
+
+## Context
+
+LLM-backed D&D operations share an execution-policy choice, but repeating a
+provider or model-named profile on every module binding ties pipeline structure
+to a deployment decision. Different environments may require different model,
+backend, timeout, or reasoning settings while retaining the same workload.
+
+Notarius also needs a usable default for maintained D&D prompts without making
+an operator profile mandatory. That default must remain owned by the D&D
+family, while generic LLM infrastructure stays unaware of domain-specific
+policy.
+
+## Decision
+
+Pipelines may name one workload-oriented default profile, inherited only by
+selected LLM-backed bindings and validators. Binding-level profile IDs remain
+intentional exceptions, and the run-wide CLI profile override has highest
+precedence.
+
+The D&D family owns an embedded fallback profile named `dnd-extraction`.
+Operators may provide a complete profile with the same ID through a PromptKit
+filesystem source. PromptKit selects the higher-precedence matching definition;
+Notarius does not merge profile documents. Production, development, and local
+deployments can therefore use different execution policy behind one unchanged
+pipeline ID.
+
+## Alternatives considered
+
+- Repeat a model-named profile on every binding. This makes routine deployment
+  policy changes noisy and obscures the shared workload intent.
+- Require every deployment to install a profile file. This adds configuration
+  friction and leaves maintained D&D prompts without an application-owned
+  fallback.
+- Put D&D profile policy in generic LLM infrastructure. This breaks domain
+  ownership and makes generic code depend on one workload.
+
+## Consequences
+
+Pipeline configuration expresses workload intent rather than a specific
+provider or model. Operators can replace the complete execution policy without
+editing bindings, while binding-level and run-wide exceptions remain available.
+Profile changes affect resolved pipeline and checkpoint identity, so they may
+intentionally cause work to be recomputed. The D&D fallback becomes a
+maintained application execution-policy asset.

docs/adr/0011-centralize-llm-assets.md

@@ -0,0 +1,69 @@
+# ADR-0011: Centralize LLM-facing assets in a content-only package
+
+**Status:** Accepted
+**Date:** 2026-08-05
+
+## Context
+
+LLM prompts, private response schemas, generic schemas, and fallback profiles
+are authored and reviewed as content, but package-local embedding scattered that
+content across implementation trees. Finding all of the assets that contribute
+to a prompt family required navigating code ownership boundaries rather than a
+single discoverable content boundary.
+
+The repository must retain module ownership of prompt semantics, schema
+identities, registration, and prompt-cache behavior. Durable artifact schemas
+and non-LLM domain data have different compatibility and ownership rules, so
+they must not move merely because they are embedded files.
+
+## Decision
+
+LLM-facing content is embedded by the root `assets` package. It is a data-only
+dependency leaf: its single `FS() fs.FS` API returns the read-only embedded
+filesystem, and the package contains no business logic or internal or PromptKit
+dependencies. The accepted import path is
+`gitea.maximumdirect.net/eric/notarius/assets`; it makes repository-owned
+content available to its consumers, not a public extension contract.
+
+Consumers scope that filesystem to the subtree they own before reading or
+registering content. Modules continue to own their manifests, prompt ordering,
+private response-schema identity, and registration. Centralizing physical files
+does not centralize domain semantics or transfer those responsibilities to the
+root package.
+
+The root package contains prompt content, private LLM response schemas, generic
+LLM schemas, shared fragments, and fallback profiles. Durable artifact schemas
+and non-LLM domain data remain with their current owners. A module fingerprint
+is derived from its manifest-selected module and shared files, rather than from
+an entire asset tree. The relocation is accepted to cause a one-time checkpoint
+invalidation.
+
+This decision supersedes only the physical asset-co-location portion of
+ADR-0004's decision that places domain-specific prompt fragments and schemas
+within the domain tree. ADR-0004's domain-first packaging and registrar
+ownership decisions remain accepted.
+
+## Alternatives Considered
+
+- Keep package-local assets. This preserves physical co-location with code but
+  makes prompt-author discovery and cross-family review unnecessarily costly.
+- Use `internal/llmassets`. This would hide content from legitimate owners
+  outside the `internal` subtree and would make the root asset boundary depend
+  on implementation-layer placement.
+- Build a behavioral central registry. This would mix content discovery with
+  prompt selection and registration behavior, moving module semantics into a
+  shared registry.
+- Use runtime filesystem overlays. This would add runtime configuration and
+  failure modes where compile-time embedded content is sufficient.
+
+## Consequences
+
+Prompt authors can find in-scope LLM content in one top-level tree while module
+packages continue to define its meaning and registration. Consumers have an
+explicit, narrow dependency on only the content they need. The root package is
+intentionally importable but must remain a stable, content-only leaf rather
+than becoming a general extension API.
+
+The initial relocation invalidates existing checkpoints once. Later checkpoint
+identity changes remain limited to the manifest-selected prompt and shared
+content, so unrelated files do not trigger recomputation.

docs/adr/0012-resolve-opaque-entity-identifiers-deterministically.md

@@ -0,0 +1,64 @@
+# ADR-0012: Resolve opaque entity identifiers deterministically
+
+**Status:** Accepted
+**Date:** 2026-08-08
+
+## Context
+
+Entity IDs in durable Notarius artifacts are application-owned, deterministic
+identifiers. They are useful to artifact consumers, but their hash-based form
+does not help a model distinguish entities and would make the model reproduce
+an opaque implementation detail. A plain name is likewise insufficient where
+multiple supplied records share that name.
+
+The LLM boundary must preserve the typed artifact and durable-schema ownership
+of [ADR-0003](0003-typed-interfaces-with-two-zone-data-model.md) and the distinction
+between disambiguating references and source evidence in
+[ADR-0009](0009-minimal-evidence-grounded-extraction-artifacts.md).
+
+## Decision
+
+Callers present a model with semantic selections: a canonical name when it is
+unique in the request, or a contextual descriptor containing the name and
+source coordinates when that context is needed to distinguish supplied
+records. The model returns only those supplied selections. The caller resolves
+each accepted selection against the request-local supplied records and attaches
+the opaque application ID deterministically.
+
+Source coordinates are permitted in a selection solely as identity context.
+They neither establish an occurrence fact nor replace that occurrence's
+current-transcript evidence. A selector must resolve exactly; unknown,
+ambiguous, partial, reordered, or otherwise unsafe selections are not mapped.
+Where an operation requires a complete grounded artifact, that failure rejects
+the complete artifact rather than accepting a partially mapped result.
+
+An explicitly scoped request-local short label is permitted only when a
+contextual descriptor would be impractical and the caller can deterministically
+map the label within that one request. Such a label is not a durable ID, must
+not escape the request boundary, and requires a concrete justification in its
+own module contract.
+
+## Alternatives considered
+
+- Ask the model to return durable IDs. This exposes opaque implementation
+  state, does not improve semantic disambiguation, and makes model output
+  depend on hash formatting.
+- Select by name alone. This cannot safely distinguish same-name records.
+- Make request-local labels durable identifiers. This would turn prompt
+  presentation into a public identity contract and create avoidable migration
+  pressure.
+- Let the model invent identifiers or resolve ambiguity. This makes identity
+  assignment non-deterministic and weakens validation.
+
+## Consequences
+
+Durable integration contracts retain their exact ID/name pairs while models
+operate on readable contextual selections. Calling modules must own selector
+construction, exact resolution, ambiguity handling, and conversion into their
+durable artifact type; PromptKit and its adapter remain transport-only.
+
+Some ambiguous or invalid proposals are deliberately omitted, retried, or
+rejected according to the caller's existing failure policy. Internal candidate
+keys may support deterministic request-local mapping, but they are not
+model-visible selectors or durable data. This adds local validation work while
+keeping identity assignment auditable and stable.

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

@@ -0,0 +1,91 @@
+# ADR-0013: Use request-local candidate handles for semantic reconciliation
+
+**Status:** Accepted
+**Date:** 2026-08-09
+
+## Context
+
+Several typed normalize stage modules need semantic reconciliation after
+deterministic preprocessing: a model can judge whether source-backed candidates
+refer to the same underlying entity, while application code remains responsible
+for constructing the normalized artifact. Requiring the model to reproduce a
+candidate's full contextual selector makes the response larger and introduces
+avoidable formatting, ordering, and transcription failure modes.
+
+Reconciliation must preserve the exact typed artifact boundary established by
+[ADR-0003](0003-typed-interfaces-with-two-zone-data-model.md), the domain-neutral
+framework and concrete-domain dependency direction established by
+[ADR-0004](0004-package-modules-by-domain.md), and the distinction in
+[ADR-0009](0009-minimal-evidence-grounded-extraction-artifacts.md) between source
+evidence and auxiliary identity context. It also needs a concrete, narrowly
+scoped application of the request-local-label exception allowed by
+[ADR-0012](0012-resolve-opaque-entity-identifiers-deterministically.md).
+
+## Decision
+
+Semantic reconciliation will be a domain-neutral framework mechanism used by
+typed normalize stage modules. A consuming artifact family will retain
+ownership of its typed records, identity rules, consolidation policy, durable
+IDs, and domain warnings; the framework mechanism will not infer those rules
+from arbitrary data.
+
+For each reconciliation request, deterministic code will assign every eligible
+model-visible candidate a contiguous, one-based integer handle. The model may
+receive the candidate's contextual label, source references, and bounded source
+context needed to judge identity, but its structured response will identify
+candidates only by those supplied handles. A handle is local to one request,
+does not represent entity identity, and must never enter a durable artifact or
+be used to derive a durable ID.
+
+The model will propose duplicate groups and select one supplied member of each
+group as canonical. Deterministic code will resolve the handles through the
+retained request mapping, validate the complete proposal, discard unsafe
+groups, and apply only validated groups through typed domain-owned policy. The
+model will not synthesize replacement records or directly mutate an artifact.
+
+Every reconciliation prompt will combine a mandatory framework-owned protocol
+and safety policy with an explicitly selected semantic policy. The semantic
+policy may be the conservative generic policy or a domain-owned policy, but it
+cannot replace the shared response protocol or deterministic safety boundary.
+
+## Alternatives considered
+
+- Return durable application IDs. Opaque IDs do not help semantic judgment,
+  expose application identity mechanics, and make model output reproduce data
+  that deterministic code already owns.
+- Return names alone or copied contextual selectors. Names can be ambiguous,
+  while reproducing labels and source ranges adds response complexity and
+  creates mismatches without adding semantic information. Request-local
+  handles preserve exact selection without either failure mode.
+- Ask the model to return synthesized canonical replacement records. This
+  would transfer typed artifact construction, provenance consolidation, and
+  durable identity policy to a probabilistic boundary.
+- Reconcile reflection-discovered fields or arbitrary JSON. This would weaken
+  the typed artifact contract and move domain semantics into generic code.
+- Hide reconciliation inside extraction or another stage. This would obscure
+  stage ownership and create cross-stage behavior outside the fixed pipeline;
+  reconciliation remains explicit normalize-stage behavior.
+- Let each domain replace the complete prompt protocol. This would duplicate
+  safety mechanics and allow domain policy to bypass the common response and
+  validation contract.
+
+## Consequences
+
+Model responses become smaller and easier to validate, while deterministic
+application code retains authority over identity, provenance, ordering, and
+typed artifact construction. The framework requires a request-local mapping,
+bounded context preparation, a private integer response contract, proposal
+assessment, and shared prompt assets. Each consuming artifact family still
+requires a typed adapter for its irreducibly domain-specific rules.
+
+Request-local handles are deliberately unsuitable for persistence, logging as
+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.

docs/cli.md

@@ -1,273 +1,175 @@
 # CLI Reference
 
 This is the canonical reference for the implemented Notarius command-line
-interface.
+interface. For the shortest successful run, see the [README](../README.md).
+Configuration fields, discovery rules, and selectable module keys are defined
+in [Configuration](config.md); runtime state and recovery procedures are
+defined in [Operations](operations.md).
 
-For the minimal end-to-end invocation, see the [README](../README.md).
+## Command Summary
 
-## Commands
+~~~
-
-```text
 notarius help
-notarius run <pipeline-id> --input path/to/source.json [--config path/to/config.yml] [--only lane-a,lane-b] [--chunk_cache auto|bypass|refresh] [--output-dir path] [--resume] [--recompute-step step-id] [--debug [--debug-dir path]] [--llm-profile id] [--session-id id] [--reference selector=path] [--without-reference selector]
+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]
-```
+~~~
 
-Running `notarius` with no arguments, `notarius help`, `notarius --help`, or
+Running Notarius without arguments, or with **help**, **--help**, or **-h**,
-`notarius -h` prints usage and exits successfully.
+writes the command summary to standard output and exits with status 0.
 
-## `run`
+## run
 
-`notarius run <pipeline-id>` executes a configured pipeline against one input
+~~~
-file.
+notarius run <pipeline-id> --input path/to/source.json [--json] [flags]
+~~~
 
-Flags:
+The **run** command executes the named pipeline for one input file. The
+pipeline ID and **--input** are required.
 
-- `--input path`: required source input file.
+| Flag | Meaning |
-- `--config path`: config file path. If omitted, Notarius uses the discovery
+| --- | --- |
-  rules in [Configuration](config.md#discovery).
+| **--config path** | Use this configuration file. When omitted, configuration discovery applies; see [Configuration](config.md). |
-- `--only lane-a,lane-b`: run only the named artifact lanes. Values are
+| **--input path** | Source input file to process. Required. |
-  comma-separated and must be non-empty. This retains its existing behavior for
+| **--output-dir path** | Override the configured output root for this run. |
-  implicit single-step pipelines; explicit multi-step pipelines reject it
+| **--json** | Write the successful run-result receipt as JSON to standard output. |
-  rather than inferring dependency closure.
+| **--chunk_cache auto\|bypass\|refresh** | Override chunk-plan cache handling for this run. |
-- `--resume`: request checkpoint reuse for this invocation. Checkpoint recording
+| **--resume** | Reuse compatible recorded checkpoints when checkpoint recording is enabled. |
-  must be enabled in configuration. See
+| **--recompute-step step-id** | With **--resume**, recompute the selected ordered step and its dependent lanes. It cannot be combined with **--only**. |
-  [Operations](operations.md#checkpoint-cache) for prerequisites and reuse
+| **--debug** | Retain a debug bundle for this run. |
-  behavior.
+| **--debug-dir path** | Override the debug-bundle root. Requires **--debug**. |
-- `--recompute-step step-id`: with `--resume` and checkpoint recording enabled,
+| **--only lane-a,lane-b** | Run only the selected comma-separated artifact lanes when that selection is valid for the configured pipeline. |
-  force the named ordered step and every transitive dependent lane to execute.
+| **--llm-profile id** | Highest-precedence configured profile for selected LLM-backed bindings and validators; it replaces binding and [pipeline](config.md#pipelines) defaults. |
-  Compatible required predecessors and unrelated lanes remain reusable. The
+| **--session-id id** | Override the generated prompt session identifier with a non-empty value for LLM-backed module calls. |
-  value may identify an explicit step or the implicit single-step ID `default`;
+| **--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. |
-  it cannot be combined with `--only`.
+| **--clear-reasoning-effort** | Clear reasoning effort inherited from the selected PromptKit profile for every LLM-backed call in this run. |
-- `--chunk_cache auto|bypass|refresh`: select chunk-plan reuse for this
+| **--reference selector=path** | Add or replace a file reference binding. Repeatable. |
-  invocation. `auto` reuses a valid plan by canonical source digest, `bypass`
+| **--without-reference selector** | Remove a configured optional reference binding. Repeatable. |
-  performs no plan-cache I/O, and `refresh` regenerates and replaces a valid
-  plan only after chunk validation succeeds. See
-  [Configuration](config.md#state-surfaces) for the persistent setting, precedence,
-  and cache-root selection.
-- `--output-dir path`: output root. Defaults to `./notarius-output`.
-- `--debug`: allocate and retain one debug bundle for this invocation.
-- `--debug-dir path`: debug-bundle root override. This flag requires `--debug`.
-- `--llm-profile id`: override every effective LLM-capable pipeline module
-  binding with one Scriptorium profile ID. Validator-specific profiles are not
-  overridden.
-- `--session-id id`: pass a stable prompt session identifier through LLM-backed
-  module calls.
-- `--reference selector=path`: bind a reference path to a chunk, extractor,
-  merger, or normalizer reference slot. Repeatable.
-- `--without-reference selector`: remove a configured optional reference binding.
-  Repeatable. It accepts the same selector forms as `--reference`, without
-  `=path`.
 
-On success, the command prints the completed pipeline ID, normalized output and
+**--chunk_cache** accepts only **auto**, **bypass**, or **refresh**.
-rejected output counts, and the output directory. A debug-enabled run also
+**--debug-dir**, **--output-dir**, **--session-id**, and
-prints `debug=<bundle-path>`. If the run completes with warnings, the warning
+**--reasoning-effort**, and **--recompute-step** reject explicit empty values.
-count is printed to stderr.
+**--reasoning-effort** and **--clear-reasoning-effort** are mutually exclusive.
+When neither is present, reasoning effort comes from the selected PromptKit
+profile. These controls apply to the shared run client, including retries and
+LLM-backed validators, and do not modify configuration or profile files.
+Persistent reasoning settings remain a PromptKit profile concern.
+**--recompute-step** requires **--resume**; checkpoint requirements and reuse
+behavior are documented in [Operations](operations.md).
 
-Reference flags are external file bindings resolved against selected chunk,
+Every run uses one effective prompt session. Without **--session-id**, Notarius
-extractor, merger, and normalizer targets before the run starts. Generated
+generates a stable `notarius:v1:` identifier from the trimmed resolved input
-artifact bindings are configured in ordered steps and cannot be introduced by a
+module key and the input file's exact raw bytes. The same module and bytes
-CLI path flag. Flat slot names are accepted only
+therefore produce the same identifier, regardless of pipeline, references,
-when exactly one selected target declares that slot. For configured reference
+profile, retries, or run settings. An explicit non-empty value replaces that
-bindings, precedence, path resolution, and validation, see
+default. Session identifiers are visible to providers; they are non-secret
-[Configuration](config.md#pipelines).
+correlation identifiers, not credential storage. See
+[Operations](operations.md#operational-limits) for privacy and workflow
+guidance.
 
-`--reference` binds or replaces one slot for one selected target. Selectors are:
+### Reference selectors
 
-- `slot=path`: valid when exactly one selected target declares `slot`;
+Use **--reference** only for a reference slot declared by the selected
-- `chunk.slot=path`: target the chunker;
+configured target. The accepted selector forms are:
-- `merge.slot=path`: valid when exactly one selected merger declares `slot`;
-- `lane.slot=path`: valid when exactly one selected extractor, merger, or
-  normalizer in that lane declares `slot`;
-- `lane.extract.slot=path`: target a lane extractor;
-- `lane.merge.slot=path`: target a lane merger;
-- `lane.normalize.slot=path`: target a lane normalizer.
 
-Use `slot=path` when the selected targets declare the slot unambiguously:
+| Form | Target |
+| --- | --- |
+| slot=path | The unique 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.extract.slot=path | The extractor in lane. |
+| lane.merge.slot=path | The merger in lane. |
+| lane.normalize.slot=path | The normalizer in lane. |
 
-```sh
+**--without-reference** uses the same selector forms without =path. Slot
-go run ./cmd/notarius run dnd-session \
+names, requiredness, and configured bindings are part of the
-  --config examples/dnd-spells.config.yml \
+[configuration contract](config.md).
-  --input examples/seriatim-minimal-transcript.json \
-  --reference roster=./campaign-roster.txt
-```
 
-Use an explicit selector when multiple selected targets declare the same slot or
+### Run output
-when you want to target a specific target:
 
-```sh
+Without **--json**, standard output contains the completed pipeline ID, counts
-go run ./cmd/notarius run dnd-session \
+of normalized and rejected outputs, and the output directory. A debug-enabled
-  --config examples/dnd-spells.config.yml \
+run also prints its debug-bundle path to standard output. A successful run with
-  --input examples/seriatim-minimal-transcript.json \
+warnings reports the warning count to standard error. The published JSON bundle
-  --reference spells.extract.glossary=./campaign-glossary.txt
+is defined by the [JSON output contract](integrations/json-output.md).
-```
 
-For the maintained NPC-grounded workflow, use the explicit ordered pipeline.
+With **--json**, successful standard output is exactly one
-The first step produces the normalized NPC artifact; the second step receives
+`notarius.run-result.v1` JSON document followed by a newline, with no
-it in memory and fans it out to spell extraction, combat extraction, and combat
+human-oriented status or debug-path line. Its fields and compatibility policy
-normalization:
+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
+leave incomplete standard-output bytes that are not a result document.
 
-```sh
+Example:
-go run ./cmd/notarius run dnd-npc-grounded \
-  --config examples/dnd-npc-grounded.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --output-dir ./npc-grounded-output
-```
 
-The generated NPC content remains contextual grounding, not spell or combat
+~~~
-evidence. It is represented in manifests and debug summaries by bounded
+OPENROUTER_API_KEY=your-api-key \
-identity and producer provenance, not by payload content or a filesystem path.
+  go run ./cmd/notarius run dnd-session \
+    --config examples/dnd-minimal.config.yml \
+    --input examples/seriatim-minimal-transcript.json
+~~~
 
-The same grammar can target chunk, merge, and normalize slots when the configured
+## config validate
-modules declare them:
 
-```sh
+~~~
-go run ./cmd/notarius run dnd-session \
+notarius config validate [--config path/to/config.yml] [--pipeline pipeline-id] [--only lane-a,lane-b]
-  --config path/to/config.yml \
+~~~
-  --input examples/seriatim-minimal-transcript.json \
-  --reference chunk.scene_guide=./campaign-scenes.txt \
-  --reference spells.merge.merge_notes=./merge-notes.txt \
-  --reference spells.normalize.normalization_notes=./normalization-notes.txt
-```
 
-Use `--without-reference` to remove a configured optional binding for a run:
+This command loads and validates a configuration. With **--pipeline**, it also
+resolves that pipeline against the production module catalog. **--only** selects
+lanes during that resolution and requires **--pipeline**.
 
-```sh
+Success is written to standard output as either config "<path>" is valid or
-go run ./cmd/notarius run dnd-session \
+config "<path>" is valid for pipeline "<pipeline-id>".
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --without-reference glossary
-```
-
-Use `--session-id` when an external orchestrator needs all prompt calls from one
-run to share an identifier:
-
-```sh
-go run ./cmd/notarius run dnd-session \
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --session-id campaign-17-session-04
-```
-
-When `cache.checkpoints.enabled` is `true`, runs record checkpoints whether or
-not `--resume` is present. Add the resume flag to load and reuse compatible
-recorded work; using it while checkpoint recording is disabled is an error:
-
-```sh
-go run ./cmd/notarius run dnd-session \
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --resume
-```
-
-To selectively rerun one ordered step and its dependent lanes, use the step ID
-from the configuration. The selected step and dependents are reported as
-`forced_recompute`; reusable predecessors are reported as `reused`:
-
-```sh
-go run ./cmd/notarius run dnd-npc-grounded \
-  --config examples/dnd-npc-grounded.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --resume --recompute-step grounded-events
-```
-
-Checkpoint decisions use these categories: `reused`, `executed`,
-`forced_recompute`, and `dependency_invalidated`. The reason code and bounded
-detail identify the decision without exposing reference content, local paths,
-or secrets. `--recompute-step` requires checkpoint recording and `--resume`;
-unknown step IDs, empty values, and combinations with `--only` are rejected.
-The operator meanings of checkpoint reason codes are maintained in
-[Operations](operations.md#resume-and-selective-recompute).
-
-Use `--debug` to retain the redacted summary and trace bundle for one run. The
-bundle is allocated before pipeline resolution; once allocated, its path is
-also printed to stderr if the command fails. Debug-write failures cause exit
-code `1`.
-
-```sh
-go run ./cmd/notarius run dnd-session \
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --debug --debug-dir ./notarius-debug
-```
-
-Use `refresh` when intentionally replacing the cached plan for the same source:
-
-```sh
-go run ./cmd/notarius run dnd-session \
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --chunk_cache refresh
-```
-
-Use `bypass` for a one-off run that must not inspect or create plan-cache state:
-
-```sh
-go run ./cmd/notarius run dnd-session \
-  --config examples/dnd-spells.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --chunk_cache bypass
-```
-
-`--diagnostics-dir` has been removed. For checkpoint behavior, durable output,
-debug-bundle lifecycle, and failure inspection, see [Operations](operations.md).
-
-## `config validate`
-
-`notarius config validate` loads and validates configuration.
-
-Flags:
-
-- `--config path`: config file path. If omitted, Notarius uses the discovery
-  rules in [Configuration](config.md#discovery).
-- `--pipeline pipeline-id`: additionally resolve one configured pipeline against
-  the production module catalog.
-- `--only lane-a,lane-b`: validate resolution for selected artifact lanes. This
-  flag requires `--pipeline`.
 
 Examples:
 
-```sh
+~~~
 go run ./cmd/notarius config validate \
-  --config examples/dnd-spells.config.yml
+  --config examples/dnd-minimal.config.yml \
+  --pipeline dnd-session
 
-go run ./cmd/notarius config validate \
+OPENROUTER_API_KEY=validation-placeholder \
-  --config examples/dnd-spells.config.yml \
+  go run ./cmd/notarius config validate \
-  --pipeline dnd-session \
+    --config examples/dnd-complete.config.yml \
-  --only spells
+    --pipeline dnd-session
-```
+~~~
 
-## `pipelines list`
+The placeholder in the second command is sufficient only for offline
+validation; it cannot run a provider-backed pipeline.
 
-`notarius pipelines list` prints configured pipeline IDs in sorted order.
+## pipelines list
 
-Flags:
+~~~
+notarius pipelines list [--config path/to/config.yml] [--json]
+~~~
 
-- `--config path`: config file path. If omitted, Notarius uses the discovery
+This command lists configured pipeline IDs in sorted order. By default, it
-  rules in [Configuration](config.md#discovery).
+writes one ID per line to standard output. **--json** writes an object shaped as
-- `--json`: print `{"pipelines":[...]}` instead of one ID per line.
+{"pipelines":[...]} instead.
 
-Examples:
+~~~
-
-```sh
 go run ./cmd/notarius pipelines list \
-  --config examples/dnd-spells.config.yml
+  --config examples/dnd-minimal.config.yml
+~~~
 
-go run ./cmd/notarius pipelines list \
+## Output Streams And Exit Statuses
-  --config examples/dnd-spells.config.yml \
-  --json
-```
 
-## Exit Codes
+Successful commands write their primary result to standard output. Warnings and
+errors are written to standard error.
 
-- `0`: command succeeded.
+For **run --json**, warnings remain on standard error and standard output is a
-- `1`: command syntax was valid, but loading config, resolving modules, running
+machine-readable success result only. Syntax and runtime diagnostics remain on
-  the pipeline, calling the provider, writing output, or writing a requested
+standard error. Parse the result only after the process exits with status 0.
-  debug bundle failed.
-- `2`: command syntax was invalid, a command was unknown, a required argument
-  was missing, or a flag value was malformed.
 
-For YAML structure, defaults, Scriptorium profile sources, environment
+| Status | Meaning |
-overrides, and selectable module and validator keys, see
+| --- | --- |
-[Configuration](config.md).
+| 0 | The command completed successfully, including root help. |
+| 1 | Command syntax was valid but configuration loading or validation, pipeline resolution or execution, provider use, output, or requested debug handling failed. |
+| 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
+parser as a usage error: it writes an error to standard error and exits with
+status 2.

docs/consumers/subprocess.md

@@ -0,0 +1,77 @@
+# Using Notarius As A Subprocess
+
+Use this workflow when an orchestrator runs Notarius and consumes its published
+artifacts. The [CLI reference](../cli.md) owns invocation syntax and exit
+statuses, while the [run-result receipt](../integrations/run-result.md) and
+[Published JSON Output contract](../integrations/json-output.md) own the
+durable result formats.
+
+## 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
+```
+
+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.
+
+```sh
+notarius run pipeline-id \
+  --config /path/to/notarius.yml \
+  --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.
+
+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
+diagnosis and ignore all standard-output bytes: a failed receipt write may have
+left a partial document.
+
+## Discover Required Artifacts
+
+Decode the successful receipt and accept the schema versions supported by the
+caller. Use its `output_directory` as the bundle root. For the production JSON
+output, resolve `index_file` under that root with a confinement check and reject
+an absolute path or a result that escapes the root.
+
+Read the resulting `index.json` and locate each artifact by `lane_id`, not by a
+guessed filename. Before decoding a selected payload, verify its descriptor's
+media type and schema identity against the relevant published artifact
+contract. The JSON bundle contract links to the available lane contracts.
+
+If `index.json` has an `evidence_context` descriptor, treat it as a
+pipeline-wide artifact rather than a lane entry. Verify its six descriptor
+fields before decoding the linked file according to the [Published Evidence
+Context contract](../integrations/evidence-context.md). Decode its top-level
+source-unit array as a reading excerpt. Obtain authoritative citations and lane
+provenance from the normalized lane artifacts; the excerpt has neither and its
+nearby units do not widen a lane artifact's cited source reference.
+
+A zero exit status may still report rejected outputs, warnings, or absent
+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.
+
+## 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
+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
+provenance records. An evidence-context artifact contains source-unit text and
+metadata and can cover most of an input; preserve and share it only when that
+source content is authorized for the recipient.

docs/development.md

@@ -17,11 +17,13 @@ implemented component map.
 | Application shape, package boundaries, contracts, dependency direction, runtime guarantees, or safety properties | [Architecture](policy/architecture.md) and relevant [ADRs](adr/) | Architecture defines the intended system and its invariants; ADRs preserve significant decision rationale. |
 | 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. |
+| 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) | It documents implemented module contracts, capabilities, assets, and registration. |
+| 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 Scriptorium integration. |
+| 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. |
-| CLI or user-visible configuration behavior | [CLI Reference](cli.md) and [Configuration](config.md) | These are the canonical user and operator references. |
 | External input formats, artifact schemas, or durable output files | [Integration Contracts](integrations/) | Integration documents define external and durable data contracts. |
 | Proposed or unimplemented behavior | [Roadmap](roadmap/) | Future work belongs only in roadmap documentation until implemented. |
 

docs/integrations/chunk-map.md

@@ -0,0 +1,81 @@
+# Accepted Chunk Map
+
+This document defines the optional durable `chunk-map.json` artifact in a
+[published JSON bundle](json-output.md). It describes the accepted,
+materialized chunk plan used by one run. It is not a lane payload and is never
+an input to a later pipeline step.
+
+## Contract Identity
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `source/chunk-map` |
+| Logical file | `chunk-map.json` |
+| Media type | `application/json` |
+| Schema ID | `notarius.source.chunk_map` |
+| Schema name | `notarius_source_chunk_map_v1` |
+| Schema version | `v1` |
+
+The optional `chunk_map` descriptor in `index.json` identifies this artifact.
+Export is controlled by the JSON output binding described in
+[Configuration](../config.md#module-bindings-and-validators).
+
+## Wire Shape
+
+Every payload has these required fields:
+
+| Field | Meaning |
+| --- | --- |
+| `source_id` | Accepted source-document identity. |
+| `source_digest` | Lower-case `sha256:` digest of that source document. |
+| `plan_digest` | Lower-case `sha256:` digest of the logical chunk plan. |
+| `requested_chunker` | Chunk module selected by the resolved pipeline. |
+| `producer` | Original accepted-plan producer. `input_module` and `chunk_module` are required; `llm_profile` is optional. |
+| `plan_annotations` | Plan-level annotation namespace map; `{}` when none are present. |
+| `chunks` | Non-empty execution-order chunk collection. |
+
+Each `chunks` entry contains non-empty `id`, zero-based `index`, `source_ref`,
+positive `unit_count`, and an explicit `annotations` map. `source_ref` contains
+the same `source_id` as the top-level value plus positive inclusive
+`start_unit_id` and `end_unit_id` values. Endpoints identify source units; their
+numeric values do not by themselves establish source-document order.
+
+Annotation namespaces are non-empty trimmed strings. Their values are arbitrary
+valid JSON and are retained without interpreting a module-specific namespace.
+
+## Ordering And Validation
+
+`chunks` are in execution order. Their indexes are contiguous, start at zero,
+and equal their array positions; chunk IDs are unique. The emitted map is built
+only after the selected plan has been accepted and materialized against the
+source document, so its ranges, unit counts, annotations, and digests describe
+that exact plan.
+
+The codec rejects malformed JSON, trailing content, unknown fixed-object
+fields, invalid identities or digests, invalid annotations, duplicate chunk
+IDs, non-contiguous indexes, and a `plan_digest` that does not match the
+reconstructed logical plan. The checked-in
+[schema](../../internal/framework/chunkmap/assets/schemas/source_chunk_map.v1.json)
+defines the strict JSON shape.
+
+## Valid Example
+
+The compact
+[source chunk-map fixture](../../internal/framework/chunkmap/testdata/source_chunk_map.v1.json)
+is decoded by the production codec and demonstrates an accepted map with
+annotations, producer identity, and ordered chunks.
+
+## Publication And Compatibility
+
+The map is present only when a chunk plan was accepted and its export is
+enabled. It remains publishable if a later lane is rejected, but is absent when
+chunk-plan validation rejects the plan. `requested_chunker` identifies the
+current pipeline selection, while `producer` identifies the component that
+originally produced the accepted plan; they may differ when an accepted plan is
+reused.
+
+The map contains structure rather than source content: it excludes transcript
+bytes, source-unit metadata, chunk text, private model output, reference
+content, debug data, and filesystem paths. Treat the exported map with the
+same care as other published output. Publication location and retention are
+defined in [Operations](../operations.md#output-bundles).

docs/integrations/dnd-combat-turn-artifacts.md

@@ -1,10 +1,10 @@
-# D&D Combat-Turn Artifact Contract
+# D&D Combat-Turn Artifact
 
-This document defines the durable artifact, serialization, extraction,
+This contract defines the durable combat-action occurrence list produced by
-candidate-validation, normalization, and production lane boundaries for D&D
+`dnd/combat-turns`. It records source-grounded turns and actions; it is not a
-combat turns.
+complete initiative tracker, combat summary, or state model.
 
-## Artifact identity
+## Identity and compatibility
 
 | Property | Value |
 | --- | --- |
@@ -14,150 +14,58 @@ combat turns.
 | Schema version | `v1` |
 | Media type | `application/json` |
 
-The top-level JSON object contains the required `combat_turns` array, which
+`v1` is a strict JSON object with required `combat_turns`; the array may be
-may be empty. Every object rejects unknown fields.
+empty. Turn and source-reference objects reject unknown fields. An incompatible
+shape change requires a new schema version.
 
-## JSON shape
+## Wire shape
 
-Each combat turn contains these required fields:
+Each combat turn has these required fields:
 
-| Field | Shape |
+| Field | Contract |
 | --- | --- |
-| `actor` | Non-empty string. |
+| `actor` | Non-empty acting character or creature name. |
-| `turn_kind` | One of `turn`, `reaction`, `legendary_action`, `lair_action`, or `other`. |
+| `turn_kind` | `turn`, `reaction`, `legendary_action`, `lair_action`, or `other`. |
-| `source_refs` | Required array with at least one source reference. |
+| `source_refs` | One or more transcript evidence ranges. |
 
-Source references use the shared source-reference shape:
+Each source reference has exactly `source_id`, `start_unit_id`, and
+`end_unit_id`. It identifies an inclusive current-transcript range; unit IDs
+are positive and the start may not follow the end.
 
 ```json
 {
-  "source_id": "session-alpha",
+  "combat_turns": [
-  "start_unit_id": 1,
+    {
-  "end_unit_id": 2
+      "actor": "Mira Thorn",
+      "turn_kind": "turn",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 31, "end_unit_id": 32}
+      ]
+    }
+  ]
 }
 ```
 
-`source_id` must be non-empty and both unit IDs must be positive integers. The
+## Eligibility, evidence, and normalized form
-codec does not resolve references against a source document or enforce source
-range ordering; those checks belong to the later source-reference validation
-boundary.
 
-## Codec behavior
+The extractor requires an approved [scene-description artifact](dnd-scene-description-artifacts.md).
+It emits combat turns only for a chunk with an exact matching scene classified
+`combat`; an exact non-combat scene produces an accepted empty list. The scene
+record controls eligibility only: its title, summary, and reference do not
+become turn evidence. No exact matching scene also produces an empty list and
+the `scene_classification_unavailable` warning.
 
-The codec exposes two representations of the same typed artifact:
+An optional normalized [NPC registry artifact](dnd-npc-registry-artifacts.md) can ground an
+actor name. Its registry references are provenance, never combat evidence.
+Normalization trims and, where possible, canonicalizes actor names; orders and
+deduplicates exact source references; orders valid-evidence turns by source
+chronology; and collapses only duplicates with the same actor identity, turn
+kind, and complete valid evidence. It does not infer turns, initiative, or
+actions from registry or scene data.
 
-- Candidate encode/decode preserves invalid actor and turn-kind values,
+The [NPC-occurrence artifact](dnd-npc-occurrence-artifacts.md) records
-  collection presence, and source references so later validators can report
+broader NPC occurrences. The [enemy-event artifact](dnd-enemy-event-artifacts.md)
-  them. Candidate decoding still requires valid JSON, one JSON value, known
+uses combat turns as grounding only; turns do not establish an enemy event or
-  fields, and compatible JSON types.
+its outcome. The [JSON output contract](json-output.md) defines publication,
-- Approved encode/decode enforces the structural rules in this contract.
+and [D&D module internals](../internal/dnd.md) describes routing and validation
-
+mechanics.
-The codec owns the durable JSON Schema, whose object layers all set
-`additionalProperties` to `false`. Codec metadata contains only
-`combat_turn_count`.
-
-The maintained compact fixture is
-`internal/modules/dnd/codec/combatturns/testdata/dnd_combat_turns.v1.json`.
-
-## Extraction boundary
-
-The standalone extractor uses these identities:
-
-| Property | Value |
-| --- | --- |
-| Extractor key | `dnd/combat-turns` |
-| Capability | `dnd.combat_turns` |
-| Prompt ID | `dnd.combat_turns` |
-| Prompt version | `v1` |
-| Private response-schema key | `dnd_combat_turns_llm` |
-| Private response-schema ID | `notarius.dnd.combat_turns.llm` |
-| Default profile | `gemini-2-flash` |
-
-It requires `chunks` and `source.transcript`, accepts no options, and makes one
-structured completion for each supplied chunk. The prompt receives the
-chunk-scoped transcript plus the existing `players`, `party`, and `glossary`
-inputs, and optionally the deprecated `roster` reference through the shared
-party mapping. The optional `npcs` reference is an approved normalized NPC
-artifact used only for identity grounding; it never supplies combat evidence.
-An external file is validated during preparation. In an ordered pipeline, the
-same slot may receive the producer's canonical generated artifact at the step
-handoff.
-
-The private response envelope has the same turn fields and JSON types as the
-durable shape except that source references contain only `start_unit_id`
-and `end_unit_id`. It enforces required field presence, types, and
-unknown-field rejection, while deterministic validators own enum membership,
-non-empty values and collections, and positive-number requirements. The
-extractor assigns the current source ID, removes exact duplicate ranges, and
-stable-sorts turns by the earliest valid source-document position. Numeric unit
-IDs are identifiers; source-document slice position determines chronology.
-Semantically malformed candidate fields remain in the typed result for the
-configured validation and retry boundary.
-
-## Deterministic candidate validation
-
-The standalone validator keys are:
-
-| Validator | Responsibility |
-| --- | --- |
-| `extract/dnd/combat-turns/shape` | Required list, actor, turn kind, and source references, plus supported turn-kind values. |
-| `extract/dnd/combat-turns/source_refs` | Source identity, source-unit existence, and range order through the source document. |
-| `extract/dnd/combat-turns/source_relatedness` | At most one advisory warning per turn when the actor is not related to cited transcript text. |
-
-Source-reference and relatedness validators defer malformed shape to the shape
-validator. Relatedness also defers when any cited source range is invalid. It
-combines overlapping cited ranges once in document order and compares actors
-with the shared Unicode-aware NPC identity policy.
-
-The production D&D registrar exposes the extractor and these validators. Its
-default extraction chain preserves this order: JSON syntax, combat shape,
-source references, private response schema, then source relatedness.
-
-## Normalization boundary
-
-The standalone normalizer uses key `dnd/combat-turns`, requires `merged`,
-provides `normalized`, accepts no options, and accepts only the optional
-structured `npcs` reference. Campaign references are LLM extraction context and
-are not normalizer inputs. For an external file, the NPC registry is resolved
-during preparation; for a generated binding, it is resolved at the operation-
-time handoff. Runtime normalization uses that immutable prepared or handed-off
-view.
-
-Normalization policy is `dnd.combat_turns.normalize.v1`. It display-normalizes
-the actor, canonicalizes exact registry actor matches, orders and deduplicates
-exact source references, stable-sorts records by earliest valid source-document
-position, and collapses only records with the same actor identity, turn kind,
-and complete valid evidence set. The first normalized record is retained.
-Invalid evidence is never eligible for duplicate collapse. Every mutation and
-collapse emits a bounded warning using the merged input index in its scope.
-
-The normalizer reports `normalization_policy` and `identity_policy` metadata
-and fingerprints. An external registry may additionally contribute
-`npc_registry_digest` and `npc_count`; generated registry identity is retained
-in framework handoff provenance and dependency fingerprints. The
-normalized-invariants validator is
-`normalize/dnd/combat-turns/invariants`; it defers shape and source-reference
-failures, then checks actor display normalization, canonical evidence ordering,
-chronology, and duplicate identity. It rejects
-with `invalid_combat_turn_normalization` under policy
-`dnd.combat_turns.validator.normalized.v1`.
-
-The production D&D registrar exposes the normalizer and normalized-invariants
-validator. Its default normalization chain is JSON syntax, combat shape,
-normalized invariants, source references, durable schema, then source
-relatedness. The lane uses the framework's typed append-order merger and has no
-merge validator chain.
-
-## Production manifest and references
-
-The selectable lane uses extractor and normalizer key `dnd/combat-turns`,
-`appendorder` for the typed merger, and the durable codec above. A bound `npcs`
-reference contributes raw-file provenance to the run manifest. A generated
-binding contributes artifact kind, schema identity, media type, canonical
-digest, size, and bounded producer provenance. Consumer metadata and checkpoint
-fingerprints contain no registry names, content, paths, or NPC source ranges.
-The component-local registry fingerprint covers only the names projected to the
-consumer, while manifest provenance retains the full artifact digest. The
-normalized lane is emitted as `lanes/<lane-id>.json` by the JSON
-output module, and warnings and rejection summaries remain in their shared
-companion files.

docs/integrations/dnd-enemy-event-artifacts.md

@@ -0,0 +1,114 @@
+# D&D Enemy-Event Artifact
+
+This contract defines the durable, source-grounded enemy-event occurrence list.
+It records enemies directly established as opposing the party and explicitly
+observed combat outcomes. It is an ordered observation artifact from which a
+consumer may derive a ledger; it is not a ledger, encounter roster, or terminal
+state model.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/enemy-event-list` |
+| Schema ID | `notarius.dnd.enemy_events` |
+| Schema name | `notarius_dnd_enemy_events_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+
+`v1` is a strict JSON object with required `events`; the array may be empty.
+Event and source-reference objects reject unknown fields. An incompatible shape
+change requires a new schema version.
+
+## Wire shape
+
+Every event has these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `name` | Non-empty display name or directly grounded collective subject label. |
+| `kind` | `engaged`, `killed`, `fled`, `captured`, or `incapacitated`. |
+| `source_refs` | One or more current-transcript evidence ranges. |
+
+Each source reference has exactly `source_id`, `start_unit_id`, and
+`end_unit_id`. It identifies an inclusive current-transcript range; unit IDs
+are positive and the start may not follow the end.
+
+```json
+{
+  "events": [
+    {
+      "name": "Ashfang",
+      "kind": "engaged",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 41, "end_unit_id": 42}
+      ]
+    },
+    {
+      "name": "Ashfang",
+      "kind": "fled",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 57, "end_unit_id": 58}
+      ]
+    }
+  ]
+}
+```
+
+## Event semantics and evidence
+
+| Kind | Required evidence |
+| --- | --- |
+| `engaged` | The subject is directly established as actively opposing the party in combat. At most one engagement is emitted for one subject in one combat scene. |
+| `killed` | The transcript explicitly establishes that the subject died or was killed. Damage, defeat, disappearance, or combat ending is insufficient. |
+| `fled` | The subject explicitly escapes, retreats, or otherwise leaves combat to avoid continued engagement. Movement or absence from later turns is insufficient. |
+| `captured` | The subject is explicitly taken prisoner or secured under the party's control. A grapple or temporary restraint alone is insufficient. |
+| `incapacitated` | The subject is explicitly rendered unable to continue acting without being established as killed or captured. A missed turn is insufficient. |
+
+The current transcript is the only event evidence. Campaign context and
+normalized NPC, scene-description, combat-turn, and NPC-occurrence artifacts
+can ground names or control combat eligibility, but none may supply event
+evidence. An outcome may share evidence with an engagement, in which case both
+events are retained.
+
+Extraction is limited to chunks with an exact combat-scene classification. An
+exact non-combat classification produces an accepted empty list. Missing or
+mismatched classification also produces an accepted empty list and a
+`scene_classification_unavailable` warning.
+
+## Subjects, normalization, and order
+
+A subject matching the normalized NPC registry uses that registry's canonical
+display name. Unmatched hostile creatures, summoned entities, and directly
+grounded groups remain valid subjects. An unnamed homogeneous group uses the
+narrowest transcript-grounded label, such as `Orcs`, `One orc`, or `Remaining
+orcs`; the artifact never invents synthetic member identities or quantities.
+Party members, allies, neutral observers, mentioned-but-absent enemies, hazards,
+traps, and environmental effects are excluded.
+
+Normalization collapses surrounding and repeated internal whitespace in subject
+display values, canonicalizes recognized registry names, canonicalizes and
+deduplicates exact source ranges, then orders events by valid evidence
+chronology, normalized subject identity, display name, kind, and reference
+sequence. The deterministic kind tie order is `engaged`,
+`incapacitated`, `captured`, `fled`, then `killed`. Only entries with the same
+normalized name, kind, and complete canonical evidence sequence are collapsed.
+Different kinds, evidence, repeated engagement in separate scenes, and later
+outcomes remain separate. A later engagement for the same named subject is
+preserved after an earlier outcome because the artifact does not assert an
+irreversible state transition.
+
+## Non-goals
+
+The artifact has no NPC or scene ID, quantity, confidence, description,
+rationale, summary, current state, or inferred terminal outcome. It does not
+emit `active` or `unresolved`; consumers may derive an unresolved ledger view
+only when an engagement has no later explicit outcome. It never infers an
+outcome from turn absence, scene termination, initiative order, hit-point
+guesses, or other artifacts.
+
+The [JSON output contract](json-output.md) defines publication. Configuration
+keys, required generated-reference slots, and validator-chain selection are
+defined in the [configuration reference](../config.md). Implementation and
+prompt-grounding mechanics are described in the
+[D&D module internals](../internal/dnd.md).

docs/integrations/dnd-item-occurrence-artifacts.md

@@ -0,0 +1,72 @@
+# D&D Item-Occurrence Artifact
+
+`dnd/item-occurrences` currently produces this source-grounded item and currency
+occurrence list. It records discoveries and possession changes, not an
+inventory, balance, or ledger.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/item-occurrence-list` |
+| Schema ID | `notarius.dnd.item_occurrences` |
+| Schema name | `notarius_dnd_item_occurrences_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+
+`v1` accepts one strict JSON object with required `occurrences`; the array may
+be empty. Each occurrence has required `item_id`, `name`, `kind`, and
+`source_refs`, and occurrence and source-reference objects reject unknown
+fields. `quantity`, `from`, and `to` appear only when their kind permits them.
+An incompatible shape change requires a new schema version.
+
+## Registry grounding
+
+Both extraction and normalization require an `item_registry` reference bound to
+an earlier normalized `dnd/item-registry` artifact. The registry is immutable
+for an operation and contributes names-only grounding after the shared evidence
+message. Notarius resolves the model's selected name into the unchanged exact
+durable ID/name pair. It is never occurrence evidence.
+
+Each occurrence must use one exact registry ID/name pair. An extraction response
+with an unknown or ambiguous selected name is rejected as invalid model output;
+the configured pipeline may retry it and never accepts a partial artifact.
+Normalization and validation remain defense in depth for artifacts entering
+through other boundaries: normalization canonicalizes a recognized name by ID,
+preserves unknown values for the registry validator, and the registry validator
+rejects unknown or mismatched pairs.
+
+## Wire shape
+
+Each source reference has exactly `source_id`, `start_unit_id`, and
+`end_unit_id`. It identifies an inclusive range in the current transcript;
+unit IDs are positive and the start may not follow the end.
+
+```json
+{
+  "occurrences": [
+    {
+      "item_id": "item:sha256:…",
+      "name": "Silver Pieces",
+      "kind": "acquired",
+      "quantity": 20,
+      "to": "party",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 2, "end_unit_id": 2}
+      ]
+    }
+  ]
+}
+```
+
+The five kinds remain `discovered`, `acquired`, `lost`, `consumed`, and
+`transferred`. Holder, quantity, currency, ordering, and exact-duplicate rules
+are unchanged: discovered has no holder; acquired requires `to`; lost and
+consumed require `from`; transferred requires distinct non-`party` holders.
+The only current downstream compatibility requirement is its registry handoff;
+the normalized occurrence list is otherwise published for callers. See
+[Configuration](../config.md#d-d-reference-slots) for the binding and
+[JSON output](json-output.md) for publication.
+
+See [item registry](dnd-item-registry-artifacts.md) for the grounding artifact
+and [D&D module internals](../internal/dnd.md) for implementation details.

docs/integrations/dnd-item-registry-artifacts.md

@@ -0,0 +1,95 @@
+# D&D Item Registry Artifact
+
+This contract defines the durable, source-grounded item registry produced by
+`dnd/item-registry`. It records transcript-established item types and unique
+designations for one source document; it is not an inventory, holder record,
+quantity ledger, or item-occurrence artifact.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/item-registry` |
+| Schema ID | `notarius.dnd.item_registry` |
+| Schema name | `notarius_dnd_item_registry_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+| Identity policy | `dnd.item_registry.identity.v1` |
+
+`v1` accepts one strict JSON object with required `items`; the array may be
+empty. Item and source-reference objects reject unknown fields. An incompatible
+artifact shape or identity-policy change uses a new version or policy.
+
+## Wire shape and identity
+
+Each item has these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `id` | `item:sha256:` followed by 64 lowercase hexadecimal characters. |
+| `name` | Non-empty transcript-established item type or unique designation. |
+| `source_refs` | One or more transcript evidence ranges that establish the item. |
+
+A source reference has exactly `source_id`, `start_unit_id`, and `end_unit_id`.
+The source ID identifies the transcript, unit IDs are positive inclusive unit
+identifiers, and the start may not follow the end.
+
+```json
+{
+  "items": [
+    {
+      "id": "item:sha256:31e73b6280ef98e4d8070e07fd4de9b2c3e842cc03af1a09ca631cb95b73e3b3",
+      "name": "Star Compass",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 4, "end_unit_id": 5}
+      ]
+    }
+  ]
+}
+```
+
+The ID is deterministic for an item name or type, rather than for one physical
+instance. Notarius normalizes the display name for comparison with Unicode
+NFKC, supported apostrophe normalization, collapsed whitespace, and case
+folding. It hashes compact JSON for this array:
+
+```text
+["dnd.item_registry.identity.v1", comparison_name]
+```
+
+The canonical ID is the lowercase SHA-256 digest of those bytes with the
+`item:sha256:` prefix. Equal comparison names represent one item identity;
+normalization unions their transcript evidence when it safely consolidates a
+candidate group.
+
+## Scope, reconciliation, and evidence
+
+The registry includes named unique items, concrete reusable item types, stable
+unique designations, and separately established currency denominations. It
+excludes vague loot or treasure, generic weapons, quantities, inferred
+properties, and inferred uniqueness. Capitalization alone does not establish
+eligibility.
+
+Normalization first applies deterministic display, evidence, and ID rules. It
+then may use a bounded LLM-assisted proposal to reconcile semantically duplicate
+records. The proposal may choose only a supplied candidate display name;
+invalid, uncertain, overlapping, or unsafe proposals retain the deterministic
+result with retry or fallback diagnostics. A proposal that mixes a recognized
+currency denomination with a non-currency item, or combines recognized
+denominations, is unsafe and retains every deterministic record. Currency
+denominations, materially different item types, and merely nearby objects
+remain distinct. Source references establish registry provenance, not evidence
+for later artifacts.
+
+## Consumers and publication
+
+`dnd/item-occurrences` requires one approved item registry through its
+`item_registry` reference slot for both extraction and normalization. Its
+consumer receives names-only grounding; Notarius resolves the selected name
+into the unchanged exact durable ID/name pair. The registry’s source references
+are never occurrence evidence. Unknown or ambiguous selections are rejected by
+the occurrence contract. See the
+[item-occurrence artifact](dnd-item-occurrence-artifacts.md) for that strict
+wire contract, [Configuration](../config.md#d-d-reference-slots) for binding
+rules and validator selection, and the [JSON output contract](json-output.md)
+for publication.

docs/integrations/dnd-location-occurrence-artifacts.md

@@ -0,0 +1,86 @@
+# D&D Location-Occurrence Artifact
+
+This contract defines the durable occurrence list produced by
+`dnd/location-occurrences`. It records source-grounded ways the party relates
+to locations in a required normalized location registry; it does not extend
+that registry or infer a place absent from it.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/location-occurrence-list` |
+| Schema ID | `notarius.dnd.location_occurrences` |
+| Schema name | `notarius_dnd_location_occurrences_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+
+`v1` accepts one strict JSON object with required `occurrences`; the array may
+be empty. Occurrence and source-reference objects reject unknown fields. An
+incompatible shape change requires a new schema version.
+
+## Wire shape
+
+Each occurrence has these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `location_id` | Exact ID from the required normalized [location registry](dnd-location-registry-artifacts.md). |
+| `name` | Exact canonical display name for `location_id` in that registry. |
+| `kind` | One of `visited`, `planned`, `recalled`, or `mentioned`. |
+| `source_refs` | One or more current-transcript evidence ranges for this occurrence. |
+
+A source reference has exactly `source_id`, `start_unit_id`, and `end_unit_id`.
+It identifies an inclusive range in the current transcript; unit IDs are
+positive and the start may not follow the end.
+
+```json
+{
+  "occurrences": [
+    {
+      "location_id": "location:sha256:fb05475da0fc7debf994b517e1906ffe7209887a6a1ec306356d84de820b1a24",
+      "name": "Moon Gate",
+      "kind": "visited",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 12, "end_unit_id": 13}
+      ]
+    }
+  ]
+}
+```
+
+## Occurrence categories
+
+| Kind | Meaning |
+| --- | --- |
+| `visited` | The transcript establishes physical party presence, including arrival, continuing presence, or departure. |
+| `planned` | The party explicitly proposes, intends, or agrees to future travel; speculation alone is not enough. |
+| `recalled` | The transcript explicitly recounts prior party presence before the current live events. |
+| `mentioned` | The location is explicit but no stronger category applies, including lore, directions, third-party activity, non-actionable speculation, a mere hypothetical reference, or out-of-character discussion. |
+
+For overlapping evidence, precedence is `visited`, then `planned`, then
+`recalled`, then `mentioned`. For example, “What if we went to Moon Gate?” is
+eligible as `mentioned` when its narrow evidence explicitly references that
+registry location, but it is not `planned` without an actual proposal,
+intention, or agreement to travel. Inferred, unstated, uncertain, and
+unsupported places or occurrences are omitted. Normalization
+canonicalizes the registry name, orders and deduplicates source references, and
+orders occurrences by source chronology, location ID, name, kind, and reference
+sequence. It collapses only exact duplicates with the same ID, kind, and
+complete canonical evidence sequence.
+
+## Required grounding and evidence
+
+Both extraction and normalization require exactly one `location_registry` reference of
+kind `dnd/location-registry`, media type `application/json`, and at most 1 MiB. The
+registry provides identity grounding only. The model selects a supplied
+contextual name-and-registry-reference descriptor, and Notarius resolves it
+into the exact durable ID/name pair. Unknown, partial, or ambiguous selections
+are rejected rather than guessed or reassigned. The current transcript is the
+only evidence source for an occurrence; registry evidence and provenance never
+become occurrence evidence.
+
+See [Configuration](../config.md#d-d-reference-slots) for the selectable slot
+and generated-handoff compatibility, [D&D module internals](../internal/dnd.md)
+for implementation behavior, and the [JSON output contract](json-output.md)
+for publication.

docs/integrations/dnd-location-registry-artifacts.md

@@ -0,0 +1,93 @@
+# D&D Location Registry Artifact
+
+This contract defines the durable, source-grounded location registry produced
+by `dnd/location-registry`. It records transcript-established physical places for one
+source document; it is not a map, location hierarchy, campaign-wide world
+registry, or location description.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/location-registry` |
+| Schema ID | `notarius.dnd.location_registry` |
+| Schema name | `notarius_dnd_location_registry_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+| Identity policy | `dnd.location_registry.identity.v1` |
+
+`v1` accepts one strict JSON object with required `locations`; the array may be
+empty. Location and source-reference objects reject unknown fields. An
+incompatible artifact shape or identity-policy change uses a new version or
+policy.
+
+## Wire shape and identity
+
+Each location has these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `id` | `location:sha256:` followed by 64 lowercase hexadecimal characters. |
+| `name` | Non-empty transcript-established display name. |
+| `source_refs` | One or more transcript evidence ranges that identify the place. |
+
+A source reference has exactly `source_id`, `start_unit_id`, and `end_unit_id`.
+The source ID identifies the transcript, unit IDs are positive inclusive unit
+identifiers, and the start may not follow the end.
+
+```json
+{
+  "locations": [
+    {
+      "id": "location:sha256:fb05475da0fc7debf994b517e1906ffe7209887a6a1ec306356d84de820b1a24",
+      "name": "Moon Gate",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 4, "end_unit_id": 5}
+      ]
+    }
+  ]
+}
+```
+
+The ID is deterministic and scoped to the source document. Notarius normalizes
+the display name for comparison with Unicode NFKC, supported apostrophe
+normalization, collapsed whitespace, and case folding. It hashes compact JSON
+for this array, using the earliest canonical source reference as the anchor:
+
+```text
+["dnd.location_registry.identity.v1", comparison_name, source_id, start_unit_id, end_unit_id]
+```
+
+The canonical ID is the lowercase SHA-256 digest of those bytes with the
+`location:sha256:` prefix. Equal display names are allowed when their evidence
+anchors differ, so a generic name does not force distinct places to collapse.
+
+## Scope, reconciliation, and evidence
+
+Locations are physical or spatial places established by the transcript with a
+stable proper name or unique in-world designation, such as named planes,
+regions, settlements, districts, buildings, rooms, landmarks, routes, and
+geographic features. Generic, temporary, relative, and descriptive phrases
+such as “the room,” “the bar,” “the hallway,” “outside,” and “upstairs” are not
+registry locations. Capitalization alone does not establish eligibility.
+Notarius does not infer an unstated place or add hierarchy, coordinates,
+descriptions, participants, or ownership.
+
+Normalization first applies deterministic display, evidence, and ID rules. It
+then may use a bounded LLM-assisted proposal to reconcile semantically duplicate
+records. The proposal is validated and applied conservatively; invalid or
+unusable proposals retain the deterministic result with retry or fallback
+diagnostics. The registry's source references establish registry provenance,
+not evidence for later artifacts.
+
+## Consumers and publication
+
+`dnd/location-occurrences` requires one approved location registry through its
+`location_registry` reference slot. Its prompt receives contextual selectors
+containing a canonical name and registry references; Notarius resolves a
+selection into the unchanged exact durable ID/name pair. Registry references
+must not be treated as occurrence evidence. See the
+[location-occurrence artifact](dnd-location-occurrence-artifacts.md)
+for that contract, [Configuration](../config.md#references-and-ordered-handoffs)
+for binding rules, and the [JSON output contract](json-output.md) for
+publication.

docs/integrations/dnd-npc-artifacts.md

@@ -1,139 +0,0 @@
-# D&D NPC Artifact
-
-This document defines the durable D&D NPC-list artifact, its JSON codec, and
-the selectable production NPC pipeline. The normalized JSON payload can be
-passed explicitly to the spell extractor as an optional caster-name registry
-or to the combat extractor and normalizer as an actor registry. It
-remains a reference, not spell or combat evidence.
-
-## Identity
-
-- Artifact kind: `dnd/npc-list`
-- Durable schema ID: `notarius.dnd.npcs`
-- Durable schema name: `notarius_dnd_npcs_v1`
-- Durable schema version: `v1`
-- Media type: `application/json`
-- Identity policy: `dnd.npcs.identity.v1`
-
-The durable JSON Schema is owned by the D&D NPC codec. NPC IDs are derived from
-the Unicode-normalized, case-folded canonical name using the identity policy.
-The durable codec enforces the artifact shape and ID syntax; registry identity
-validation remains a separate deterministic concern.
-
-The extractor's private LLM response schema is a separate structural transport
-contract. It omits framework-assigned NPC and source IDs and admits semantic
-candidates for the deterministic shape and source-reference validators; it is
-not part of this durable contract.
-
-## Output Shape
-
-The payload is one object with a required top-level `npcs` array:
-
-```json
-{"npcs": []}
-```
-
-The array may be empty. Every object and nested object rejects unknown fields.
-
-## NPC Fields
-
-Each NPC contains exactly these required fields:
-
-- `id`: `npc:sha256:` followed by 64 lowercase hexadecimal characters;
-- `name`: the canonical display name;
-- `source_refs`: at least one source reference supporting the NPC record.
-
-Each source reference contains required `source_id`, `start_unit_id`, and
-`end_unit_id`; unit IDs are positive integers. Source document identity, unit
-existence, and range ordering are validated by the source-reference validator
-when the artifact is used by a pipeline.
-
-## Codec Boundary
-
-`EncodeCandidate` and `DecodeCandidate` provide strict single-value JSON
-serialization while preserving typed values that still need semantic
-validation. `Encode` and `Decode` are the approved-artifact boundary and
-require all durable structural fields, non-empty required strings, valid source
-reference shapes, and the NPC ID pattern.
-
-Codec metadata contains only `npc_count`. Schema bytes and returned metadata
-are independent values so callers cannot mutate codec-owned state.
-
-## Production Pipeline
-
-The production identities are:
-
-- extractor: `dnd/npcs`;
-- artifact kind: `dnd/npc-list`;
-- normalizer: `dnd/npcs`; and
-- durable schema: `notarius.dnd.npcs`, version `v1`, media type
-  `application/json`.
-
-The extractor maps private model records to the current source identity and
-assigns deterministic IDs. Extraction validation checks shape, source
-references, and source relatedness. The normalizer then consolidates records
-only when their normalized canonical names match, preserves the first record's
-display and output position, unions exact evidence, and validates the retained
-registry's identity. No LLM is used for consolidation.
-
-The extraction prompt asks only for individually identifiable NPC names backed
-by source evidence. Groups, generic roles, invented labels, and descriptive or
-relationship enrichment are outside the contract.
-
-The default extraction chain is `generic/valid_json`,
-`extract/dnd/npcs/shape`, `extract/dnd/npcs/source_refs`,
-`generic/valid_json_schema`, and `extract/dnd/npcs/source_relatedness`. The
-default normalize chain is `generic/valid_json`, `extract/dnd/npcs/shape`,
-`normalize/dnd/npcs/identity`, `extract/dnd/npcs/source_refs`,
-`generic/valid_json_schema`, and `extract/dnd/npcs/source_relatedness`.
-Relatedness emits bounded warnings when an NPC canonical name is not present
-near its cited transcript text; opaque campaign
-references may explain such a warning but do not become evidence.
-
-## Manifest And Artifact Handoff
-
-The NPC extractor records prompt and response-schema identities. The durable
-codec records only `npc_count`; raw names, source references, and payload bytes
-stay in the lane file rather than manifest
-metadata. The normalized lane can be consumed by a later ordered step through
-the registered canonical codec:
-
-```yaml
-steps:
-  - id: identify-npcs
-    artifacts:
-      npcs:
-        extract: dnd/npcs
-        normalize: dnd/npcs
-  - id: grounded-events
-    references:
-      npcs:
-        artifact:
-          step: identify-npcs
-          lane: npcs
-    artifacts:
-      spells:
-        extract: dnd/spells
-        normalize: dnd/spells
-      combat:
-        extract: dnd/combat-turns
-        normalize: dnd/combat-turns
-```
-
-The framework hands only an accepted normalized artifact across the barrier. It
-validates the canonical bytes against each consumer slot and clones the
-operation-time reference for the spell and combat consumers. Generated
-provenance records the artifact kind, schema identity, media type, canonical
-digest, size, and producer step/lane/module, but not names, source
-ranges, or payload bytes. External normalized files remain supported as
-explicit references and retain their file provenance.
-
-NPC source references are registry provenance and are never accepted as spell
-or combat evidence. Current transcript units remain the only event evidence.
-
-Consumers receive a separate names-only projection in normalized registry
-order, for example `{"npcs":[{"name":"Mira Thorn"}]}`. The projection omits
-IDs and evidence. Its digest covers the exact projected bytes and is used for
-consumer-local checkpoint identity, while the full durable artifact digest
-remains the manifest and generated-reference provenance identity. The unbound
-projection is exactly `{"npcs":[]}` and also has a projection digest.

docs/integrations/dnd-npc-occurrence-artifacts.md

@@ -0,0 +1,90 @@
+# D&D NPC Occurrence Artifact
+
+This contract defines the durable occurrence list produced by
+`dnd/npc-occurrences`. It records discrete, source-grounded occurrences with
+NPCs already present in a normalized registry; it does not extend that registry
+or summarize the session.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/npc-occurrence-list` |
+| Schema ID | `notarius.dnd.npc_occurrences` |
+| Schema name | `notarius_dnd_npc_occurrences_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+
+`v1` is a strict JSON object with required `occurrences`; the array may be
+empty. Occurrence and source-reference objects reject unknown fields. An
+incompatible shape change requires a new schema version.
+
+## Wire shape
+
+Each occurrence has these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `npc_id` | Exact durable ID from the required NPC registry. |
+| `name` | Non-empty canonical display name from the required NPC registry. |
+| `kind` | One of the occurrence categories below. |
+| `source_refs` | One or more transcript evidence ranges. |
+
+Each source reference has exactly `source_id`, `start_unit_id`, and
+`end_unit_id`. It identifies an inclusive range in the current transcript;
+unit IDs are positive and the start may not follow the end. Extraction evidence
+for an occurrence is confined to its accepted chunk.
+
+```json
+{
+  "occurrences": [
+    {
+      "npc_id": "npc:sha256:example",
+      "name": "Mira Thorn",
+      "kind": "dialogue",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 12, "end_unit_id": 13}
+      ]
+    }
+  ]
+}
+```
+
+## Occurrence categories
+
+| Kind | Meaning |
+| --- | --- |
+| `mentioned` | The NPC is referred to but is not established as present or communicating. |
+| `noncombat_presence` | The NPC is present and relevant without meaningful dialogue or combat participation. |
+| `dialogue` | The NPC speaks, responds, or meaningfully participates in a non-combat exchange. |
+| `combat_ally` | The NPC actively participates in combat on the party's side. |
+| `combat_opponent` | The NPC actively participates in combat against the party. |
+| `other` | A clearly evidenced direct occurrence not covered by another category. |
+
+The categories do not represent motives, relationships, state, or events that
+the cited transcript does not establish. An `other` entry is not a substitute
+for uncertain classification.
+
+## Identity, evidence, and order
+
+The required normalized [NPC registry artifact](dnd-npc-registry-artifacts.md)
+supplies names-only contextual grounding to the model. Notarius resolves the
+selected name and writes the exact `{npc_id, name}` pair. An unknown or
+ambiguous selection rejects the complete model result; normalization does not
+repair names by similarity. Registry references are provenance only and never
+replace an occurrence's own evidence.
+The registry may include an identity established by a factual third-party
+mention; that provenance alone does not create a `mentioned` occurrence. Each
+occurrence remains a separately cited fact in the current transcript.
+Normalization validates the exact pair, orders and
+deduplicates exact source references, then orders occurrences by valid source
+chronology, NPC comparison identity, display name, kind, and reference sequence.
+Only entries with the same NPC ID, canonical name, kind, and complete valid evidence
+sequence are collapsed; distinct categories or evidence remain separate.
+
+See the [combat-turn artifact](dnd-combat-turn-artifacts.md) for combat-action
+occurrences. The [enemy-event artifact](dnd-enemy-event-artifacts.md) consumes
+only `combat_opponent` occurrences as grounding; they never establish an enemy
+event or outcome. The [JSON output contract](json-output.md) defines
+publication. Pipeline mechanics are described in
+[D&D module internals](../internal/dnd.md).

docs/integrations/dnd-npc-registry-artifacts.md

@@ -0,0 +1,92 @@
+# D&D NPC Registry Artifact
+
+This contract defines the durable NPC registry produced by `dnd/npc-registry`. It is a
+minimal, source-grounded identity registry for other D&D artifacts, not a
+character sheet or a relationship summary.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/npc-registry` |
+| Schema ID | `notarius.dnd.npc_registry` |
+| Schema name | `notarius_dnd_npc_registry_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+| Identity policy | `dnd.npc_registry.identity.v1` |
+
+`v1` accepts one strict JSON object with required `npcs`; the array may be
+empty. NPC and source-reference objects reject unknown fields. An incompatible
+artifact shape or identity-policy change uses a new version or policy.
+
+## Wire shape and identity
+
+Each NPC has these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `id` | `npc:sha256:` followed by 64 lowercase hexadecimal characters. |
+| `name` | Non-empty canonical display name. |
+| `source_refs` | One or more transcript evidence ranges for the identity. |
+
+A source reference has exactly `source_id`, `start_unit_id`, and `end_unit_id`.
+The source ID identifies the transcript, unit IDs are positive inclusive unit
+identifiers, and the start may not follow the end.
+
+```json
+{
+  "npcs": [
+    {
+      "id": "npc:sha256:35ba5f679aee69e07ae3bd65c44278f29539d5dc9bb5225db1c0060555b23221",
+      "name": "Mira Thorn",
+      "source_refs": [
+        {"source_id": "session-7", "start_unit_id": 4, "end_unit_id": 5}
+      ]
+    }
+  ]
+}
+```
+
+The ID is deterministic: normalize the name to Unicode NFKC, normalize the
+supported apostrophe forms, collapse whitespace, case-fold it, then serialize
+`["dnd.npc_registry.identity.v1", comparison_name]` as compact JSON. SHA-256
+those UTF-8 bytes and prefix the lowercase hexadecimal digest with
+`npc:sha256:`. Each canonical identity and ID appears at most once.
+Normalization collapses records with the same canonical identity, retains their
+earliest position, and merges
+their canonicalized evidence; it does not add aliases, roles, descriptions, or
+relationship fields.
+
+When evidence supports a semantically duplicate group, the canonical display
+name is one of that group's supplied candidates. A complete, stable proper name
+is preferred over an abbreviation. An unadorned proper name is preferred over
+the same name plus a contextual class, role, title, or relationship descriptor
+unless the transcript establishes that descriptor as part of the person's
+name. A longer candidate is not preferred solely because it includes such a
+descriptor.
+
+## Scope and consumers
+
+Only individually identifiable NPC names with transcript evidence belong in
+this artifact. A factual third-party mention can establish an identity even if
+the NPC is not present, speaking, or acting in the cited passage. Names used
+only in hypothetical, speculative, or imagined examples are excluded, as are
+groups, generic roles, invented labels, and descriptive enrichment. Its source
+references prove registry provenance; they do not become evidence for a spell,
+occurrence, combat, or enemy-event occurrence.
+
+Registry evidence establishes an identity, not an [NPC occurrence](dnd-npc-occurrence-artifacts.md).
+That later artifact independently records any current-transcript occurrence
+with its own cited evidence and category.
+
+This registry can ground actor or caster names in the [spell](dnd-spell-artifacts.md)
+and [combat-turn](dnd-combat-turn-artifacts.md) artifacts. It is required to
+resolve the canonical `name` in an [NPC occurrence](dnd-npc-occurrence-artifacts.md).
+Occurrence consumers receive names-only grounding; Notarius resolves the
+selected canonical name and writes the unchanged exact durable ID/name pair.
+Spells, combat turns, and the [enemy-event artifact](dnd-enemy-event-artifacts.md)
+also receive names-only grounding for actor or subject display. None of these
+projections supply later-artifact evidence. [Configuration](../config.md#d-d-reference-slots)
+owns the `npc_registry` binding rules.
+The [JSON output contract](json-output.md) defines publication, and
+[D&D module internals](../internal/dnd.md) owns pipeline mechanics.

docs/integrations/dnd-scene-description-artifacts.md

@@ -0,0 +1,70 @@
+# D&D Scene-Description Artifact
+
+This contract defines the durable output of `dnd/scene-descriptions`. Each
+record classifies one accepted transcript chunk and gives it a minimal
+source-grounded title and summary.
+
+## Identity and compatibility
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `dnd/scene-description-list` |
+| Schema ID | `notarius.dnd.scene_descriptions` |
+| Schema name | `notarius_dnd_scene_descriptions_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+
+`v1` is a strict JSON object with required non-empty `scenes`. Scene and
+source-reference objects reject unknown fields. An incompatible shape change
+requires a new schema version.
+
+## Wire shape
+
+Each scene has exactly these required fields:
+
+| Field | Contract |
+| --- | --- |
+| `id` | Non-empty accepted chunk ID, assigned by Notarius. |
+| `source_ref` | The assigned inclusive source range for that chunk. |
+| `kind` | `combat`, `narrative`, `recap`, or `meta`. |
+| `title` | Non-empty, trimmed, source-grounded title. |
+| `summary` | Non-empty, trimmed, source-grounded summary. |
+
+`source_ref` has exactly `source_id`, `start_unit_id`, and `end_unit_id`.
+Its source ID identifies the input transcript; its positive unit IDs identify
+the chunk's inclusive range, with the start no later than the end.
+
+```json
+{
+  "scenes": [
+    {
+      "id": "chunk-000001",
+      "source_ref": {"source_id": "session-7", "start_unit_id": 1, "end_unit_id": 3},
+      "kind": "narrative",
+      "title": "Arrival at the watchtower",
+      "summary": "The party reaches the ruined watchtower and begins to investigate it."
+    }
+  ]
+}
+```
+
+## Meaning and normalized form
+
+`combat` identifies a chunk where active combat is the central activity.
+`narrative` is current in-world play that is not principally combat, recap, or
+meta discussion. `recap` is primarily a recounting of an earlier session, and
+`meta` is primarily out-of-character discussion. The artifact does not add
+participants, confidence, events, or information absent from the chunk.
+
+Normalization trims title and summary, orders scenes by source position and
+then ID, and removes exact duplicate records. A reused ID with different
+durable fields, or the same source range with different kind, title, or
+summary, is invalid. It does not merge adjacent ranges, alter prose, or infer
+missing scenes.
+
+The [combat-turn artifact](dnd-combat-turn-artifacts.md) and
+[enemy-event artifact](dnd-enemy-event-artifacts.md) use an exact matching
+`combat` scene only as eligibility control; scene title, summary, and source
+reference never become their evidence. Publication is defined by the
+[JSON output contract](json-output.md); implementation details live in
+[D&D module internals](../internal/dnd.md).

docs/integrations/dnd-spell-artifacts.md

@@ -1,204 +1,73 @@
 # D&D Spell Artifact
 
-This document is the durable serialized artifact contract for the production
+This contract defines the durable output of the `dnd/spells` extractor and
-D&D spell extractor. Selectable extractor keys are cataloged in
+normalizer. It records source-grounded spell-casting occurrences; it is not a
-[Configuration](../config.md#implemented-production-modules).
+spellbook, a rules lookup result, or a record of hypothetical casts.
 
-## Identity
+## Identity and compatibility
 
-- Artifact kind: `dnd/spell-list`
+| Property | Value |
-- Prompt ID: `dnd.spells`
-- Response schema key: `dnd_spells`
-- Response schema ID: `notarius.dnd.spells`
-- Response schema name: `notarius_dnd_spells_v1`
-- Response schema version: `v1`
-- Media type: `application/json`
-
-The durable JSON Schema is owned by the D&D spell artifact codec. The
-extractor's private LLM response schema is a separate transport contract: its
-source-reference objects omit `source_id`, which the extractor assigns while
-mapping the response to the canonical artifact. The LLM DTO and transport
-schema are not part of this durable contract. The private schema owns required
-fields, JSON types, object and array shapes, and unknown-field rejection;
-deterministic validators own the durable artifact's semantic constraints.
-
-The output contains canonical spell casts derived from transcript evidence.
-Source IDs are assigned from the input identity; source-unit ranges identify
-the evidence location.
-
-## Output Shape
-
-The extractor payload is a JSON object with one required top-level array. Its
-structure is:
-
-```text
-{"spell_casts": [<spell-cast object>, ...]}
-```
-
-`spell_casts` must be present. It may be empty when no spell casts are found.
-When multiple chunk results are combined, spell casts remain in chunk order.
-When the payload is written as durable output, its logical path is derived from
-the configured artifact lane ID as defined by the
-[JSON output contract](json-output.md#output-payload-files).
-
-## Spell-Cast Fields
-
-Each spell cast contains exactly these required fields:
-
-- `caster`: in-world character or creature casting the spell;
-- `spell`: spell name;
-- `source_refs`: transcript source references with extractor-assigned source
-  IDs and evidence unit ranges. It must contain at least one entry.
-
-Both string fields must be non-empty. `caster` is the canonical in-world
-caster, not the human player, transcript speaker, or GM when the associated
-character or creature can be identified. Player and party references may
-disambiguate that identity, but do not independently establish that a cast
-occurred. The `spell` value must resolve through the effective SRD-plus-overlay
-catalog as either a canonical name or alias. Catalog validation accepts aliases
-but does not rewrite them; unknown fields are rejected.
-
-The artifact includes an actual casting event or an unambiguous declared
-casting attempt. Spell mentions, hypothetical plans, rules discussion, and
-catalog matches without a casting event are excluded. The spell catalog is a
-name-recognition policy and never evidence that a cast occurred.
-
-## Source References
-
-Each source reference contains exactly three required fields: `source_id`,
-`start_unit_id`, and `end_unit_id`. The source ID must match the input identity.
-The unit IDs must be positive integers present in the input, and the start unit
-must not appear after the end unit. Unknown fields are rejected.
-
-For each cast, the complete `source_refs` collection identifies the transcript
-evidence for the caster, spell name, and occurrence of the cast or declared
-attempt. The deterministic validators establish that ranges are structurally
-valid and that the spell name is related to cited text. Semantic evidence
-sufficiency is an extraction policy and remains subject to evaluation rather
-than deterministic proof.
-
-Reference slot keys and accepted file types are defined in
-[Configuration](../config.md#implemented-production-modules). References are
-supporting disambiguation material, not source evidence, and are not
-addressable through `source_refs`.
-
-## Optional NPC Grounding
-
-The `dnd/spells` extractor accepts an optional `npcs` reference containing one
-normalized NPC artifact as `application/json`, up to 1 MiB. An external file is
-validated during preparation; an ordered generated binding is validated at the
-step handoff. Both paths use the approved NPC codec and identity policy,
-re-encode canonical durable JSON for registry provenance, and supply only the
-registry's ordered names as the operation-time spell prompt input. It helps the
-model prefer canonical caster names; it does not establish that a spell was
-cast.
-
-NPC source references may identify the run that produced the registry or any
-other session. They remain registry provenance and are never copied into a
-spell cast's `source_refs`; every spell evidence range must still identify the
-current transcript. Generated provenance records producer and canonical
-artifact identity without payload content or a path. When the slot is absent,
-the prompt receives exactly `{"npcs":[]}` with its projection digest, and the
-run has no NPC reference provenance.
-
-## Normalization Behavior
-
-When the `dnd/spells` normalizer is selected, each recognized spell name is
-rewritten to the effective catalog's canonical display name. Lookup uses the
-catalog's case-insensitive, whitespace-normalizing, apostrophe-normalizing, and
-alias rules. Unknown names are preserved exactly for the normalize validators;
-the normalizer does not guess or apply fuzzy matching.
-
-Each cast's `source_refs` is copied, sorted by exact `source_id`,
-`start_unit_id`, and `end_unit_id`, and stripped of exact structural
-duplicates. Adjacent or overlapping ranges are not merged, and the normalizer
-does not synthesize references or change their boundaries.
-
-After those per-cast changes, duplicate identity requires the same canonical
-spell name, the same caster after case folding and whitespace normalization,
-and the same complete, non-empty set of source references valid for the source
-document. Only the first occurrence is retained, in stable order. Its caster
-and canonical references are preserved. Unknown names, empty or invalid
-evidence, and casts with different evidence remain separate.
-
-Mutation and duplicate decisions are returned through the normal warnings
-surface. Warning scopes use the merged input index, such as `spell_casts[0]`,
-so they remain meaningful even when a later duplicate is removed. The
-normalizer uses these reason codes:
-
-| Reason code | Meaning |
 | --- | --- |
-| `spell_name_canonicalized` | A catalog lookup replaced an input name with its canonical display name. |
+| Artifact kind | `dnd/spell-list` |
-| `spell_name_unresolved` | A name was not found in the effective catalog and was retained unchanged. |
+| Schema ID | `notarius.dnd.spells` |
-| `source_references_normalized` | Reference order changed or exact duplicate references were removed. |
+| Schema name | `notarius_dnd_spells_v1` |
-| `duplicate_spell_cast_collapsed` | A later cast matched the retained cast's complete duplicate identity. |
+| Schema version | `v1` |
+| Media type | `application/json` |
 
-Only warnings from an accepted normalize attempt are promoted to
+`v1` is a single strict JSON object. It requires `spell_casts`; the array may
-`warnings.json`. If an unresolved name reaches the default normalize validator
+be empty. Each spell-cast object and source-reference object rejects unknown
-chain, the catalog validator rejects the candidate with `unknown_spell`; the
+fields. An incompatible shape change requires a new schema version.
-`spell_name_unresolved` warning remains in the attempt's debug artifact. An
-explicit validator override that accepts the candidate promotes the unresolved
-warning normally.
 
-The default extraction and normalization chains both preserve this registered
+## Wire shape
-order: JSON syntax, spell shape, catalog membership, source references, JSON
-Schema, then source relatedness. Extraction validates the private response
-schema; normalization validates the durable artifact schema.
 
-## Manifest Metadata
+Each `spell_casts` entry has these required fields:
 
-The extractor adds prompt and response-schema provenance under the artifact lane
+| Field | Contract |
-manifest metadata:
+| --- | --- |
+| `caster` | Non-empty in-world character or creature name. |
+| `spell` | Non-empty spell name. |
+| `source_refs` | One or more transcript evidence ranges. |
+
+Every source reference has exactly `source_id`, `start_unit_id`, and
+`end_unit_id`. The source ID identifies the input transcript; the unit IDs are
+positive inclusive unit identifiers, and the start may not follow the end in
+that source. References are evidence for the cast, not campaign-reference or
+NPC-registry provenance.
 
 ```json
 {
-  "metadata": {
+  "spell_casts": [
-    "extractor": {
+    {
-      "prompt_id": "dnd.spells",
+      "caster": "Mira Thorn",
-      "prompt_version": "v1",
+      "spell": "Fireball",
-      "prompt_sha256": "sha256:...",
+      "source_refs": [
-      "response_schema_key": "dnd_spells",
+        {"source_id": "session-7", "start_unit_id": 12, "end_unit_id": 13}
-      "response_schema_id": "notarius.dnd.spells",
+      ]
-      "response_schema_name": "notarius_dnd_spells_v1",
-      "response_schema_version": "v1",
-      "response_schema_sha256": "sha256:...",
-      "catalog_base_id": "dnd-5e-2014-srd-spells",
-      "catalog_digest": "sha256:...",
-      "catalog_overlay_ids": ["campaign.example"],
-      "npc_registry_digest": "sha256:...",
-      "npc_count": 3
-    },
-    "normalizer": {
-      "catalog_base_id": "dnd-5e-2014-srd-spells",
-      "catalog_digest": "sha256:...",
-      "catalog_overlay_ids": ["campaign.example"]
     }
-  }
+  ]
 }
 ```
 
-`catalog_digest` identifies the effective semantic catalog, while
+## Evidence and normalized form
-`catalog_overlay_ids` is sorted and empty for a base-only configuration. Raw
-prompt, schema, catalog, alias, and local overlay-file content are not
-included in manifest metadata. The `normalizer` metadata uses the same catalog
-identity fields when that module is selected. Overlay origin, media type, byte
-size, and raw digest are recorded separately in the manifest's reference
-provenance; see the [JSON output contract](json-output.md#manifestjson).
 
-The `npc_registry_digest` and `npc_count` fields in the example are present for
+An entry represents an actual cast or an unambiguous declared attempt. A spell
-an external NPC registry when the extractor publishes its prepared module
+mention, rules discussion, plan, or catalog match alone is not an occurrence.
-metadata. They contain no NPC names, source references, paths, or raw
+The configured catalog checks the name; it does not establish evidence.
-bytes. A generated registry's identity is instead represented by the framework
-handoff provenance and dependency fingerprint, so the consumer module metadata
-does not duplicate it.
 
-The extractor's prompt hash, private response-schema hash, and effective catalog
+When normalization is selected, recognized spell names use the effective
-digest also contribute independently scoped semantic checkpoint fingerprints.
+catalog's canonical display name. Source references are put in canonical source
-Changing any of those prepared contracts intentionally produces a cold
+order and exact duplicate references are removed. A later entry is collapsed
-checkpoint miss. Fingerprints contain only digests, never prompt, schema,
+only when it has the same canonical spell, the same case- and
-catalog, or reference content. When an NPC registry is bound, its semantic
+whitespace-insensitive caster identity, and the same complete valid reference
-digest contributes an additional local `npc_registry` fingerprint for an
+sequence. Remaining entries retain their merged order.
-external binding; the manifest metadata contains only that digest and
+
-`npc_count`. Raw NPC file provenance remains independently recorded in the
+The optional normalized [NPC registry artifact](dnd-npc-registry-artifacts.md) can ground a
-manifest's `references` list. Generated bindings contribute the canonical
+caster name. Its own references remain registry provenance and are never copied
-artifact dependency fingerprint and bounded producer provenance instead.
+into `source_refs`.
+
+## Related contracts
+
+The [spell-catalog overlay contract](dnd-spell-catalog-overlays.md) defines
+the configured catalog additions. The [JSON output contract](json-output.md)
+defines where this logical artifact is published; [D&D module internals](../internal/dnd.md)
+describes extraction and validation mechanics.

docs/integrations/dnd-spell-catalog-overlays.md

@@ -1,19 +1,27 @@
-# D&D Spell-Catalog Overlay Contract
+# D&D Spell-Catalog Overlays
 
-This document defines the JSON format accepted by the D&D spell catalog
+This document defines the optional JSON overlay consumed by the D&D spell
-resolver. An overlay supplies campaign-specific spell names and aliases for
+extractor. An overlay contributes campaign spell names and aliases for
-recognition. It does not supply spell rules, levels, classes, effects, or
+recognition. It does not define spell rules, effects, levels, classes, or
-source evidence.
+transcript evidence. Bind the optional `spell_catalog` reference as described
+in [Configuration](../config.md#references-and-ordered-handoffs).
 
-The `dnd/spells` extractor accepts one optional UTF-8 `application/json` overlay
+## Contract Identity
-bundle through its `spell_catalog` reference slot. The framework materializes
-that file relative to the configuration or command-line binding, enforces the
-1 MiB slot limit, and records its origin and raw digest separately from the
-effective catalog digest.
 
-## Shape
+| Property | Value |
+| --- | --- |
+| Consumer | D&D spell extraction and normalization |
+| Reference slot | `spell_catalog` |
+| Media type | `application/json` |
+| Required schema version | `notarius.dnd.spell-catalog-overlay.v1` |
+| Base catalog | Embedded D&D 5e 2014 SRD catalog |
 
-An overlay bundle has this shape:
+At most one overlay document may be bound. The maintained example is
+[dnd-spell-catalog.json](../../examples/dnd-spell-catalog.json).
+
+## Wire Shape
+
+This is a minimal valid overlay:
 
 ```json
 {
@@ -22,49 +30,49 @@ An overlay bundle has this shape:
     {
       "id": "campaign.example",
       "ruleset": "dnd-5e-2014",
-      "source": {
+      "source": {"title": "Example campaign spells"},
-        "title": "Example campaign spells",
+      "spells": [{"name": "Aegis of Emberfall"}]
-        "version": "1",
-        "url": "",
-        "license": ""
-      },
-      "spells": [
-        {
-          "name": "Aegis of Emberfall",
-          "aliases": ["Emberfall Aegis"]
-        }
-      ]
     }
   ]
 }
 ```
 
-The top-level `schema_version` and `catalogs` fields are required. The schema
+| Field | Required | Meaning and constraints |
-version must be exactly `notarius.dnd.spell-catalog-overlay.v1`, and at least
+| --- | --- | --- |
-one catalog is required. Catalogs require a unique, non-empty, trimmed `id`,
+| `schema_version` | Yes | Exactly `notarius.dnd.spell-catalog-overlay.v1`. |
-the exact `dnd-5e-2014` `ruleset`, a `source`, and a non-empty `spells` array.
+| `catalogs` | Yes | Non-empty array of catalog objects with unique IDs. |
+| `catalogs[].id` | Yes | Non-empty trimmed string. |
+| `catalogs[].ruleset` | Yes | Exactly `dnd-5e-2014`. |
+| `catalogs[].source.title` | Yes | Non-empty trimmed string. |
+| `catalogs[].source.version` | No | String when present. |
+| `catalogs[].source.url` | No | String when present. |
+| `catalogs[].source.license` | No | String when present. |
+| `catalogs[].spells` | Yes | Non-empty array of spell objects. |
+| `catalogs[].spells[].name` | Yes | Non-empty trimmed string. |
+| `catalogs[].spells[].aliases` | No | Array of non-empty trimmed strings when present. |
 
-`source.title` is required and must be non-empty and trimmed. `source.version`,
+Unknown fields are rejected at every object level. The document must contain
-`source.url`, and `source.license` are optional strings and may be empty.
+one JSON value; `null` is not accepted for optional strings or aliases.
-Each spell requires a non-empty, trimmed `name`. `aliases` may be omitted or
-may be an array of trimmed, non-empty strings; JSON `null` is not an alias
-array. Overlay objects contain no other supported spell fields.
 
-Decoding is strict: unknown fields, malformed JSON, trailing JSON values, and
+## Composition And Compatibility
-non-string optional source fields are rejected.
 
-## Composition
+Notarius starts with the embedded base catalog, then applies overlay catalogs
+in ascending catalog-ID order. A new canonical spell name adds a recognition
+entry. If an overlay names an existing canonical spell, it augments that spell
+with aliases while retaining the established display spelling.
 
-The resolver always starts with the embedded D&D 5e 2014 SRD catalog. Overlay
+Repeated aliases for the same spell are accepted. A canonical-name, canonical-
-catalogs are sorted by `id` before composition, so the input order does not
+to-alias, or alias-to-alias collision between different spells is rejected,
-affect the result. A new canonical name adds a recognition entry. A canonical
+including a collision with the embedded catalog. Matching uses the catalog’s
-name matching an existing canonical name augments that spell and keeps the
+case, whitespace, and apostrophe normalization, so authors should avoid names
-established canonical display spelling. Repeated aliases for the same spell
+or aliases that normalize to another spell.
-are idempotent.
 
-Canonical-name display conflicts and canonical/alias or alias/alias collisions
+Spell extraction receives the effective catalog as deterministic canonical-name
-between different spells are errors, including collisions with the embedded
+and alias pairs. An alias in the transcript selects its associated canonical
-catalog. Canonical names and aliases use the catalog's case, whitespace, and
+name; the extractor is instructed to return that canonical spelling. The
-common-apostrophe normalization rules. The effective catalog returns canonical
+projection contains no catalog source metadata or provenance, and aliases
-names in sorted order and produces a semantic SHA-256 digest that is stable
+remain recognition context rather than transcript evidence.
-under JSON formatting, object-key, catalog, spell, and alias reordering.
+
+The overlay is a recognition aid only. The durable spell-artifact schema and
+source-evidence rules are defined by the
+[D&D spell artifact contract](dnd-spell-artifacts.md).

docs/integrations/evidence-context.md

@@ -0,0 +1,107 @@
+# Published Evidence Context
+
+This contract defines the optional `source/evidence-context` artifact emitted
+by the production JSON output. It is a selected source-unit excerpt for
+convenient reading alongside normalized lane artifacts; it is not a second
+citation or provenance model. Its configuration is owned by
+[Configuration](../config.md#module-bindings-and-validators), and its
+logical-file discovery is owned by [Published JSON Output](json-output.md).
+
+## Identity And Discovery
+
+When enabled, the JSON bundle contains `evidence-context.json` and an
+`index.json` `evidence_context` descriptor with the same six fields as other
+pipeline-wide artifact descriptors.
+
+| Property | Value |
+| --- | --- |
+| Artifact kind | `source/evidence-context` |
+| Media type | `application/json` |
+| Schema ID | `notarius.source.evidence_context` |
+| Schema name | `notarius_source_evidence_context_v1` |
+| Schema version | `v1` |
+| Logical file | `evidence-context.json` |
+
+Consumers must discover the file from the descriptor, verify all six descriptor
+fields, and decode only a supported schema version. The descriptor is optional:
+its absence means evidence publication was not enabled for that bundle.
+
+## Payload
+
+The v1 payload is a top-level JSON array of generic source units. There is no
+wrapper, source-level metadata, context grouping, lane identifier, or evidence
+reference in the payload. An enabled configuration with no contributing
+accepted evidence publishes `[]`.
+
+```json
+[
+  {
+    "id": 10,
+    "kind": "transcript_segment",
+    "text": "Aria casts Cure Wounds.",
+    "ref": {
+      "source_id": "session-alpha",
+      "start_unit_id": 10,
+      "end_unit_id": 10
+    }
+  },
+  {
+    "id": 20,
+    "kind": "transcript_segment",
+    "text": "The party regroups.",
+    "ref": {
+      "source_id": "session-alpha",
+      "start_unit_id": 20,
+      "end_unit_id": 20
+    }
+  }
+]
+```
+
+Each source unit has required `id`, `kind`, `text`, and self `ref` fields.
+`ref` contains `source_id`, `start_unit_id`, and `end_unit_id`, and both unit
+endpoints identify that unit's `id`. A unit may also contain source-owned
+`metadata`, an open-ended JSON object. Fixed unit and reference fields are
+strict: consumers must reject unknown fixed fields, malformed units, invalid
+self-references, units whose `source_id` differs from other units in the same
+excerpt, and a payload that is not the array described here.
+
+The excerpt preserves each selected unit exactly as represented by the
+validated generic source document. It does not add evidence-context-specific
+annotations or reshape source-owned metadata.
+
+## Selection And Citations
+
+The framework obtains direct source references only through typed evidence
+projections of accepted normalized artifacts in the configured lane allowlist.
+It validates each reference against the current source document, expands its
+range by `window_units` source-unit positions on each side, clamps at document
+boundaries, and takes the union of all expanded ranges. The output contains
+each selected source unit once in source-document position order, regardless
+of numeric unit IDs. Repeated references, overlapping windows, and citations
+from multiple lanes do not duplicate a unit. Rejected, failed, absent,
+inactive, and unselected lanes contribute nothing.
+
+Normalized lane artifacts remain authoritative for citations and for which lane
+cited a range. The excerpt has no lane attribution and must not be used to
+reconstruct it. Its included nearby units provide reading context only; they
+do not widen any citation in a lane artifact.
+
+The excerpt contains at most every generic source unit once. It can therefore
+equal the complete generic source document when coverage is broad or the
+window is large. No byte-, token-, or compression-size guarantee is made, and
+the framework does not truncate the excerpt to meet an arbitrary size limit.
+
+## Consumer Responsibilities And Data Handling
+
+The artifact is additive to the JSON bundle and is not a lane payload,
+normalized-output count, checkpoint, or generated reference. Consumers that
+do not need it must tolerate an absent descriptor. Consumers that do use it
+should validate the descriptor and payload before use, retain the artifact with
+its schema identity when needed for a run record, and read citations from the
+corresponding normalized lane artifacts.
+
+The excerpt contains source-unit text and source-owned metadata and is durable
+output. Treat it as sensitive source content, apply appropriate access controls
+and retention, and do not assume its selected form is materially smaller or
+less sensitive than the original input.

docs/integrations/json-output.md

@@ -1,191 +1,142 @@
-# JSON Output
+# Published JSON Output
 
-This document is the durable JSON output file-format contract produced by the
+This document defines the logical JSON bundle emitted by the production JSON
-production JSON encoder and written by the CLI. Selectable output-encoder keys
+output encoder. The bundle’s physical destination, atomic publication, and
-are cataloged in
+retention are operational concerns; see [Operations](../operations.md#output-bundles).
-[Configuration](../config.md#implemented-production-modules).
+Output configuration, including chunk-map and evidence-context publication, belongs in
+[Configuration](../config.md#module-bindings-and-validators).
 
-The output module produces the logical bundle described here. The CLI's
+## Bundle Layout
-physical placement and lifecycle for that bundle are defined in
-[Operations](../operations.md#output-directory).
 
-## Files
+All paths below are logical, relative, slash-separated bundle paths. The
+encoder always emits the first four JSON files below and adds lane or
+pipeline-wide artifact files when their corresponding artifacts are available:
 
-The encoder writes:
+A subprocess caller first obtains the physical bundle root from the
+[run-result receipt](run-result.md), then resolves `index.json` beneath that
+root for the logical discovery described here.
 
-- `index.json`
+| Path | Purpose |
-- `manifest.json`
+| --- | --- |
-- `lanes/<lane-id>.json`, one file per normalized serialized artifact
+| `index.json` | Entry point that names the other published files and lane payloads. |
-- `rejected.json`
+| `manifest.json` | Run provenance and result summaries. |
-- `warnings.json`
+| `rejected.json` | Rejected pipeline outputs. |
+| `warnings.json` | Accepted-output and run warnings. |
+| `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. |
 
-Files are pretty-printed JSON with a trailing newline when the payload is JSON.
+JSON files are pretty-printed with a trailing newline. Lane payloads are
-Logical file paths are relative, slash-separated, and may not contain `..`.
+accepted only when their media type is `application/json`.
 
 ## `index.json`
 
-Shape:
+`index.json` is the bundle’s discovery document. An approved run with no
+normalized lanes has this valid minimal index:
 
 ```json
 {
   "manifest_file": "manifest.json",
-  "output_files": [
+  "output_files": [],
-    {
-      "lane_id": "spells",
-      "media_type": "application/json",
-      "file": "lanes/spells.json",
-      "module_key": "noop",
-      "schema_id": "notarius.dnd.spells",
-      "schema_name": "notarius_dnd_spells_v1",
-      "schema_version": "v1"
-    }
-  ],
   "rejected_file": "rejected.json",
   "warnings_file": "warnings.json"
 }
 ```
 
-`output_files` is sorted by lane ID. Output file names are produced by
+| Field | Required | Meaning |
-sanitizing the lane ID:
+| --- | --- | --- |
+| `manifest_file` | Yes | Always `manifest.json`. |
+| `output_files` | Yes | Lane descriptors sorted by `lane_id`. |
+| `rejected_file` | Yes | Always `rejected.json`. |
+| `warnings_file` | Yes | Always `warnings.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. |
 
-- characters outside `A-Z`, `a-z`, `0-9`, `.`, `_`, and `-` become `_`;
+Each lane descriptor has required `lane_id` and `file`. It may also include
-- repeated `..` sequences are replaced;
+`media_type`, `module_key`, `schema_id`, `schema_name`, and `schema_version`
-- leading and trailing `.`, `_`, and `-` are trimmed;
+when supplied by the normalized artifact. Each pipeline-wide artifact
-- empty sanitized names are rejected;
+descriptor (`chunk_map` or `evidence_context`) contains `artifact_kind`,
-- two lanes that sanitize to the same output file are rejected.
+`file`, `media_type`, `schema_id`, `schema_name`, and `schema_version`. Their
+payloads are defined by the [Accepted Chunk Map contract](chunk-map.md) and
+[Published Evidence Context](evidence-context.md), respectively.
 
-`manifest_file`, `rejected_file`, and `warnings_file` contain the fixed paths
+The lane path is derived from its lane ID. Characters outside letters, digits,
-shown above. Each `output_files` entry requires `lane_id` and `file`. It also
+periods, underscores, and hyphens become underscores; `..` sequences are
-contains the normalized payload `media_type`, normalizer `module_key`, and
+neutralized; leading and trailing periods and underscores are removed. A lane
-response `schema_id`, `schema_name`, and `schema_version` when those values are
+that produces an empty name, or two lanes that produce the same path, makes
-available.
+output encoding fail.
+
+## Lane Payloads
+
+Each `lanes/<safe-lane-id>.json` file is the codec-owned normalized JSON for
+that lane. Consumers should use the index descriptor’s schema identity rather
+than infer a lane schema from its name. The current D&D payload contracts are
+[spells](dnd-spell-artifacts.md), [NPC registry](dnd-npc-registry-artifacts.md),
+[NPC occurrences](dnd-npc-occurrence-artifacts.md),
+[combat turns](dnd-combat-turn-artifacts.md),
+[item registry](dnd-item-registry-artifacts.md),
+[item occurrences](dnd-item-occurrence-artifacts.md),
+[scene descriptions](dnd-scene-description-artifacts.md),
+[enemy events](dnd-enemy-event-artifacts.md),
+[location registry](dnd-location-registry-artifacts.md), and
+[location occurrences](dnd-location-occurrence-artifacts.md).
 
 ## `manifest.json`
 
-`manifest.json` contains a run manifest. This abridged example shows its core
+`manifest.json` is published provenance, not a copy of lane payloads or a
-structure:
+checkpoint store. Fields without a value may be omitted. Its top-level fields
+group into the following externally observable summaries:
 
-```json
+| Group | Fields |
-{
+| --- | --- |
-  "run_id": "run-123",
+| Run identity and result | `run_id`, `pipeline_id`, `pipeline_digest`, `schema_version`, `validation_status`, `started_at`, `completed_at` |
-  "pipeline_id": "dnd-session",
+| Resolved components | `input_module`, `chunker`, `extractors`, `merger`, `normalizer`, `output_encoder`, `artifact_lanes`, `validator_chains`, `module_metadata` |
-  "artifact_lanes": [
+| Source and references | `source_digests`, `references` |
-    {
+| Published result summaries | `normalized_outputs`, `rejected_outputs` |
-      "id": "spells",
+| Execution summaries | `chunk_plan`, `checkpoint_decisions`, `llm_profiles`, `metadata` |
-      "extractor": "dnd/spells",
-      "merger": "appendorder",
-      "normalizer": "noop"
-    }
-  ],
-  "validation_status": "approved",
-  "started_at": "2026-01-01T00:00:00Z",
-  "completed_at": "2026-01-01T00:00:01Z"
-}
-```
 
-Fields with empty values may be omitted by JSON encoding.
+`references` records provenance such as the target, slot, origin, digest,
+media type, size, and generated-artifact identity. It does not contain
+reference content. `normalized_outputs` and `rejected_outputs` likewise
+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.
 
-The manifest fields are:
+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
+override behavior are defined by the [CLI reference](../cli.md#run).
 
-- `run_id`, `pipeline_id`, and `pipeline_digest`: run and resolved-pipeline
+Each `llm_profiles` entry identifies effective, non-secret LLM execution
-  identity;
+provenance:
-- `input_module`, `chunker`, `extractors`, `merger`, `normalizer`, and
-  `output_encoder`: resolved module keys;
-- `chunk_plan`: payload-free provenance for the effective chunk plan. `mode`
-  is the effective cache mode; `action` is `reused`, `generated`,
-  `refreshed`, or `bypassed` when a plan was materialized. `requested_module`
-  is the current pipeline chunker, while `producer_input_module`,
-  `producer_module`, `producer_llm_profile`, `producer_references`,
-  `producer_metadata`, `source_digest`, `plan_digest`, `plan_schema_version`,
-  and `created_at` describe the stored or generated producer when available.
-  A cached plan can therefore identify a producer different from the requested
-  module. This object never embeds ranges, units, annotations, prompts,
-  responses, or reference content;
-- `module_metadata` and `artifact_lanes`: module and per-lane provenance,
-  including prompt and response-schema provenance when provided;
-- `validator_chains`: resolved validation points and validators;
-- `source_digests` and `references`: source and reference provenance;
-- `normalized_outputs` and `rejected_outputs`: payload-free result summaries;
-- `llm_profiles`: selected profile IDs and provider or model names when
-  available;
-- `metadata`: the effective prompt `session_id`;
-- `validation_status`: `approved` or `rejected`;
-- `started_at` and `completed_at`: UTC run timestamps.
 
-`source_digests` contains source document digests only. Bound references are
+| Field | Required | Meaning |
-recorded separately under `references`, which contains provenance only: target
+| --- | --- | --- |
-stage, lane ID when present, slot name, origin type and URI, digest, media
+| `id` | Yes | Selected PromptKit profile identifier. |
-type, byte size, and binding source. Reference content is not written to
+| `provider` | No | Notarius adapter provider identifier. |
-durable output.
+| `model` | No | Effective provider model identifier. |
+| `backend_id` | No | Effective PromptKit backend registration identifier. Endpoint-only profiles omit it. |
+| `reasoning_effort` | No | Effective opaque provider reasoning setting. An empty or explicitly cleared setting is omitted. |
 
-Reference `stage` is `chunk`, `extract`, `merge`, or `normalize`. `lane_id` is
+These values describe observed execution; they are not a backend-registration
-omitted for chunk references and present for extract, merge, and normalize
+interface. Entries that differ by backend or effective reasoning remain
-references.
+distinct even when their profile, provider, and model are otherwise equal.
 
-`validation_status` is `approved` when no outputs were rejected and `rejected`
+## Rejections And Warnings
-when one or more outputs were rejected.
 
-Producer warnings and the current run's chunk-validation warnings remain in
+`rejected.json` is always an object with a `rejected` array. Each entry has
-`warnings.json`. The manifest records only provenance and decision summaries;
+required `stage` and `message`; `step_id`, `lane_id`, `module_key`, `chunk_id`,
-empty producer-only values are omitted for compatibility with existing readers.
+`chunk_index`, `validator_name`, `reason_code`, `attempt_count`, and
+`diagnostic_artifact_path` are present only when applicable.
 
-`validator_chains` records the resolved validator chain for each validation
+`warnings.json` is always an object with a `warnings` array. Each warning has
-point. Entries include stage, lane ID when applicable, module key, and validators
+`reason_code` and `message`; `scope` is optional. Both arrays are empty when
-with key and execution class. Empty chains are recorded with an empty
+there is nothing to report.
-`validators` array, including chains resolved from explicit empty config
-overrides.
 
-`normalized_outputs` summarizes each normalized lane output without embedding
+## Compatibility
-payload bytes. Entries include lane ID, normalizer module key, source ID, media
-type, and response schema provenance where available.
 
-`rejected_outputs` summarizes rejected module outputs without embedding raw
+The index is the authoritative map from a logical lane to its published
-payload bytes. Entries include stage, lane, module, chunk, validator or reason,
+payload. Consumers must tolerate omitted optional manifest and descriptor
-message, attempt count, and optional diagnostic artifact path.
+fields, and should rely on the linked artifact contract for each lane’s JSON
-
+shape. This contract describes the published logical bundle only; it does not
-## Output Payload Files
+promise a filesystem layout or expose internal state formats.
-
-Each normalized serialized artifact is written to
-`lanes/<sanitized-lane-id>.json`. The JSON output encoder is domain-neutral and
-accepts only artifacts whose codec media type is `application/json`. The file
-contains the codec-owned JSON bytes pretty-printed.
-
-The schema of each lane payload is owned by that artifact contract. For the
-current D&D lanes, see [D&D Spell Artifact](dnd-spell-artifacts.md),
-[D&D NPC Artifact](dnd-npc-artifacts.md), and
-[D&D Combat-Turn Artifact](dnd-combat-turn-artifacts.md).
-
-## `rejected.json`
-
-Shape:
-
-```json
-{
-  "rejected": []
-}
-```
-
-When output validation rejects an output, each entry contains `stage` and
-`message`. It includes `lane_id`, `module_key`, `chunk_id`, `chunk_index`,
-`validator_name`, `reason_code`, `attempt_count`, and
-`diagnostic_artifact_path` when applicable.
-
-## `warnings.json`
-
-Shape:
-
-```json
-{
-  "warnings": [
-    {
-      "scope": "extract",
-      "reason_code": "example",
-      "message": "human-readable warning"
-    }
-  ]
-}
-```
-
-`warnings` is an empty array when no warnings are reported.
-Each warning requires `reason_code` and `message`; `scope` is omitted when it is
-empty.

docs/integrations/pkg-promptkit.md

@@ -0,0 +1,108 @@
+# PromptKit Integration
+
+Notarius pins
+[`gitea.maximumdirect.net/eric/promptkit` v0.5.0](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0)
+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)
+owns the public engine API, and the upstream
+[format reference](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/formats.md)
+owns prompt, profile, and schema file contracts.
+
+## Supported Boundary
+
+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
+  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;
+- register the optional conventional `local` backend through `BackendLocal`,
+  `LocalBackend`, and `WithBackend`;
+- distinguish structured-output validation failure from execution failure; and
+- identify a missing explicit profile through `ErrProfileNotFound` and backend
+  admission exhaustion through `ErrCapacityExceeded`.
+
+The pinned
+[`BackendLocal`, `LocalBackend`, and `WithBackend` API](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/backends.go)
+owns the registration and backend-capacity contract.
+
+For one completion, the adapter calls `PrepareExecution`, takes a
+caller-owned `Details` snapshot, and calls `RunPrepared` for that same opaque
+prepared execution. It defers `Discard` for every unexecuted handle. Explicit
+profile preflight uses `Engine.InspectProfile`; it does not prepare a synthetic
+prompt. PromptKit's prepared handle, inspection result, and capacity-error
+types stay inside the Notarius LLM adapter.
+
+When a PromptKit profile and runtime override leave `temperature`, `max_tokens`,
+or `top_p` unset, Notarius leaves that control unset as well. Compatible
+providers therefore apply their own defaults; an operator that requires a
+specific sampling value must select it explicitly in the profile or runtime
+override.
+
+Notarius does not use PromptKit's optional `ArtifactReader`. It materializes
+source and reference content itself and supplies owned inline artifacts at the
+adapter boundary. It also retains responsibility for pipeline retries,
+scheduling, debug persistence, redaction, profile provenance, and conversion
+from private model responses into durable domain artifacts.
+
+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
+correlation identifiers and may be exposed to providers and provider
+observability. The CLI contract owns generation and override behavior.
+
+Notarius records PromptKit's selected backend ID and effective reasoning
+setting as optional run-manifest provenance. Endpoint-only profiles have no
+backend ID. Debug prompt material also retains the selected backend ID and
+PromptKit's stable lower-case `effective_model_params` JSON, which may include
+`backend_id`. Notarius production configuration exposes one optional
+conventional `local` registration. It does not expose a general user-defined
+PromptKit backend registry. Endpoint-only profiles remain supported unchanged.
+
+Notarius retains its application-wide scheduled client around the PromptKit
+adapter. PromptKit may apply a narrower limit for the selected backend;
+endpoint-only profiles have no such backend limit. The adapter translates
+PromptKit capacity rejection into the provider-neutral Notarius
+`ErrLLMCapacityExceeded` contract. It may include the normalized selected
+backend ID in safe diagnostic context, without exposing PromptKit's capacity
+error type, and leaves retries to the calling pipeline stage.
+
+## Profile Sources And Compatibility
+
+Notarius gives PromptKit the configured operator profile source, registered
+application fallback profile assets, and optional backend registration through
+the same construction path for inspection and execution. PromptKit owns the
+resulting source precedence and strict profile parsing: a matching operator
+profile is a complete replacement for a fallback or built-in profile, while an
+invalid matching document fails instead of falling through. The operator
+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,
+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,
+profile-source construction, and this compatibility statement against the
+pinned upstream documentation.
+
+## Notarius Ownership
+
+[LLM Runtime Internals](../internal/llm.md) describes how Notarius mounts
+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.
+[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
+configuration selects one PromptKit profile source and optionally registers
+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.

docs/integrations/run-result.md

@@ -0,0 +1,68 @@
+# Run Result Receipt
+
+`notarius run --json` writes this receipt to standard output when a run
+completes successfully. It lets a subprocess caller discover the physical root
+of the published output bundle without parsing interactive command output.
+Command syntax, streams, and exit statuses are defined in the
+[CLI reference](../cli.md); logical files within the bundle are defined in the
+[Published JSON Output contract](json-output.md).
+
+## Schema
+
+The current schema version is `notarius.run-result.v1`.
+
+| Field | Required | Meaning |
+| --- | --- | --- |
+| `schema_version` | Yes | Exactly `notarius.run-result.v1`. |
+| `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. |
+| `validation_status` | Yes | The final run manifest validation status. |
+| `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.
+
+```json
+{
+  "schema_version": "notarius.run-result.v1",
+  "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"
+}
+```
+
+## Paths And Bundle Discovery
+
+`output_directory` and `debug_directory`, when present, are lexical absolute
+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.
+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
+lane payloads.
+
+## Delivery And Compatibility
+
+Notarius writes the receipt only after the output bundle has been published and
+any requested debug terminal reporting has completed. Standard output is not
+transactional: a result-write failure returns a nonzero status and can leave
+partial bytes. Consumers must ignore standard output unless the process exits
+with status 0.
+
+Future versions may add optional fields to this schema. Consumers must tolerate
+unknown fields. An incompatible field or semantic change requires a new
+`schema_version` value.

docs/integrations/seriatim.md

@@ -1,69 +1,73 @@
-# Seriatim Transcript JSON
+# Seriatim Transcript Input
 
-This document is the external input contract consumed by the production
+This document defines the JSON transcript accepted by the production Seriatim
-Seriatim input adapter. Selectable input-adapter keys are cataloged in
+input adapter. It is a source input, not a durable lane artifact. Configure the
-[Configuration](../config.md#implemented-production-modules).
+input adapter through [Configuration](../config.md#production-module-keys).
 
-## Adapter
+## Contract Identity
 
-- Source format: `application/vnd.seriatim+json`
+| Property | Value |
+| --- | --- |
+| Consumer | Seriatim input adapter |
+| Media type | `application/vnd.seriatim+json` |
+| Source document kind | `transcript` |
+| Source-unit kind | `transcript_segment` |
 
 ## Accepted Shape
 
-The input must be one JSON object with top-level `metadata` and `segments`
+The input is one JSON object containing `metadata` and a non-empty `segments`
-fields. This covers the maintained minimal fixture and Seriatim intermediate
+array. This minimal document is valid:
-output that provides the same required segment fields.
 
-The maintained example is
+```json
-[examples/seriatim-minimal-transcript.json](../../examples/seriatim-minimal-transcript.json).
+{
+  "metadata": {"id": "session-alpha"},
+  "segments": [
+    {
+      "id": 1,
+      "start": 0,
+      "end": 4,
+      "speaker": "Aria",
+      "text": "Aria casts Cure Wounds."
+    }
+  ]
+}
+```
 
-Required top-level fields:
+The maintained two-segment input is
+[seriatim-minimal-transcript.json](../../examples/seriatim-minimal-transcript.json).
 
-- `metadata`: an object. Its entries are accepted as source metadata.
+| Field | Required | Meaning and constraints |
-- `segments`: a non-empty array of segment objects.
+| --- | --- | --- |
+| `metadata` | Yes | JSON object. Its entries become source metadata; no particular metadata key is otherwise required. |
+| `segments` | Yes | Non-empty array of segment objects, kept in input order. |
+| `segments[].id` | Yes | Positive canonical decimal integer, supplied as a JSON number or string. IDs must be unique. |
+| `segments[].start` | Yes | Finite, non-negative numeric value, supplied as a JSON number or string. |
+| `segments[].end` | Yes | Finite, non-negative numeric value that is not earlier than `start`. |
+| `segments[].speaker` | Yes | String that is non-empty after trimming. |
+| `segments[].text` | Yes | String that is non-empty after trimming. Its original text is retained. |
 
-Required segment fields:
+Additional top-level and segment fields are ignored. A missing required field,
+`null` in place of an object or array, malformed JSON, or more than one
+top-level JSON value is rejected.
 
-- `id`: a positive integer JSON number or canonical decimal string without
+## Source Identity And References
-  leading zeros or surrounding whitespace;
-- `start`: a finite, non-negative JSON number or numeric string;
-- `end`: a finite, non-negative JSON number or numeric string that is not less
-  than `start`;
-- `speaker`: a non-empty string;
-- `text`: a non-empty string.
 
-Other top-level and segment fields, such as `categories`, are ignored.
+The adapter chooses the source ID in this order:
 
-Multiple top-level JSON values are rejected.
+1. a non-empty source ID supplied by the calling request;
+2. non-empty string `metadata.id`;
+3. non-empty string `metadata.source_id`;
+4. `seriatim:` followed by the first 16 hexadecimal characters of the raw
+   input’s SHA-256 digest.
 
-## Validation
+Each accepted segment becomes one source unit whose unit ID is `segments[].id`.
+Its self-reference uses the derived source ID and the same segment ID for both
+range endpoints. Artifact contracts use those segment IDs when they cite
+transcript evidence.
 
-The adapter rejects empty input, malformed JSON, multiple top-level JSON values,
+## Compatibility
-non-object segment values, duplicate segment IDs, and any violation of the
-shape or field constraints above.
 
-Segment text is preserved as provided, but it must not be empty after trimming.
+This adapter accepts only the shape described here. A broader Seriatim export
-
+is usable only when it supplies this object, metadata, and segment shape with
-## Derived Identity
+the stated types and constraints. Unknown additional fields do not add
-
+Notarius behavior.
-Notarius identifies the parsed source in this order:
-
-1. `metadata.id`, when it is a non-empty string after trimming;
-2. `metadata.source_id`, when it is a non-empty string after trimming;
-3. `seriatim:<first-16-hex-chars-of-raw-sha256>`.
-
-The exact raw input SHA-256 remains the basis of the fallback source ID. The
-source digest recorded in output provenance is instead the SHA-256 of the
-canonical generic source document, excluding the digest field itself. It covers
-the derived source identity, document kind and format, ordered units and their
-self-references, and accepted metadata. Segment IDs become the unit IDs used by
-artifact source references; each produced unit carries a self-reference whose
-source ID is the derived document ID and whose start and end IDs both equal the
-segment ID.
-
-## Compatibility Limit
-
-This contract covers only Seriatim transcript JSON with the top-level
-`metadata` object and `segments` array described here. Broader Seriatim output
-schemas are compatible only when they provide these required fields with the
-accepted types.

docs/internal/cli.md

@@ -0,0 +1,172 @@
+# CLI Internals
+
+This document describes **internal/cli**, Notarius's production composition
+root. The [CLI reference](../cli.md) owns command syntax and exit statuses;
+[Configuration](../config.md) owns configuration values; and
+[Operations](../operations.md) owns filesystem layout, recovery, and operator
+procedures.
+
+## Inputs, Outputs, And Boundaries
+
+The CLI accepts process arguments, standard streams, and injectable options
+used by tests and embedding code. It writes command results to the supplied
+streams and returns a process exit status. For a run, it also creates the
+production catalog and runtime collaborators, hands a prepared pipeline and
+source bytes to the framework, and places the logical files returned by the
+runner.
+
+It is the only boundary allowed to compose concrete registries, LLM clients,
+cache/checkpoint collaborators, debug recorders, and physical output paths.
+Pipeline modules receive interfaces and request data rather than CLI streams or
+filesystem roots. The [Architecture](../policy/architecture.md) defines this
+composition-root boundary; [Pipeline Internals](pipeline.md) owns resolution,
+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.
+
+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
+[Configuration](../config.md), while the loading and resolution mechanics are
+in [Configuration Internals](configuration.md).
+
+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
+profiles are inspected
+against the configured PromptKit source and backend registrations without
+loading a prompt or performing generation, so an unknown or invalid profile
+fails before pipeline preparation. Credential availability remains an
+execution-time concern. Pipeline listing validates configuration before
+returning normalized, sorted identifiers.
+
+## Production Composition
+
+The production composition helper allocates every framework registry and the
+prompt-asset registry, then registers the generic, Seriatim, and D&D module
+families in that order. The resulting registries provide both the module
+catalog used for resolution and the concrete constructors used for preparation.
+Tests may provide a catalog or registries instead; production code must not
+silently merge an injected partial catalog with production registrations.
+
+The production LLM factory builds one PromptKit-backed client from the resolved
+**promptkit.profile_dir** or **promptkit.profile_file** source, attaches the
+profile-provenance recorder, creates one scheduler from the effective global
+LLM limit, and wraps the client before it reaches modules. Registration and LLM
+construction errors are returned before a pipeline is prepared. Configuration
+field definitions remain in [Configuration](../config.md#promptkit-profiles);
+the D&D registrar's fallback profile assets and the adapter mechanics remain in
+[LLM Runtime](llm.md).
+
+The factory also accepts `LLMRuntimeOverrides`, whose reasoning pointer
+preserves inherit, replace, and clear states across the composition boundary.
+Run orchestration constructs this value from the mutually exclusive
+`--reasoning-effort` and `--clear-reasoning-effort` controls. Absence preserves
+a nil pointer, replacement is trimmed, and clear uses a non-nil empty string.
+The same override reaches the one shared production client, checkpoint
+identity, and debug invocation metadata. Persistent reasoning configuration
+remains owned by PromptKit profiles; Notarius configuration has no reasoning
+field.
+
+## Run Orchestration
+
+After parsing and validating a run invocation, the CLI performs this ordered
+handoff:
+
+1. load and validate configuration, then apply command-level operational
+   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
+   the effective pipeline, and inspect its explicit effective PromptKit
+   profiles;
+4. materialize external or generated references and record redacted invocation
+   and resolution provenance when debug capture is enabled;
+5. construct registries, the scheduled LLM client, and prepared modules;
+6. read the source input once, resolve its effective session from the explicit
+   override or resolved input module and raw bytes, then construct requested
+   checkpoint collaborators and invoke the framework runner with that same
+   value; and
+7. write the runner's logical output files only after a successful run, then
+   complete the command report and user-facing result.
+
+Preparation happens before source parsing, so module construction and
+dependency failures cannot begin stage execution. The CLI also preserves the
+framework's result and warning information when it writes summaries and the
+final command result. Detailed state lifecycle, resume handling, and physical
+path confinement are maintained in [Run State Internals](state.md) and
+[Operations](../operations.md).
+
+The CLI owns the versioned generated-session policy and resolves the sole
+effective value before checkpoint construction. It records that value in the
+final debug invocation summary when capture is enabled and passes it unchanged
+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.
+
+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
+output publication and requested debug terminalization succeed. A receipt-write
+failure exits with runtime status 1 and may leave partial standard-output bytes,
+but the already-published output bundle remains complete and requested debug
+reporting remains successfully terminalized. The CLI reports a bounded
+command-owned error and does not repeat terminal reporting. The receipt remains
+a CLI reporting concern rather than a framework or output-module responsibility;
+its public contract is the
+[run-result receipt](../integrations/run-result.md).
+
+## Failure Mapping And Terminal Reporting
+
+Argument, flag, and invocation-combination failures are reported to standard
+error before runtime composition and use the syntax error class. Once an
+invocation is syntactically valid, configuration loading and validation,
+resolution, registration, profile checks, reference materialization, module
+construction, input reads, runner failures, output publication, and requested
+debug handling use the runtime failure class. The public status numbers and
+stream contract are defined in the [CLI reference](../cli.md#output-streams-and-exit-statuses).
+
+When debug capture has been allocated, one command-state value records the
+known run result. Guarded terminalization writes a success report once, or
+attempts a failure report and error record once. A persistence failure is
+reported in addition to the original failure and never replaces it. If a debug
+path exists, failure output includes that path so the retained diagnostic data
+is discoverable.
+
+## Invariants To Preserve
+
+- Only the CLI composes production implementations and physical runtime roots.
+- Configuration and resolved composition failures occur before module
+  preparation or source parsing.
+- A runner's logical files are published only after a successful run.
+- Production registries and a caller-supplied catalog or registries are
+  alternative composition sources, not an implicit mixture.
+- A requested debug bundle has one terminal report attempt; its persistence
+  errors supplement rather than obscure the primary command error.
+- User-facing flags, paths, exit codes, and configuration fields are defined
+  by their public documentation, not duplicated here.
+
+## Focused Tests
+
+- **internal/cli/command_contract_test.go** covers dispatch, help, syntax and
+  runtime error classes, discovery, validation, and listing.
+- **internal/cli/run_contract_test.go** covers the run handoff, publication,
+  debug reporting, and command-owned state collaborators.
+- **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/state_hardening_test.go** covers safe run identity, state
+  roots, and failure ordering.
+
+Run **go test ./internal/cli** after changing command composition or command
+behavior. Pair it with **go test ./internal/core/config** when the configuration
+handoff changes.

docs/internal/configuration.md

@@ -0,0 +1,144 @@
+# Configuration Internals
+
+This document describes the maintainer-facing configuration boundary in
+**internal/core/config**. The [Configuration](../config.md) reference owns the
+file format, fields, defaults, precedence contract, and selectable keys. The
+[CLI reference](../cli.md) owns command syntax; this document does not redefine
+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
+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. |
+| 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. |
+
+The CLI discovers a configuration file, invokes this package, and supplies the
+result to the framework. Configuration never reads an input file, constructs a
+module, or creates output, cache, or debug paths. Those responsibilities remain
+at their respective [CLI](cli.md), [pipeline](pipeline.md), and
+[run-state](state.md) boundaries.
+
+## 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**.
+
+**Config.Validate** checks configuration-only invariants before resolution. It
+rejects incompatible profile sources, invalid state-surface values, unsupported
+concurrency settings, malformed bindings and references, invalid retries, and
+invalid pipeline, step, or lane structure. PromptKit local-backend validation
+accepts only an absolute HTTP or HTTPS endpoint with a host and no user
+information, query, or fragment, and rejects a negative local concurrency
+limit. Its errors retain the closest known pipeline, lane, and binding context.
+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.
+
+## Effective Resolution
+
+**Config.Resolve** first recomputes derived concurrency defaults and validates
+the configuration. 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.
+
+After module and validator selection, the resolver applies the effective
+profile policy to LLM-backed bindings only: command override, binding profile,
+pipeline profile, then the prompt default. Deterministic bindings remain
+profile-free, and no second inheritance decision occurs during execution. The
+public field definitions and precedence are owned by
+[Configuration](../config.md#pipelines).
+
+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
+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
+configuration through the effective value. This ownership includes the nested
+PromptKit local-backend value.
+
+Resolution failures stop before module construction and source parsing. They
+include an error path for an unconfigured pipeline, missing module, missing
+capability, incompatible artifact variant, invalid option, invalid reference,
+or invalid lane selection. CLI code maps these valid-invocation failures to the
+runtime error class described in the [CLI reference](../cli.md#output-streams-and-exit-statuses).
+
+## Resolved Identity And Redaction
+
+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
+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.
+
+## Invariants To Preserve
+
+- Defaults, YAML values, and environment values are applied in one direction;
+  later sources may override only their supported operational settings.
+- 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.
+- 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/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.
+
+Run **go test ./internal/core/config** after changing this boundary. Changes to
+the handoff or resolved-composition semantics also need the focused framework
+pipeline tests.

docs/internal/dnd.md

@@ -0,0 +1,247 @@
+# D&D Module Internals
+
+This guide records the conventions shared by the production D&D module family.
+It complements [Module Internals](modules.md), which owns generic registration
+and extension mechanics, and [Configuration](../config.md), which owns the
+selectable keys, bindings, reference syntax, and default validator chains.
+
+## Durable Artifact Contracts
+
+The ten lanes have separate durable wire contracts. This guide deliberately
+does not repeat their JSON shapes or schemas.
+
+| Lane | Durable contract |
+| --- | --- |
+| Spells | [spell artifacts](../integrations/dnd-spell-artifacts.md) |
+| NPC registry | [NPC registry artifacts](../integrations/dnd-npc-registry-artifacts.md) |
+| Combat turns | [combat-turn artifacts](../integrations/dnd-combat-turn-artifacts.md) |
+| Item occurrences | [item-occurrence artifacts](../integrations/dnd-item-occurrence-artifacts.md) |
+| Item registry | [item-registry artifacts](../integrations/dnd-item-registry-artifacts.md) |
+| NPC occurrences | [NPC-occurrence artifacts](../integrations/dnd-npc-occurrence-artifacts.md) |
+| Scene descriptions | [scene-description artifacts](../integrations/dnd-scene-description-artifacts.md) |
+| Enemy events | [enemy-event artifacts](../integrations/dnd-enemy-event-artifacts.md) |
+| Location registry | [location-registry artifacts](../integrations/dnd-location-registry-artifacts.md) |
+| Location occurrences | [location-occurrence artifacts](../integrations/dnd-location-occurrence-artifacts.md) |
+
+## Family Composition
+
+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,
+profile precedence, and validator order.
+
+Private structured-LLM response schemas are deliberately minimal. They reject
+invalid JSON structure, missing required fields, incompatible types, and
+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.
+
+## Prompt Construction
+
+D&D LLM-facing content lives beneath `assets/dnd/`. Each module contributes a
+local `prompt.yaml` declaration and `instructions.md`; input-specific files
+such as a catalog, registry, grounding projection, or candidate collection are
+local only when that module needs them. New extractor content uses its feature
+subtree, while families with both extraction and normalization content use their
+`extract` and `normalize` subtrees. Shared visual-provenance fragments use
+the `common-dnd-` prefix. Production lane code belongs with its D&D codec,
+extractor, normalizer, and validator packages; registry projections and
+identity helpers remain in their owning entity packages rather than in a
+consumer lane.
+
+The owning module’s manifest is the source of truth for which local and shared
+assets are selected, their mount paths, their message order, cache controls,
+and the files included in its prompt fingerprint. Shared fragments belong to
+the D&D shared implementation and are selected by name rather than copied into
+module directories. The root `assets` package is a content-only boundary; its
+physical ownership and rationale are defined by
+[ADR-0011](../adr/0011-centralize-llm-assets.md).
+
+Put each rule at its narrowest owner:
+
+- universal behavior belongs in the shared system asset;
+- D&D-family behavior belongs in a selected `common-dnd-` asset;
+- rules for an input projection belong with that input asset;
+- lane-specific policy belongs in the module’s `instructions.md`; and
+- transport-envelope shape belongs in the private response schema.
+
+A rule is eligible for the system prompt only when every D&D LLM prompt needs
+it regardless of lane, inputs, or response shape. Module instructions must not
+repeat rules selected from shared assets or schemas. Reintroduce such repetition
+only after observational evaluation with representative transcripts shows that
+it improves results at the intended target models and cost; structural prompt
+tests alone are not that evidence.
+
+Every maintained D&D LLM prompt selects `dnd-extraction` as its default
+profile. The D&D registrar registers the fallback, while an operator can
+replace it with a complete profile of the same ID from the configured PromptKit
+source. Deployment profile selection is documented in
+[Configuration](../config.md#promptkit-profiles).
+
+The D&D transcript assets have distinct consumers. Scene chunking consumes the
+complete-session `common-dnd-transcript-full.md`, while extraction prompts
+consume the current-chunk `common-dnd-transcript-chunk.md`. NPC, location, and
+item normalization instead mount the generic semantic-reconciliation
+candidate and transcript-window presentation assets. Player, party, glossary,
+and compatible campaign references provide disambiguating context only when
+declared by the active prompt; they never establish evidence. Reference
+material is canonically ordered before rendering so equivalent inputs remain
+stable.
+
+Extraction prompts render the common system and identity messages first, then
+cached campaign references and the cached chunk transcript. Evidence policy and
+any lane-specific registry, catalog, or grounding projection follow that
+prefix. The final module instructions message is ephemeral. This keeps the
+reusable extraction prefix identical while preserving the lane-specific suffix.
+
+Scene chunking intentionally uses a different order: system, cached campaign
+references, uncached module instructions, then the final ephemeral full
+transcript. Entity normalization also has its own order: D&D system, mandatory
+generic protocol, ephemeral domain semantic instructions, generic candidate
+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.
+
+## 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.
+
+Default chains keep responsibilities separate: structural validators assess the
+candidate, source-reference validators resolve cited ranges against the current
+source and require extraction evidence to stay within the current chunk,
+durable-schema validation checks an approved representation, and
+relatedness validators report advisory evidence concerns. The configured order
+is documented in
+[Configuration](../config.md#production-validator-keys-and-default-chains).
+
+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.
+
+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,
+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.
+
+## Semantic Registry Reconciliation
+
+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
+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.
+
+The generic core owns the mandatory handle protocol, candidate and transcript
+presentation, the private response schema, source-reference validation,
+candidate and combined-material limits, structured completion, proposal
+assessment, stable group ordering, and typed plan-application mechanics. The
+D&D prompt contributes its system message and registry-specific semantic
+instructions. The generic registrar registers the shared prompt and schema;
+the D&D registrar registers each consuming prompt and the fallback profile.
+
+Fewer than two eligible candidates skips the LLM without a semantic warning.
+An exceeded bound also skips the call and preserves the deterministic
+preprocessed registry, adding the registry's bounded fallback warning. Invalid
+structured output or discarded proposal groups use the normalizer's existing
+retry contract; retry exhaustion preserves the safe deterministic or
+partially applied result and emits its bounded fallback warning. Provider,
+transport, cancellation, and context-material failures remain execution
+errors.
+
+Application remains typed and registry-owned. All three policies select the
+canonical member's normalized display name, union member evidence in source
+order, preserve ungrouped records, and derive durable identity only after
+consolidation. NPC IDs derive from the final name. Item IDs also derive from
+the final name, and a typed guard prevents currency aliases from crossing
+denominations or mixing currency with non-currency records. Location IDs
+derive from the final name and final evidence, preserving same-name,
+parent/child, and distinct physical-place identities. Registry warning scopes,
+reason codes, and postconditions remain outside the generic core.
+
+## Generated References And Grounding
+
+Normalized D&D artifacts can be handed to a later step through a generated
+reference binding. The framework verifies artifact compatibility and retains
+producer provenance; consumers resolve the handed-off artifact into an
+immutable, validated projection for each operation. External files are checked
+during preparation, while generated artifacts are resolved at the handoff.
+
+NPC and item registry consumers receive names-only grounding. Location
+consumers receive a contextual selector containing the canonical name and the
+registry references needed to distinguish same-name places. The calling module
+resolves those supplied selections locally and maps them into the unchanged
+durable ID/name pair; an unknown or ambiguous selection rejects the complete
+occurrence result rather than accepting a partial mapping. The NPC registry
+additionally supplies names-only actor grounding to spells, combat turns, and
+enemy events.
+
+Registry references establish a registry identity and may disambiguate a
+selection, but never become occurrence evidence. Each occurrence keeps its own
+current-transcript source references, even when it was grounded through the
+same registry record.
+Scene descriptions are eligibility-only projections: they retain current-chunk
+classification data, not scene prose or evidence, and exist to route combat
+extraction. Enemy-event extraction also projects combat turns to `actor` and
+`turn_kind` and filters NPC occurrences to `combat_opponent` names and kinds.
+These projections are guidance only and never event evidence.
+
+## Lane-Specific Rules
+
+The following differences are intentional and should remain explicit when a
+shared helper changes.
+
+Shared D&D text comparison is identified by `dnd.text_comparison.v1`. Any
+semantic change requires an explicit policy-version review for every affected
+identity, mapping, normalization, and validator policy; helper source is not a
+checkpoint fingerprint.
+
+| Lane | Intentional behavior |
+| --- | --- |
+| Spells | May use a spell-catalog overlay and optional NPC grounding; the catalog validator supplies domain-specific semantic checks. |
+| NPC registry | Establishes transcript-grounded NPC identities, including factual third-party mentions, without assigning occurrence categories. It does not consume an NPC registry, and its normalizer is the LLM-assisted reconciliation exception described above. |
+| Combat turns | Requires a scene-description artifact. It calls the LLM only for an exact `combat` classification; exact non-combat classifications return an accepted empty result, while missing or mismatched classifications return an empty result with a bounded warning. Optional NPC grounding never becomes evidence. |
+| Item occurrences | Requires the normalized item registry for exact deterministic grounding at extraction and normalization. Campaign context may disambiguate, but the registry never becomes occurrence evidence. |
+| Item registry | Produces source-grounded item types and unique designations. Its LLM-assisted reconciliation is proposal-only, preserves distinct currency denominations and item types, and does not create per-instance identities. |
+| NPC occurrences | Requires the normalized NPC registry at extraction and normalization, using it for canonical actor grounding only. It separately emits cited current-transcript occurrence facts, including `mentioned`, rather than deriving them from registry provenance. |
+| Scene descriptions | Produces the classifications consumed by combat routing; it does not consume an NPC registry or provide evidence for combat artifacts. |
+| Enemy events | Requires NPC, scene-description, combat-turn, and NPC-occurrence artifacts. It calls the LLM only for an exact `combat` classification, records ordered observations rather than terminal state, and normalizes recognized names through the NPC registry while preserving grounded collective labels. |
+| Location registry | Produces a source-anchored, session-scoped registry from stable proper names or unique in-world designations. Its LLM-assisted reconciliation is proposal-only and never collapses same-name places without validated identity and evidence rules. |
+| Location occurrences | Requires the normalized location registry for both extraction and normalization. Its [durable occurrence categories](../integrations/dnd-location-occurrence-artifacts.md#occurrence-categories) distinguish explicit speculation from unsupported inference; the deterministic normalizer enforces exact registry grounding and never turns registry provenance into occurrence evidence. |
+
+The combat and scene-description contracts describe their exact handoff and
+empty-result behavior in more detail:
+[combat turns](../integrations/dnd-combat-turn-artifacts.md) and
+[scene descriptions](../integrations/dnd-scene-description-artifacts.md).
+The [enemy-event contract](../integrations/dnd-enemy-event-artifacts.md)
+defines its durable semantics; [Configuration](../config.md) owns its
+selectable bindings and validation chains.
+
+## Focused Verification
+
+When changing D&D behavior, test the affected codec, extractor, normalizer,
+validator, prompt-asset manifest, and registry projection. Also test generated
+handoffs at the integration boundary and run the full D&D module suite:
+
+~~~sh
+go test ./internal/modules/dnd/...
+go test ./internal/modules/integration/...
+~~~

docs/internal/llm.md

@@ -1,222 +1,282 @@
 # LLM Runtime Internals
 
-`internal/framework/llm` implements Notarius's transport boundary for structured
+`internal/framework/llm` is Notarius’s provider-independent structured
-completion. It contains the Scriptorium adapter, concurrency scheduler,
+completion boundary. It adapts framework requests to PromptKit, bounds
-prompt/schema registries, selected-profile recording, and provider-error
+provider calls, assembles registered prompt and schema assets, records selected
-redaction.
+profiles, and redacts provider errors. The architectural boundary is defined in
+[Architecture](../policy/architecture.md#llm-boundary); profile sources,
+credentials, and concurrency settings belong in
+[Configuration](../config.md#promptkit-profiles) and
+[Configuration](../config.md#concurrency-output-cache-and-debug).
 
-Provider-neutral ownership rules are defined in
+## Structured Completion Boundary
-[Architecture](../policy/architecture.md#llm-boundary). Profile sources,
-credentials, and concurrency settings are defined in
-[Configuration](../config.md).
 
-## Structured Contract
+Modules and LLM-backed validators depend only on
+`contracts.StructuredLLMClient`. A completion request supplies a prompt ID and
+version, optional profile and session IDs, named input material, variables, and
+a caller-owned decode target. The successful response returns the validated raw
+structured bytes together with non-secret provider, model, profile, and token
+metadata.
 
-Modules and LLM-backed validators depend on
+The caller owns the domain behavior: it chooses the prompt, prepares inputs,
-`contracts.StructuredLLMClient.CompleteStructured`. A request identifies a
+selects the private response schema, and interprets the decoded result. The
-prompt and optional profile/session, supplies named input materials and
+adapter does not own source evidence, artifact conversion, normalization, or
-variables, and provides a caller-owned decoding target. A successful response
+durable schemas. Those responsibilities remain with the module and its
-contains the validated raw structured bytes plus non-secret provider, model,
+[integration contract](../integrations/).
-profile, and token metadata.
 
-The caller owns prompt selection, response-schema selection, and interpretation
+The calling module also resolves contextual entity selections and attaches any
-of the decoded result. `LLMInputMaterial` keeps source and reference bytes with
+application identity; PromptKit and this adapter do not own entity identity.
-their origin metadata so the adapter can pass named artifacts to Scriptorium
-without exposing Scriptorium types through stage contracts.
 
-## Production Construction
+`PromptKitClient` validates the request target and prompt identity, maps each
+named material to a PromptKit inline artifact while preserving its origin URI,
+passes the supplied request session through to PromptKit's direct per-run
+session field, retains the same value as the `session_id` prompt variable for
+maintained prompt compatibility, and forwards profile selection. It does not
+derive or replace session values; the CLI owns that policy. It then creates one
+frozen prepared execution, captures its caller-owned credential-redacted
+details for debug material, and executes that exact snapshot through
+PromptKit's prepared-execution boundary. The direct field
+is authoritative for provider session behavior. A session ID is a stable,
+non-secret correlation identifier and may be exposed to providers and provider
+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.
 
-`internal/cli` constructs the production runtime by:
+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
+inherits the selected profile, a non-empty value replaces it, and an empty
+value clears inherited reasoning. The CLI's mutually exclusive
+`--reasoning-effort` and `--clear-reasoning-effort` controls select those
+states. With neither flag, profile behavior remains unchanged. Because
+production constructs one shared client, the selected state applies uniformly
+to module calls, retries, and LLM-backed validators for the whole run.
 
-1. allocating the asset registry populated by the generic, Seriatim, and D&D
+An empty request profile lets the prompt select its configured default. Before a
-   package-family registrars;
+run begins, the CLI asks the adapter to inspect every explicit profile on the
-2. creating a `ScriptoriumClient` from the effective profile source;
+resolved selected LLM-backed bindings and validators, including inherited
-3. attaching an `LLMProfileRecorder`;
+pipeline profiles. Inspection resolves the profile and its selected backend and
-4. creating a scheduler from the effective concurrency limit;
+target without loading a prompt, reading credentials, admitting capacity, or
-5. returning a `ScheduledClient` wrapper;
+contacting a provider, so a missing or invalid explicit profile fails before
-6. decorating that shared client before preparation when debug recording is
+stage execution while a valid `api_key_env` may remain unset. Calls record the
-   enabled; and
+profile actually selected by PromptKit. The recorder trims and deduplicates
-7. injecting that one shared client into complete pipeline preparation before
+non-secret profile identity, provider, model, selected backend ID, and
-   the source file is read or the runner is invoked.
+effective reasoning values for manifest use. Entries that differ in backend or
+reasoning remain distinct and deterministically ordered. Endpoint-only profiles
+retain an empty backend ID, which the published JSON omits. Successful
+completion responses and recorded profile manifests identify the adapter
+provider as `promptkit`.
 
-The D&D scene chunker and spell, NPC, and combat extractors retain this
+The CLI's profile-inspection engine and the production adapter use the same
-injected client and use it for every structured completion. Operation requests
+profile-source construction to apply the configured profile directory or file,
-do not carry an LLM client.
+the optional registered fallback profile assets, and the optional conventional
+`local` backend. Preflight therefore resolves the same profile sources and
+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
+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.
 
-The CLI separately gathers explicit profile IDs from resolved LLM-capable stage
+Before execution, the adapter also contributes a non-secret checkpoint
-and validator bindings. It prepares a small internal check prompt for each ID so
+fingerprint for the effective PromptKit profile source. It combines the
-missing or invalid profiles fail before pipeline execution. The runtime profile
+identity of PromptKit's compiled-in profile catalog with a deterministic digest
-override syntax and scope are defined in the
+of every YAML profile in the configured profile directory, or of the configured
-[CLI reference](../cli.md#run); binding rules are defined in
+profile file, and a deterministic digest of the flattened fallback profile
-[Configuration](../config.md#module-bindings).
+assets. The fingerprint contains neither profile content nor source paths. It
+covers inherited pipeline profiles, explicit binding profiles, and
+prompt-selected defaults, so changing a model or other profile setting cannot
+reuse checkpoints created under the
+prior profile source. This cache identity is independent of durable
+profile provenance: run manifests continue to list only profiles actually
+observed during LLM calls. When the local backend is registered, a second
+fingerprint hashes its trimmed endpoint behind a stable marker. Changing that
+semantic execution target invalidates checkpoint reuse. The raw endpoint is not
+stored in checkpoint identity, and the local concurrency limit is excluded
+because it changes scheduling rather than execution semantics.
 
-## Scriptorium Adapter
+## Shared Provider-Call Limit
 
-`ScriptoriumClient` converts a Notarius request into a Scriptorium `RunRequest`.
+Production construction creates one PromptKit client and wraps it in one
-It validates the decoding target and prompt identity, maps named input materials
+scheduled client. The scheduler has a fixed, positive permit limit, serves
-to inline artifacts, forwards explicit profile and session context, delegates
+queued calls in FIFO order, and removes a queued call when its context is
-rendering/provider execution/structured validation, and unmarshals successful
+cancelled. It rechecks the caller context after admission and before dispatch.
-JSON into the caller target.
+A granted permit is released exactly once on every completion path.
 
-Empty optional input material is represented by a single space so Scriptorium
+The scheduled wrapper surrounds every `CompleteStructured` call, so concurrent
-retains the named input. The client returns Scriptorium's validated structured
+lanes, pipeline retries, and LLM-backed validators share the same provider-call
-bytes rather than re-encoding the caller target, allowing modules to preserve
+ceiling. This ceiling is independent of pipeline worker concurrency; changing
-the runtime result exactly.
+worker counts cannot exceed the configured LLM limit. The configuration field
+and its effective default are owned by
+[Configuration](../config.md#concurrency-output-cache-and-debug).
 
-Selected profile, provider, model, and token metadata are mapped into the
+PromptKit applies a second, independent admission limit when the selected
-Notarius response. The recorder deduplicates profiles by identity and supplies
+profile names a limited backend. It sits beneath the Notarius scheduled client,
-manifest-safe profile summaries after actual calls; manifest population does
+so it may narrow but cannot expand the application-wide limit. Built-in
-not guess the selected prompt default in advance.
+OpenRouter profiles select PromptKit's reserved backend and its upstream
-
+capacity policy. A positive configured local-backend limit bounds active local
-Generated-output validation failures and provider failures are wrapped with
+generations inside PromptKit; zero leaves that backend unlimited there.
-prompt context. Error strings pass through bearer-token redaction before they
+Endpoint-only profiles do not select a PromptKit backend and remain limited
-cross the runtime boundary.
+only by the Notarius scheduler.
-
-## Scheduling
-
-`Scheduler` uses a bounded permit count and a FIFO waiter queue. Immediate
-acquisition increments the in-flight count; queued acquisition waits for a
-permit or context cancellation. Cancellation removes a queued waiter, while a
-cancelled waiter that has already received a permit releases it.
-
-`ScheduledClient` acquires a permit around each structured completion and
-defers release on every result path. The effective limit and default are
-configuration facts in [Configuration](../config.md#defaults).
-
-This provider-call ceiling is independent of the pipeline's extract worker
-limit. Concurrent lanes, retries, and validators all use the same scheduled
-client, so increasing framework workers cannot exceed `total_llm`. Pipeline
-dispatch and cancellation mechanics are documented in
-[Pipeline Internals](pipeline.md#execution-flow).
 
 ## Prompt And Schema Assets
 
-`AssetRegistry` combines caller-owned prompt filesystems under stable prefixes
+An `AssetRegistry` collects prompt, schema, and optional fallback-profile
-and rejects invalid or conflicting registrations. Production module packages
+filesystems from production module families. It flattens registered roots into
-register their own prompt and schema assets; generic framework code contains no
+the corresponding PromptKit filesystems and rejects invalid roots, unreadable
-D&D prompt content. `internal/framework/promptfs` provides the domain-neutral
+assets, duplicate paths, and missing prompt or schema files during preparation.
-filesystem composition helper used to combine module-owned files with shared
+Fallback assets receive a safe content digest for checkpoint identity; raw
-domain prompt fragments.
+paths and bytes are never included. The framework’s `promptfs` helper combines
+module-selected prompt files with reusable domain fragments without making the
+framework depend on D&D content.
 
-The D&D scene chunker and spell, NPC, and combat-turn extractors each declare an
+LLM-facing content is embedded once by the root `assets` package. Each consumer
-ordered prompt asset manifest. The manifest lists the package-owned YAML and
+uses only its scoped subtree, while the module retains ownership of its prompt
-Markdown files, then the exact shared fragments rendered by that prompt; the
+declaration, ordered manifest, private response-schema identity, and
-same ordered list drives both filesystem mounting and the prompt fingerprint.
+registration. Shared D&D fragments are selected by D&D's shared implementation;
-Unused shared assets are neither mounted nor fingerprinted. Universal
+the detailed convention is in [D&D Module Internals](dnd.md). This physical
-extraction-evidence and output policy lives only in the shared extraction
+arrangement and its data-only boundary are defined by
-assets; package-owned prompt files retain artifact-specific rules. The scene
+[Architecture](../policy/architecture.md) and
-prompt keeps its separate output rule because it does not render the
+[ADR-0011](../adr/0011-centralize-llm-assets.md), rather than by this runtime
-extraction-evidence asset.
+guide.
 
-### D&D Extraction Prompt Ordering And Cache Boundaries
+The generic registrar is the sole production registration owner for the
+semantic-reconciliation default prompt and private response schema. The
+domain-neutral reconciliation package also exposes only its mandatory protocol
+and candidate/transcript presentation files for domain prompt manifests. D&D
+registry normalizers mount those files while retaining ownership and hashing
+of their D&D system message, semantic instructions, and complete prompt
+declaration. The response schema is therefore registered once even though
+several typed normalizers select it.
 
-D&D extraction prompts order messages from the most reusable content to the
+Mounted prompt assets determine a module's fingerprint. The fingerprint hashes
-most variable content. New extraction lanes use these tiers in order:
+only the module and shared files explicitly selected by its manifest, so an
+unrelated asset does not invalidate a checkpoint. Schema loaders validate JSON,
+attach identity and digest metadata, make defensive copies, and expose
+diagnostics without raw schema bytes.
 
-1. universal shared content, including the system, extraction-evidence, and
+Semantic-reconciliation normalizers extend this identity with the shared
-   in-world identity messages;
+response-schema digest, framework policy version, and complete limit-policy
-2. stable campaign or run context shared across lanes, including campaign
+digest. Their manifest metadata records the same content-free prompt, schema,
-   references;
+policy, and limit identities together with domain identity and normalization
-3. stable subset- and lane-specific context and instructions, including an NPC
+policies. Request-local handles, source material, proposal content, and raw
-   registry, catalog, task, or extraction instructions when applicable;
+asset bytes are not checkpoint metadata.
-4. the chunk transcript as the final user message.
 
-This ordering lets requests reuse the longest identical prefix before the
+Private response schemas validate a model transport envelope. They are not the
-per-chunk transcript changes. Cache reuse requires the preceding message
+durable artifact schema and should not be documented as an external wire
-sequence and content to be exactly identical; semantic similarity is not
+contract. Durable formats and compatibility rules remain in the
-sufficient. Cache boundaries belong at the ends of reusable stable tiers,
+[integration contracts](../integrations/).
-subject to the provider's cache-boundary limit. The shared identity and
-campaign-reference messages form the first two extraction boundaries. Spell
-and combat prompts add a boundary at the shared NPC registry. Each extraction
-prompt places its final boundary on its lane-specific instructions, immediately
-before the transcript. The transcript does not carry cache control because no
-reusable content follows it.
 
-Accordingly, the common prefix of all three extraction prompts is system,
+## Prompt Maintenance And Backend Caching
-extraction evidence, identity, and campaign references. The NPC prompt then
-renders task, instructions, and transcript. Spell renders the NPC registry,
-catalog, task, instructions, and transcript. Combat renders the NPC registry,
-task, instructions, and transcript. The
-scene chunker is not an extraction lane: it retains its separate system,
-transcript, campaign-reference, task, and instruction order and marks its
-transcript and campaign-reference messages ephemeral.
 
-Shared wording belongs in the canonical assets under
+Prompt message order and shared asset bytes are runtime behavior. Backend cache
-`internal/modules/dnd/shared`; extraction packages reference those assets in
+reuse depends on identical preceding roles, rendered bytes, and cache-control
-their manifests instead of copying similar text into package-local files.
+metadata—not merely equivalent meaning. Keep reusable shared assets
-Package-local assets contain only lane-specific content. An extraction lane may
+byte-identical and preserve each prompt’s declared ordering and cache controls
-depart from the tier order only when prompt-quality evidence or a provider
+when editing it.
-constraint makes the exception necessary; document the exception and rationale
-here when it becomes implemented behavior.
 
-Schema helpers load embedded JSON Schema with identity and digest metadata,
+For sibling prompts that can reuse the same source material, order universal
-return defensive copies, and expose a diagnostics map that omits schema bytes.
+shared context first, request source material next, and module-specific
-The small framework registry contains only generic test schemas; production
+suffixes last. Put a cache boundary at a reusable prefix that is useful to the
-schemas remain package-owned.
+backend. Redundant intermediate cache boundaries do not extend that reusable
+prefix and add no value.
 
-The spell, NPC, and combat extractors' package-owned prompts declare their
+Prompt-family owners may choose a different sequence when their inputs and
-structured JSON inputs and private response schemas. Each private response
+reuse pattern differ. The D&D family’s extraction, scene-chunking, and NPC
-schema remains separate from its durable artifact codec schema; this work does
+normalization policies are maintained in [D&D Module Internals](dnd.md#prompt-construction).
-not use shared schema fragments or schema generation. Those private schemas own
+Do not add tests that enforce prompt prose; prompt tests should verify the
-the transport envelope—required fields, JSON types, nullability, and
+meaningful input placement and cache controls of the prompt being changed.
-unknown-field rejection—while deterministic validators own semantic constraints
-such as enum membership, non-empty values and collections, and positive
-numbers. The spell extractor's prompt declares a required
-`application/json` `spell_catalog` input and an optional `application/json`
-`npcs` input. The extractor generates
-the catalog input from its prepared
-effective catalog as `{"spell_names":[...]}` using sorted canonical names only.
-The shared D&D prompt assets include a generic NPC grounding fragment directly
-after the campaign reference message for spell and combat prompts. When an NPC
-registry is bound, the
-domain registry boundary strictly decodes and identity-validates one durable
-artifact, re-encodes canonical JSON for provenance, and separately generates a
-names-only prompt projection. The unbound projection is exactly `{"npcs":[]}`.
-Prompt input and component-local checkpoint digests cover the projected bytes;
-manifests retain the optional full registry digest/count rather than names,
-overlay bytes, registry paths, or source metadata. Combat prompt,
-response-schema, mapping, normalization, identity, and registry-projection
-fingerprints remain separate semantic inputs to checkpoint identity.
 
-## Debug And Redaction Boundaries
+## Validation, Repair, And Retries
 
-The pipeline may wrap the client with a debug recorder that captures prepared
+PromptKit performs prompt rendering, provider execution, and the prompt’s
-prompt/response material for an explicitly requested debug run. Debug summaries
+structured-output validation. The adapter reports an empty result, validation
-and manifests receive identities, hashes, usage, and selected profile summaries
+failure, empty structured body, or decode failure as
-rather than prompt, source, reference, schema, or response content.
+`ErrInvalidStructuredOutput`, while retaining the returned raw bytes and debug
+material when they exist. Provider failures remain operational errors rather
+than output-validation failures. Apart from documented context, capacity, and
+invalid-output categories, provider error values and types do not cross the
+adapter error chain; callers receive only a credential-redacted diagnostic.
 
-The Scriptorium error wrapper removes bearer credential values from surfaced
+When PromptKit rejects backend admission before generation, the adapter maps
-provider errors; `RedactSecrets` and `ErrorWithSecretsRedacted` support known
+`promptkit.ErrCapacityExceeded` to
-secret values elsewhere in the runtime. Config summaries use a separate
+`contracts.ErrLLMCapacityExceeded`, retaining prompt context and a redacted
-clone-and-redact path in `internal/core/config`. These mechanisms implement the
+upstream diagnostic without exposing the PromptKit sentinel or capacity-error
-security invariant in
+type as a framework contract. When supplied, the normalized selected backend
-[Architecture](../policy/architecture.md#state-output-and-safety); operator
+ID appears only in that safe application-owned diagnostic context. A canceled
-handling of debug data is defined in [Operations](../operations.md#debug).
+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.
 
-## Failure Behavior
+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
+[Pipeline Internals](pipeline.md#validation-retries-and-output) and the
+[binding reference](../config.md#module-bindings-and-validators).
 
-- Invalid targets, missing prompt IDs, malformed structured output, and
+## Timeout Ownership
-  Scriptorium failures return contextual errors to the calling module.
-- Scheduler construction rejects non-positive limits; acquisition respects
-  context cancellation.
-- Asset registration rejects invalid roots, missing content, and path conflicts.
-- Schema loading distinguishes missing assets, invalid JSON, and invalid
-  metadata.
-- Profile validation errors occur during CLI preparation when an explicit
-  selected ID cannot be prepared.
 
-## Tests To Inspect
+The caller context remains the outer cancellation authority. PromptKit applies
+a positive effective generation timeout as an inner request deadline; an
+explicit zero disables only that generation deadline. The HTTP client timeout
+is a separate transport-wide cap. Notarius forwards the caller context and
+does not install another timeout wrapper around PromptKit.
 
-- `internal/framework/llm/scriptorium_client_test.go`: adapter mapping and local
+The selected PromptKit profile owns generation settings. Notarius binding
-  HTTP integration.
+retries remain outside the adapter and repeat the complete module operation
-- `internal/framework/llm/scheduler_test.go` and
+and validation chain. PromptKit does not add a provider retry loop.
-  `scheduled_client_test.go`: permits, FIFO behavior, cancellation, and wrapper
+Operator-facing behavior is summarized in
-  release.
+[Operations](../operations.md#operational-limits), and the pinned upstream
-- `internal/framework/llm/asset_registry_test.go` and
+contract is identified in
-  `schema_registry_test.go`: asset composition, validation, and defensive
+[PromptKit Integration](../integrations/pkg-promptkit.md).
-  copies.
+
-- `internal/framework/llm/secrets_test.go`: provider-error redaction.
+## Observability And Redaction
-- `internal/cli/run_contract_test.go`: profile validation, production client
+
-  wiring, manifest recording, and debug integration.
+When debug recording is enabled, the pipeline decorates the shared client. The
-- Module-local `scriptorium_assets_test.go` files: prompt inputs and package
+wrapper records prepared prompt and response material, timing, selected profile
-  asset registration.
+and backend, effective model parameters, and call identifiers in the run’s
+debug bundle, including material available from a failed structured completion.
+Effective parameters use PromptKit's stable lower-case JSON field names and may
+include `backend_id`. For a successful completion, a debug-write failure is
+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).
+
+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
+to the [JSON output contract](../integrations/json-output.md#manifestjson).
+Provider error text is wrapped with prompt context and bearer credentials are
+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.
+
+## Failure Boundaries
+
+- Construction fails for missing asset registries, mutually exclusive profile
+  sources, invalid asset registration, or a non-positive scheduler limit.
+- Preparation failures, unavailable explicit profiles, provider failures, and
+  context cancellation propagate to the calling stage with context.
+- Backend admission exhaustion is a provider-neutral operational error and is
+  not classified as invalid structured output or validator rejection.
+- Malformed or schema-invalid provider output is classified separately as
+  invalid structured output so the module or pipeline can apply its own retry
+  and rejection policy.
+- Domain semantic checks, evidence decisions, and deterministic normalization
+  run outside the provider adapter.
+
+## Focused Verification
+
+Read the LLM adapter, scheduler, asset registry, schema loader, and redaction
+tests when changing this boundary. Prompt changes also require the owning
+module’s asset tests, and retry or debug changes require focused pipeline or
+CLI coverage. The focused runtime and D&D checks are:
+
+~~~sh
+go test ./internal/framework/llm/... ./internal/modules/dnd/...
+~~~

docs/internal/modules.md

@@ -1,497 +1,152 @@
-# Module And Validator Internals
+# Module Internals
-
+
-Production module and validator implementations live under their domain-first
+This guide owns the mechanics for implementing and registering production
-trees in `internal/modules`.
+modules. [Configuration](../config.md) owns selectable keys, binding syntax,
-The selectable keys, configuration options, reference slots, and default
+reference configuration, and default validator chains. Durable input and output
-validator chain are canonical in the
+shapes belong in [integration contracts](../integrations/).
-[module](../config.md#implemented-production-modules) and
+
-[validator](../config.md#implemented-production-validators) catalogs in
+The D&D family has additional shared conventions and domain-specific
-Configuration.
+exceptions. See [D&D Module Internals](dnd.md) rather than adding them here.
-
+
-## Extension Pattern
+## Module Boundary
-
+
-A stage module package provides a stable key, constructor, contract
+A module is a typed implementation registered for one pipeline stage. Its
-implementation, `ModuleSpec`, `Register`, and focused behavior and registration
+`ModuleSpec` is the public-to-the-framework declaration of its stable key,
-tests. A validator package follows the same pattern with `ValidatorSpec` and the
+stage, execution class, required and provided capabilities, artifact kind, and
-validator registry. Package-family registrars compose those leaf registrations
+accepted reference slots. The execution class states whether a module is
-into the production catalog and own family-level policy such as default
+`deterministic` or `llm_backed`; registries retain it for catalog inspection and
-validator chains and prompt asset collection.
+resolved-pipeline debug data without constructing the module. The framework
-
+uses the declaration to resolve a configured binding before it builds the
-Production input, chunk, output, and D&D spell- and combat-extract packages
+implementation. After selection, the resolver applies profile inheritance only
-register strict option decoders and run-local builders. Preparation decodes their options into
+to bindings whose declared execution class is `llm_backed` and rejects a
-implementation-owned values and injects dependencies plus the materialized
+binding-specific profile on a deterministic module. The user-facing precedence
-reference set for the selected target. Each builder receives an isolated clone
+contract belongs in [Configuration](../config.md#pipelines).
-of that set; input and output builders receive no references. The spell and
+
-combat extractors are typed over the canonical D&D model. D&D validators, merge,
+Implementations that accept options must provide both an option validator and
-and normalize use typed variants; JSON representation validators use serialized
+a builder. The validator is used while resolving configuration; the builder
-requests; and unconditional validators expose separate chunk and typed
+decodes the same options and constructs the implementation from the prepared
-variants. The D&D production registrar registers the canonical typed spell,
+`BuildRequest`. Reject unknown options in both paths. A builder receives only
-NPC, and combat implementations, including their kind-specific merge and
+the dependencies and materialized references that the framework prepared for
-normalize behavior.
+that operation, so it must not re-read configuration or files.
-
+
-For D&D artifact defaults, generic JSON syntax validation runs first. Rejecting
+Registry helpers register the typed builder for a stage-specific registry.
-domain validators then own semantic diagnostics before generic JSON Schema
+They are preferable to hand-written untyped registration because they retain
-validation provides the final rejecting representation backstop; warning-only
+the artifact type at the framework boundary. Registrars validate the registries
-relatedness validators run last. This default composition does not reorder an
+they need, register each leaf implementation, and add any family-owned assets
-explicitly configured validator chain.
+or default validator chains. They return contextual errors so production
-
+composition fails at startup rather than at the first run.
-Prepared extractors, extract validators, and codecs may be reused concurrently
+
-by the run-wide extract pool. Production implementations are immutable after
+An artifact family can register an optional typed evidence projector alongside
-construction: they retain only typed options, immutable assets, or the shared
+its codec. The projector returns defensive copies of the artifact's direct
-concurrency-safe LLM client. Implementations that introduce mutable state must
+generic source references and must use the codec's exact Go type. It does not
-synchronize that state without creating a separate provider scheduler.
+interpret surrounding context or publish files; the pipeline validates the
-
+capability during preparation and the output boundary owns publication. See
-Specs expose capability and execution metadata without constructing an
+the [Published Evidence Context contract](../integrations/evidence-context.md)
-implementation. Registry entries separately expose option validation and
+for the durable source-unit excerpt. Lane artifacts retain citation and lane
-run-local construction. Chunk, extract, merge, and normalize modules that accept
+provenance; the framework does not add either to that published excerpt.
-auxiliary material declare identical reference slots from both
+
-`ReferenceSlots()` and `ModuleSpec().ReferenceSlots`; registration tests enforce
+An artifact family is broader than a module: it owns the cohesive domain
-that agreement. Runtime delivery uses the corresponding stage request's
+feature across its artifact type, codec, stage modules, validators, prompt
-`References` field.
+policy, schemas, identity helpers, and reference projections. An extractor and
-
+normalizer in one artifact family remain independently registered modules in
-LLM-backed extensions own their prompt definitions and response schemas under
+their respective pipeline stages. This ownership vocabulary does not create a
-package-local embedded assets. Shared filesystem composition belongs in
+new registry or change the fixed pipeline.
-`internal/framework/promptfs`; reusable D&D prompt fragments, reference
+
-declarations, prompt-input assembly, and source-unit/citation helpers belong in
+## Production Composition
-`internal/modules/dnd/shared`, which also owns bounded D&D diagnostics. The
+
-D&D scene chunker and spell, NPC, and combat-turn extractors use ordered
+Production composition is intentionally split by family:
-package-local prompt manifests for both rendering and prompt fingerprinting, so
+
-only the shared fragments each prompt actually renders participate in either
+- The generic registrar provides the unit chunker, generic JSON validators,
-operation. Extraction prompts place stable shared and lane-specific context
+  JSON output encoder, and shared semantic-reconciliation prompt and response
-before the variable transcript and use shared assets for wording common across
+  schema assets.
-lanes. The canonical ordering and cache-boundary policy is documented in
+- The Seriatim registrar provides the transcript input adapter. Its external
-[LLM Runtime](llm.md#dd-extraction-prompt-ordering-and-cache-boundaries). Stage
+  input behavior is defined by the [Seriatim contract](../integrations/seriatim.md).
-contracts expose only Notarius structured-completion types, not Scriptorium
+- The D&D registrar provides its codecs, extractors, mergers, normalizers,
-public types.
+  validators, prompt assets, fallback profile asset, and default chains. Its behavioral conventions
-
+  are documented in [D&D Module Internals](dnd.md).
-The shared `ChunkPromptMaterial` helper owns common transcript material
+
-preparation for the spell, NPC, and combat-turn extractors. It clones supplied
+The CLI owns the composition that invokes these registrars. A module package
-source metadata, falls back to the materialized chunk when content is absent,
+may register its own family but must not assemble the CLI or make framework
-checks that content remains chunk-identical, and fills only the common default
+packages depend on production extensions.
-fields. Extractors retain their request validation and wrap helper errors with
+
-their module context.
+## Semantic Reconciliation
-
+
-Reference material may inform a module or prompt but must not become source
+`internal/framework/semanticreconcile` is a domain-neutral strategy used by a
-evidence. The resolver and materializer behavior is described in
+typed normalize module; it is not itself a selectable stage module. A
-[Pipeline Internals](pipeline.md#reference-materialization).
+source-backed artifact-family normalizer projects its deterministic records
-
+into contextual candidates and owned typed record envelopes, supplies its
-## Domain Reference Data
+chosen prompt identity and resolved LLM profile, and constructs an engine with
-
+explicit limits. The core filters invalid evidence, assigns contiguous
-### `internal/modules/dnd/spells/catalog`
+request-local integer handles, renders bounded candidate and transcript
-
+materials, invokes the structured-completion boundary, and assesses the
-The spell catalog package owns the embedded, versioned D&D 5e 2014 SRD spell
+returned duplicate groups into a stable non-overlapping plan.
-reference data. Its strict JSON asset contains one canonical record per spell,
+
-including spell level and all applicable class memberships. `LoadSRD5E2014`
+The normalizer then applies that plan through a typed `ApplicationPolicy`. The
-validates catalog identity, provenance metadata, ordering, uniqueness, levels,
+core preserves ungrouped records, contribution order, and provenance while the
-classes, aliases, and lookup-key collisions before exposing immutable copies.
+artifact family owns group guards, field and evidence consolidation, durable
-
+ID derivation, retry and fallback presentation, warnings, and postconditions.
-Lookup is case-insensitive and normalizes whitespace and common apostrophe
+Request-local handles do not enter the typed value or durable artifact. Fewer
-variants while preserving source punctuation in canonical display names. The
+than two eligible candidates skips model invocation; exceeding a candidate or
-catalog contains 319 unique spells and 779 class memberships. Source and
+combined-material bound preserves the deterministic result under the family's
-license details live beside the asset in `SOURCES.md`. This domain-owned data is
+fallback policy. Provider, transport, cancellation, and context-construction
-separate from `internal/modules/dnd/shared`, which is reserved for reusable
+failures remain execution errors.
-prompt and source-reference machinery.
+
-
+The core supplies a conservative generic prompt and the single private
-`ResolveEffectiveCatalog` builds the immutable recognition view used by the
+response schema. A domain prompt may substitute its semantic instructions but
-spell extractor and catalog validator. It starts with the embedded SRD catalog
+mounts the core-owned protocol and candidate/transcript presentation assets.
-and optionally applies one strict JSON overlay from the `spell_catalog` item in
+Prompt, schema, policy, and limit identities participate in manifest metadata
-a materialized reference set. Overlay catalogs are ordered by ID, may add names
+and checkpoint fingerprints. The generic registrar owns production
-and aliases, and may augment an existing canonical spell without replacing its
+registration of those shared assets; a consuming domain registrar owns only
-display name. Cross-spell lookup collisions are errors. The effective view
+its domain prompt.
-exposes sorted canonical names, normalized lookup, overlay identities, and a
+
-semantic digest; overlay content remains contextual reference material rather
+## Adding Or Changing A Module
-than source evidence. Its external JSON contract is defined in the
+
-[spell-catalog overlay contract](../integrations/dnd-spell-catalog-overlays.md).
+1. Choose the pipeline stage and the typed artifact boundary. Put external
-
+   input or durable artifact formats in the relevant integration contract,
-### `internal/modules/dnd/npcs/identity`, `internal/modules/dnd/npcs/registry`, and `internal/modules/dnd/codec/npcs`
+   not in this guide or in a private LLM response type.
-
+2. Define a stable `ModuleSpec` with an explicit execution class, the exact
-The NPC identity package owns Unicode comparison keys, deterministic
+   capabilities, and reference slots needed for the operation. Model a
-`npc:sha256:` IDs, display normalization, and whole-registry collision issues.
+   producer/consumer handoff as an artifact-compatible slot; configuration
-The registry package resolves one optional normalized artifact through the
+   then chooses an external file or a generated binding.
-strict codec, validates whole-registry identity, canonicalizes its JSON, and
+3. Implement strict option decoding, construction, and the typed stage
-provides immutable records, a names-only prompt projection, distinct durable
+   interface. Preserve caller ownership: do not retain mutable request data
-and projection digests, count, and exact canonical-name lookup. External files cross this boundary during
+   and return defensive copies where an implementation exposes stored data.
-preparation; generated artifacts cross it at the ordered step handoff. It owns
+4. Register the module through its typed registry helper and add it to the
-the `npcs` slot and its bounded, content-safe validation failures. NPC source
+   owning family registrar. Add a default validator chain only when that
-references are durable provenance and are not treated as evidence for a
+   family owns the behavior; otherwise require an explicit compatible chain.
-consuming pipeline. The NPC codec owns the strict durable `dnd/npc-list` JSON
+5. Update the selectable-key and chain reference in
-boundary and exposes candidate versus approved encode/decode operations.
+   [Configuration](../config.md#production-module-keys), the applicable
-
+   integration contract, and focused tests. Keep the configuration document
-The `internal/modules/dnd/codec/combatturns` package owns the durable
+   as the sole list of production keys and validator order.
-`dnd/combat-turn-list` schema and candidate versus approved JSON boundary. It
+
-is registered by the production D&D family registrar for the selectable combat
+## Validation And References
-lane.
+
-
+Validators operate on the value produced at their configured stage. A default
-## Input Adapter
+chain is ordered behavior, not a set: JSON parsing, structural checks,
-
+domain-specific checks, durable-schema checks, and advisory checks may have
-### `internal/modules/seriatim/input/transcript`
+different responsibilities and failure handling. The active default chains and
-
+override rules are maintained in
-The adapter decodes the supported transcript JSON, selects the source identity,
+[Configuration](../config.md#production-validator-keys-and-default-chains).
-computes canonical source provenance, validates segments, and maps each segment
+
-into a generic source unit with a self-reference plus speaker and timestamp
+Reference slots are part of the module specification. They describe the
-metadata. It accepts no module options. Its spec advertises the transcript
+accepted artifact kind, media type, size, and whether a binding is required;
-capabilities consumed by D&D modules.
+the framework validates those constraints before construction. An external
-
+reference is materialized during preparation. A generated reference is a
-Parsing is strict about required values and duplicate unit IDs but deliberately
+compatible normalized artifact handed from an earlier pipeline step at
-ignores unrelated Seriatim fields. The external format and derived-identity
+operation time. The configuration reference rules, including precedence and
-rules are defined in the
+ordered-handoff requirements, are maintained in
-[Seriatim contract](../integrations/seriatim.md).
+[Configuration](../config.md#references-and-ordered-handoffs).
-
+
-## Chunkers
+## Focused Verification
-
+
-Chunkers implement `contracts.Chunker.Plan`. A plan identifies ordered source
+Exercise the leaf implementation and its registration path when changing a
-unit ranges and may carry optional namespaced JSON annotations; it does not
+module. Registry and registrar tests cover duplicate keys, required registries,
-contain materialized chunk content. The framework canonicalizes annotations,
+and typed construction; pipeline resolution tests cover capabilities, options,
-validates ranges against the current source, and materializes chunk IDs,
+and reference compatibility. Domain packages should additionally test their
-indexes, references, content, units, and generic metadata. Materialized source
+codecs, validators, normalizers, and any integration handoffs they own.
-unit metadata is independently owned. Annotation
+
-namespaces remain optional data: generic framework code and downstream modules
+Run the affected package tests while iterating. The complete module suite is:
-must not require D&D scene annotations or import `dnd/scenes`.
+
-
+~~~sh
-### `internal/modules/generic/chunk/units`
+go test ./internal/modules/...
-
+~~~
-The generic chunker validates the source document and returns ranges over units
-in configured windows. Overlap changes the next window start but never reorders
-units. Framework materialization derives the resulting chunk identity and
-generic metadata from those ranges.
-
-The accepted options and defaults are defined in
-[Configuration](../config.md#implemented-production-modules). Generic
-framework validation canonicalizes the returned unit slices before extraction.
-The chunker decodes its options during construction and retains only the typed
-window settings used by `Plan`.
-
-### `internal/modules/dnd/chunk/scenes`
-
-The scene chunker prepares a structured Scriptorium request from the full
-transcript, session, and optional D&D reference inputs. It validates the model's
-scene boundaries against source-unit IDs and converts them into deterministic
-plan ranges with optional scene annotations. Preparation injects the shared
-structured LLM client into the chunker; `Plan`
-supplies only the run-specific profile, session, source, references, and
-metadata.
-
-Scene validation requires sequential, contiguous, non-overlapping coverage from
-the first source unit through the last. Scene descriptions, boundaries,
-confidence, and participants are module-owned annotations. Boundary caveats
-become warnings. Malformed
-structured output is returned as an error; there is no fallback chunker.
-
-The package embeds its prompt and response schema and reports their non-secret
-identity and hashes through singleton module metadata. Shared D&D assets supply
-reference declarations and prompt inputs; their user-facing keys and accepted
-file types remain canonical in [Configuration](../config.md).
-
-## Extractor
-
-### `internal/modules/dnd/extract/spells`
-
-The spell extractor prepares a structured request from one chunk, the
-chunk-scoped source input, the session, and optional D&D reference inputs. It
-decodes the model response, assigns the generic source identity to every source
-reference, canonicalizes duplicate references, orders spell casts by their
-earliest cited unit, and returns `dnd.SpellList`.
-
-Its private response schema admits only the structural transport envelope:
-required fields, JSON types, array and object shapes, and unknown-field
-rejection. It maps integer source-unit candidates directly without repairing
-semantic values, so the deterministic shape, catalog, and source-reference
-validators own blank values, empty evidence, and invalid or unresolved ranges.
-
-The extractor owns its private model-response DTO, embedded prompt, LLM response
-schema, strict option decoder, injected shared LLM client, and prompt/schema
-manifest metadata. During preparation it resolves the optional `spell_catalog`
-reference into an immutable effective catalog and adds a generated
-canonical-name-only JSON input to every structured completion request. Overlay
-failures therefore stop construction before source parsing or an LLM call;
-campaign references remain separate disambiguation inputs and never become
-source evidence.
-
-The prompt includes only actual casting events and unambiguous declared casting
-attempts. Spell mentions, plans, rules discussion, and catalog matches without
-a casting event are excluded. Shared extraction-evidence and identity rules
-require transcript-supported caster and spell facts, while the catalog,
-campaign references, and NPC names only disambiguate source text. Structural
-source validation remains deterministic; semantic evidence sufficiency is
-enforced through extraction policy and evaluation.
-
-Both the extractor and deterministic catalog validator expose
-the effective base-plus-overlay semantic digest as scoped prepared-component
-checkpoint identity. Raw overlay provenance independently covers file-byte
-changes, while the semantic digest also invalidates reuse when the embedded
-catalog or catalog composition changes. The extractor additionally fingerprints
-its complete prompt assets and private response schema, so either semantic
-contract changing invalidates previously recorded extraction checkpoints. The
-separate `internal/modules/dnd/codec/spells` package
-owns the durable schema and stable JSON representation for artifact kind
-`dnd/spell-list`. The runner keeps the result typed through validators and later
-stages, using the codec only for checkpoint, debug, and output boundaries.
-Shared D&D helpers keep prompt input names and source-unit reference conversion
-consistent with the scene chunker.
-
-The extractor also declares the optional `npcs` registry slot and consumes the
-immutable registry boundary from `internal/modules/dnd/npcs/registry`. An
-external registry is prepared before execution; a generated registry is
-validated and supplied at operation time. Bound external registries add only
-the full `npc_registry_digest` and `npc_count` to module metadata. The local
-`npc_registry` checkpoint fingerprint always covers the names-only projection,
-including its exact unbound value. Generated bindings are represented by
-framework handoff provenance and dependency fingerprints. The unbound prompt
-input is exactly `{"npcs":[]}` and has no registry provenance.
-The shared NPC grounding fragment is placed immediately after the common
-campaign reference message and is included in the spell prompt fingerprint.
-
-The durable payload and manifest metadata shapes are defined in the
-[D&D spell artifact contract](../integrations/dnd-spell-artifacts.md).
-
-### `internal/modules/dnd/extract/npcs`
-
-The NPC extractor maps private model output to the canonical `dnd.NPCList`,
-assigns source identity and deterministic NPC IDs, and preserves source
-references for deterministic validation. It uses the shared campaign
-references only for disambiguation and does not consume the optional NPC
-registry slot. Its prompt and private response schema are package-owned. The
-private response contains only a name and model-facing evidence ranges for each
-record; anonymous groups, generic roles, invented labels, descriptions,
-aliases, and relationships are outside its contract. The
-prompt follows the shared D&D extraction ordering and cache policy documented
-in [LLM Runtime](llm.md#dd-extraction-prompt-ordering-and-cache-boundaries).
-
-The private response schema owns only structural transport validation and maps
-integer source-unit candidates unchanged. Required semantic content, non-empty
-evidence, and valid source ranges are rejected by the deterministic shape and
-source-reference validators.
-
-### `internal/modules/dnd/extract/combatturns`
-
-The combat extractor prepares one structured request per supplied chunk using
-the shared extraction-evidence, identity, campaign-reference, NPC-grounding,
-and transcript prompt inputs. It
-maps the private response to `dnd.CombatTurnList`, assigns the current source
-identity, removes exact duplicate source ranges, and orders turns by valid
-source-document position while preserving malformed candidate fields for
-deterministic validators. Its package-owned private response schema enforces
-only the structural JSON envelope; semantic artifact constraints remain with
-the validator chain. Its prepared metadata and checkpoint fingerprints contain
-only prompt/schema/mapping identities plus an optional NPC registry digest.
-The prompt follows the shared D&D extraction ordering and cache policy
-documented in
-[LLM Runtime](llm.md#dd-extraction-prompt-ordering-and-cache-boundaries). The
-package exposes typed registration and is included in the production D&D
-registrar with the default combat extraction chain.
-
-The combat normalizer accepts only the optional structured NPC registry.
-Campaign references remain extractor-only LLM context and are not materialized
-for deterministic normalization.
-
-### `internal/modules/dnd/normalize/npcs`
-
-The NPC normalizer performs deterministic identity-aware consolidation in
-merged input order. It consolidates only equal canonical-name comparison keys,
-retains the first display record, and unions exact source references. It exposes
-the identity policy as its local checkpoint fingerprint and emits bounded
-normalization warnings.
-
-## Merger And Normalizer
-
-### `internal/modules/generic/merge/appendorder`
-
-The merger passes typed values to an injected combine function in framework
-source-chunk order. The D&D registrar specializes it with a spell-list append
-function.
-
-### `internal/modules/generic/normalize/noop`
-
-The normalizer returns the merged domain value unchanged and is reusable for
-any registered artifact type.
-
-### `internal/modules/dnd/normalize/spells`
-
-The typed spell normalizer resolves the optional `spell_catalog` reference into
-the same immutable SRD-plus-overlay effective catalog used by spell extraction
-and catalog validation. It performs no LLM calls. For each spell cast it
-canonicalizes recognized names using the catalog's case, whitespace,
-apostrophe, and alias rules; sorts source references by source identity and
-unit boundaries; removes only exact reference duplicates; and emits bounded,
-scoped warnings for each mutation or unresolved name.
-
-After those per-cast changes, it collapses only casts with the same canonical
-spell, case-folded and whitespace-normalized caster, and complete non-empty
-valid source-reference set. It retains the first occurrence and its caster,
-source references, and stable order. Unknown names, empty or invalid evidence,
-and adjacent or overlapping but different ranges remain unchanged for
-validation.
-
-The normalizer exposes the effective catalog digest as its independently scoped
-`effective_catalog` checkpoint fingerprint and reports catalog base ID, digest,
-and overlay IDs as manifest metadata. Catalog contents, reference paths, and
-raw overlay bytes are not included in either surface. The normalize-stage
-reference is stage-local, so an overlay-capable pipeline binds the catalog
-independently for extraction and normalization.
-
-### `internal/modules/dnd/normalize/combatturns`
-
-The combat normalizer prepares an external NPC registry before execution or
-receives a generated registry at the ordered step handoff, then uses the
-immutable view during runtime. It display-normalizes actors,
-rewrites canonical-name matches for actors, orders and deduplicates source
-references, stable-sorts records by source-document position, and collapses
-only exact duplicate identities with fully valid evidence. It deep-clones
-output storage and emits bounded warnings scoped to merged input indexes. Its
-metadata and fingerprints identify the normalization and NPC identity policies.
-External bindings may contribute registry
-digest/count metadata; generated identity is retained in framework provenance
-and dependency fingerprints. The normalizer is included in the production D&D
-registrar with the default combat normalization chain.
-
-## Output Encoder
-
-### `internal/modules/generic/output/json`
-
-The JSON encoder sorts normalized results by lane, derives collision-checked
-safe logical names, pretty-prints JSON payloads, and assembles the logical index,
-manifest, rejected-result, warning, and lane files. Invalid JSON, unsupported
-media types, unsafe names, and sanitized-name collisions are errors.
-
-The encoder returns logical files only. The CLI places them on disk, and the
-[JSON output contract](../integrations/json-output.md) defines their external
-paths and schemas.
-
-## Generic Validators
-
-The generic validator implementations live under
-`internal/modules/generic/validate`.
-
-The unconditional accept and reject validators provide explicit chunk and
-typed-artifact variants used primarily for controlled composition and tests.
-
-The serialized JSON syntax validator uses `encoding/json` to reject malformed
-representation bytes. The serialized JSON Schema validator requires schema
-bytes, parses the instance and schema with `jsonschema`, and distinguishes
-payload rejection from schema loading or compilation errors. The framework
-serialized-validation request carries either canonical chunk bytes or artifact
-codec bytes according to its target context. Neither validator calls the LLM.
-
-## D&D Spell Validators
-
-All four validators receive `dnd.SpellList` directly. The shape validator
-rejects a missing list, blank caster or spell names, and empty reference lists.
-The catalog validator defers when shape is invalid, then checks every non-empty
-spell name against the immutable effective SRD and overlay catalog. It accepts
-normalized canonical names and aliases without rewriting the artifact; unknown
-names reject the complete result with bounded, stable index/name diagnostics. The
-source-reference validator defers malformed shapes, validates every cited
-range, and reports all range defects through a bounded aggregate while
-preserving `invalid_source_refs`. The relatedness validator resolves all cited
-ranges through the shared document-order traversal, then warns when a normalized
-consecutive spell-name token sequence is absent from the cited source text.
-Invalid shape
-or cited ranges produce no relatedness warnings; the shape and source-reference
-validators own those defects.
-
-These validators are deterministic. Shape, source-reference, and relatedness
-each expose a local semantic `policy` checkpoint fingerprint. The catalog
-validator instead exposes its effective catalog digest as its semantic
-checkpoint identity and does not add a separate policy fingerprint. Their
-selectable keys and production order are defined in
-[Configuration](../config.md#implemented-production-validators); their durable
-payload rules are defined in the
-[artifact contract](../integrations/dnd-spell-artifacts.md).
-
-## D&D NPC Validators
-
-NPC shape validation checks the required ID and name strings, list presence, and source-reference
-shape. The source-reference validator defers malformed shapes, checks
-current-document identity, unit existence, and range ordering, and reports all
-defects through bounded aggregates. Source relatedness uses the shared
-document-order traversal and normalized consecutive-token matching, emitting at
-most one bounded warning per record when the canonical name does not occur near
-its cited text. Invalid shape or cited ranges produce no relatedness warnings.
-Normalize identity validation checks deterministic IDs, canonical names, and
-duplicate canonical-name or ID ownership.
-All are deterministic and expose the policy fingerprints used by the
-production chains.
-
-## D&D Combat Validators
-
-Combat shape validation owns the required list, actor, supported turn kind, and
-non-empty source-reference collection. Combat source-reference validation defers invalid
-shape, checks source identity, unit existence, and range order, and reports all
-defects through bounded aggregates. Combat source-relatedness defers invalid
-shape or ranges, uses the shared traversal to combine overlapping cited units
-in document order, and emits at most one bounded advisory warning per turn for
-an unrelated actor. Actors use normalized consecutive-token matching. The
-normalized-invariants validator owns actor display normalization, canonical
-source-reference order, chronology, and exact duplicate identity; it defers
-shape and source-reference failures. All four validators are deterministic and
-expose local policy fingerprints. In the registered defaults, JSON syntax runs
-first; combat shape, normalized invariants when applicable, and source-reference
-validation precede JSON Schema validation; warning-only relatedness runs last.
-
-## Production Registration
-
-Production composition occurs through family registrars. The CLI allocates one
-complete framework registry set and one LLM asset registry. It invokes
-`internal/modules/generic/register`,
-`internal/modules/seriatim/register`, and `internal/modules/dnd/register` in
-that order, then exposes the matching catalog for resolution. The generic and
-Seriatim registrars own their production leaf registrations. The D&D registrar
-owns D&D leaf registrations, typed spell, NPC, and combat default-validator
-chains, typed append-order specializations, and D&D prompt/schema asset
-collection. Its registration helpers group module, validator, prompt-asset, and
-chain composition while retaining artifact-specific merge and clone behavior in
-the registrar.
-
-Concrete implementation packages do not import generic implementation
-packages directly. A concrete family's `register` package is its composition
-point for specializing reusable generic implementations, while the generic
-registrar composes only generic children.
-
-Core and framework production packages do not import production extensions.
-CLI production code is the sole application composition root for extensions
-and imports only exact family registrar packages. Other production packages,
-including commands and newly introduced package trees, do not import module
-packages directly. Compatibility tests in the CLI, core, and framework trees
-may import roots and implementation leaves directly. Other non-module tests do
-not receive that exemption. White-box tests within module families retain the
-production family boundaries. `internal/modules/integration` is test
-infrastructure: its black-box tests may compose multiple families, but it is
-not a production module family or production dependency target.
-
-## Adding An Extension
-
-When adding a production module or validator:
-
-1. implement the stage or validator contract and package-local key;
-2. expose and test its spec, constructor, and registration function;
-3. keep format or domain parsing inside the concrete package;
-4. add package-owned prompt/schema assets when the extension is LLM-backed;
-   new LLM-backed D&D extraction modules must follow the stable-to-variable
-   prompt ordering, shared-asset ownership, and cache-boundary policy in
-   [LLM Runtime](llm.md#dd-extraction-prompt-ordering-and-cache-boundaries), or
-   document the implemented exception and its evidence there;
-5. register it through its package-family registrar and add a default chain
-   there only when production policy requires one;
-6. add resolution and composition coverage for capabilities, options,
-   references, and validation behavior;
-7. update the selectable-key catalog in [Configuration](../config.md), the
-   relevant external contract, this inventory, and maintained examples when
-   user-visible behavior changes.
-
-Do not add the extension to `docs/development.md`; that file routes by task and
-does not inventory implementations.
-
-## Tests To Inspect
-
-- Package-local `*_test.go` files under the module or validator being changed.
-- `internal/framework/pipeline/typed_resolution_test.go`: typed registry, spec,
-  and heterogeneous artifact composition.
-- `internal/framework/pipeline/profile_test.go`: framework binding defaults and
-  profile resolution.
-- `internal/cli/production_contract_test.go`: production catalog, config
-  resolution, and composition smoke coverage.
-- `internal/cli/example_contract_test.go`: maintained example ownership.
-- `internal/framework/promptfs/*_test.go` and
-  `internal/modules/dnd/shared/*_test.go`: shared prompt and reference assembly.
-- `internal/modules/integration/*_test.go`: black-box composition across
-  production extension domains.

docs/internal/overview.md

@@ -1,171 +1,61 @@
 # Internal Overview
 
-This document inventories the implemented Notarius components. Normative
+This document is the implemented component map for Notarius. Normative
 boundaries and dependency direction belong in
-[Architecture](../policy/architecture.md); external behavior belongs in the
+[Architecture](../policy/architecture.md). User and operator contracts belong
-[CLI](../cli.md), [Configuration](../config.md),
+in the [CLI](../cli.md), [Configuration](../config.md),
 [Operations](../operations.md), and [integration contracts](../integrations/).
 
 ## Execution Path
 
-`cmd/notarius` delegates to `internal/cli`, the production composition root.
+~~~
-The CLI loads configuration, builds the production catalogs and runtime
+cmd/notarius -> internal/cli -> configuration and production composition
-collaborators, invokes `internal/framework/pipeline`, and places the logical
+             -> internal/framework/pipeline -> logical output files
-output files returned by the runner. Cache and debug collaborators are supplied
+             -> internal/cli -> durable output and optional state/debug data
-at this boundary.
+~~~
 
-Resolution produces a fixed ordered workflow of steps and globally unique,
+The CLI is the application boundary: it discovers configuration, composes
-sorted artifact lanes. Preparation constructs the complete module and validator
+production registries and runtime collaborators, invokes the framework, and
-set before the runner receives source bytes. Source parsing and chunking are
+places returned files. The framework resolves and prepares a fixed extraction
-serial. Each step then uses a bounded run-wide extraction pool followed by
+pipeline, then returns logical results without owning process behavior or
-serial per-lane merge and normalize continuations. A step barrier prevents
+physical state roots.
-later consumers from starting until all earlier lanes are terminal and their
-required normalized artifacts have crossed the typed handoff.
 
-## Application Boundary
+## Components
 
-| Package | Implemented responsibility |
+| Area | Implemented owners | Responsibility |
-| --- | --- |
-| `cmd/notarius` | Executable entry point and process exit delegation. |
-| `internal/cli` | Command parsing, config discovery, package-family registrar invocation, LLM client construction, reference materialization, state collaborator setup, durable writes, and user-facing results. |
-
-## Core Packages
-
-| Package | Implemented responsibility |
-| --- | --- |
-| `internal/core/artifacts` | Run-manifest and provenance models. |
-| `internal/core/config` | Defaults, YAML parsing, environment overrides, validation, redaction, and effective pipeline resolution. |
-| `internal/core/debugbundle` | Explicit per-run debug-bundle allocation and redacted summary writing. |
-| `internal/core/fileio` | Generic confined atomic file and JSON writes with caller-selected permissions. |
-| `internal/core/source` | Generic source documents, units, chunks, canonical references, validation, deterministic source digests, and independent metadata materialization. |
-
-## Framework Packages
-
-| Package | Implemented responsibility |
-| --- | --- |
-| `internal/framework/contracts` | Source-stage contracts plus artifact identity, schema, serialized representation, codec, validator, reference, output, and structured-completion interfaces and data types. |
-| `internal/framework/pipeline` | Module and artifact-codec registries, ordered-step and generated-reference resolution, option validation, profile resolution, capability checks, external reference materialization, complete pipeline preparation, typed handoff, retries, orchestration, warnings, checkpoint decisions, and manifest population. |
-| `internal/framework/validate` | Shared validator decision and cardinality helpers. |
-| `internal/framework/llm` | Scriptorium-backed structured completions, prompt/schema registration, scheduling, profile recording, and secret redaction. |
-| `internal/framework/promptfs` | Builds module prompt filesystems from module-owned and caller-provided shared prompt assets. |
-| `internal/framework/checkpoint` | Root-based checkpoint loading, recording, identity, and payload serialization. |
-| `internal/framework/chunkplan` | Source-addressed chunk-plan filesystem storage, envelope validation, and atomic publication. |
-| `internal/framework/debug` | Root-based framework and LLM debug recording. |
-
-Framework contracts provide typed artifact, provenance-wrapper, chunk-validator,
-serialized-validator, and
-typed-validator interfaces. The runner owns handoff provenance, validation
-sequencing, rejection handling, checkpoint and debug boundaries, and final
-manifest assembly.
-
-Artifact registries support heterogeneous typed extraction entries and
-kind-specific merger, normalizer, and validator variants. Resolution derives a
-lane's kind from its extractor, requires the matching codec, verifies exact Go
-type equality across the lane, and records schema identity in the resolved lane
-and pipeline digest. Registry entries carry separate option-validation and
-run-local construction closures. Preparation injects shared dependencies and
-constructs input, chunk, validators, ordered lanes, and output before source
-parsing. Production modules use strict construction-time option decoding, and
-LLM-backed modules retain the injected shared client. The D&D family registers
-the canonical `dnd/spell-list`, `dnd/npc-list`, and `dnd/combat-turn-list`
-codecs, typed spell, NPC, and combat extractors and normalizers, validators,
-plus kind-specific generic merge strategies; generic JSON validators use the
-serialized-validation contract. The runner executes lanes through
-private exact-type-checked closures, coordinates extract results independently
-of completion timing, and serializes artifacts only through their codec at
-checkpoint, debug, and output boundaries.
-
-## Production Extensions
-
-The canonical catalogs of user-selectable
-[module](../config.md#implemented-production-modules) and
-[validator](../config.md#implemented-production-validators) keys are in
-Configuration. The implemented module packages are:
-
-| Package | Implemented responsibility |
-| --- | --- |
-| `internal/modules/seriatim/input/transcript` | Parses the supported Seriatim transcript format into the generic source model. |
-| `internal/modules/generic/chunk/units` | Splits ordered source units by unit count and overlap. |
-| `internal/modules/dnd/chunk/scenes` | Produces contiguous D&D scene chunks from structured model output. |
-| `internal/modules/dnd` | Owns the canonical D&D spell-list, spell-cast, NPC-list, NPC, combat-turn-list, and combat-turn artifact types. |
-| `internal/modules/dnd/codec/spells` | Strictly decodes and stably encodes the durable D&D spell-list representation. |
-| `internal/modules/dnd/codec/npcs` | Strictly decodes and stably encodes the durable D&D NPC-list representation. |
-| `internal/modules/dnd/codec/combatturns` | Strictly decodes and stably encodes the durable D&D combat-turn-list representation. |
-| `internal/modules/dnd/extract/spells` | Maps private structured model output to canonical source-grounded D&D spell lists. |
-| `internal/modules/dnd/extract/npcs` | Maps private structured model output to canonical source-grounded D&D NPC lists. |
-| `internal/modules/dnd/extract/combatturns` | Maps private structured model output to source-grounded D&D combat-turn candidates and preserves chronology and invalid candidate values for validators. |
-| `internal/modules/dnd/normalize/combatturns` | Canonicalizes and orders merged combat turns, applies exact NPC identity matches, and collapses only exact valid-evidence duplicates. |
-| `internal/modules/dnd/validate/combatturns` | Provides deterministic shape, source-reference, source-relatedness, and normalized-invariant validation for the production combat chains. |
-| `internal/modules/dnd/npcs/registry` | Resolves validated normalized NPC references into immutable grounding data and exact identity lookup. |
-| `internal/modules/dnd/npcs/identity` | Owns Unicode-aware NPC identity, ID derivation, and registry collision validation. |
-| `internal/modules/dnd/spells/catalog` | Embeds and validates the versioned D&D 5e 2014 SRD catalog, composes optional overlays, and provides immutable effective lookup. |
-| `internal/modules/generic/merge/appendorder` | Combines accepted extraction results in chunk order. |
-| `internal/modules/generic/normalize/noop` | Preserves accepted merged output. |
-| `internal/modules/dnd/normalize/spells` | Canonicalizes catalog-backed spell names and exact source references, conservatively collapses duplicate casts, and reports deterministic warnings and independently scoped catalog checkpoint identity. |
-| `internal/modules/dnd/normalize/npcs` | Consolidates NPC records deterministically by canonical name, unions exact evidence, and reports bounded warnings. |
-| `internal/modules/generic/output/json` | Encodes manifests, lane payloads, warnings, and rejections as logical JSON files. |
-
-`internal/modules/dnd/shared` owns reusable D&D prompt fragments,
-reference declarations, prompt input assembly, source-unit reference helpers,
-and bounded diagnostics under `internal/modules/dnd/shared/diagnostics`.
-The shared NPC grounding fragment is mounted for D&D prompts and is owned by
-this package. Domain-neutral prompt filesystem composition lives in
-`internal/framework/promptfs`.
-
-The `dnd/npcs/registry` package owns the optional `npcs` registry boundary.
-External references are strictly decoded and identity-validated during
-preparation; generated references are decoded and identity-validated at the
-ordered step handoff. Both paths retain canonical registry JSON for provenance
-and emit a names-only projection to operation-time spell and combat prompts.
-Combat normalization uses the canonical registry for exact actor lookup. The
-framework records generated identity and bounded producer provenance, while
-the raw external reference remains independently tracked by pipeline
-provenance. An absent registry is represented only by the empty prompt value
-`{"npcs":[]}`. Spell
-and combat consumers use this shared boundary without changing their public
-module contracts.
-
-Generic validators under `internal/modules/generic/validate` provide
-unconditional test decisions, JSON syntax validation, and JSON Schema
-validation. D&D spell validators under `internal/modules/dnd/validate/spells`
-consume the canonical spell-list type directly to provide shape,
-effective-catalog, source-reference, and source-relatedness decisions.
-
-Production composition is grouped behind package-family registrars, and every
-implemented production extension uses its domain-first tree:
-
-| Package | Implemented responsibility |
-| --- | --- |
-| `internal/modules/generic/register` | Registers domain-neutral chunk, merge, normalize, output, and validator implementations. |
-| `internal/modules/seriatim/register` | Registers the Seriatim input adapter. |
-| `internal/modules/dnd/register` | Registers D&D modules, validators, default validator policy, and prompt/schema assets. |
-
-The CLI allocates the framework registries and asset registry, then invokes
-these registrars in generic, Seriatim, and D&D order.
-
-Implementation details for all production extensions are in
-[Module Internals](modules.md).
-
-## Run-State Components
-
-| Surface | Implemented owners | Internal purpose |
 | --- | --- | --- |
-| Durable output | Output module, pipeline runner, and CLI writer | Return logical consumer files and place them for a run. |
+| 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. |
-| Cache checkpoints | `internal/framework/checkpoint` and `internal/cli` | Validate and serialize reusable extract, merge, and normalize outcomes, including ordered-step scope and generated-artifact dependency decisions. |
+| Configuration | **internal/core/config** | Defaults, strict YAML parsing, environment overrides, structural validation, effective resolution, redaction, and resolved-composition summaries. |
-| Chunk-plan cache | `internal/framework/chunkplan` and `internal/cli` | Persist and select source-addressed plans before framework materialization. |
+| 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. |
-| Debug bundles | `internal/core/debugbundle`, `internal/framework/debug`, and pipeline instrumentation | Persist redacted summaries and application-owned traces. |
+| 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. |
+| Semantic reconciliation | **internal/framework/semanticreconcile** | Bounded source-backed candidate preparation, request-local handle proposals, deterministic assessment, typed plan application, and reconciliation identity metadata; see [Module Internals](modules.md#semantic-reconciliation) and [D&D Module Internals](dnd.md#semantic-registry-reconciliation). |
+| Embedded LLM content | **assets** | Read-only centralized LLM-facing content, scoped by its consuming package; see [LLM Runtime](llm.md#prompt-and-schema-assets) and [D&D Module Internals](dnd.md#prompt-construction). |
+| Runtime state | **internal/core/fileio**, **internal/core/debugbundle**, **internal/framework/checkpoint**, **internal/framework/chunkplan**, **internal/framework/chunkmap**, **internal/framework/debug** | Confined atomic files, debug bundles, checkpoint and chunk-plan state, accepted chunk maps, and pipeline-facing debug recording. |
+| Production extensions | **internal/modules/generic**, **internal/modules/seriatim**, **internal/modules/dnd** | Domain-neutral extensions, Seriatim input support, and D&D extraction families registered into the production catalog. |
 
-Physical layout, cleanup, recovery, and sensitive-data handling are defined
+Generic core and framework packages do not depend on production extensions.
-in [Operations](../operations.md). Concrete modules receive recorder
+Concrete extensions depend inward on their contracts and are registered only at
-interfaces and request data, not physical state roots.
+the CLI composition boundary.
 
 ## Focused Documentation
 
-- [Pipeline Internals](pipeline.md): resolution, execution, validation, retries,
+- [Configuration Internals](configuration.md): loading, validation, effective
-  checkpoint/debug hooks, and result assembly.
+  resolution, redaction, and resolved-composition identity.
-- [Module Internals](modules.md): production modules, validators, assets,
+- [CLI Internals](cli.md): command dispatch, production composition, run
-  registration, and the contributor recipe for adding an extension.
+  orchestration, and terminal reporting.
-- [LLM Runtime](llm.md): structured completion contracts, Scriptorium adapter,
+- [Pipeline Internals](pipeline.md): resolution, preparation, execution,
-  assets, scheduling, profile recording, and redaction.
+  validation, typed handoff, and framework state hooks.
 - [Run State Internals](state.md): output, cache, debug collaborator
   composition, and path safety.
+- [LLM Runtime](llm.md): structured completion, scheduling, prompt assets,
+  profiles, and secret handling.
+- [Module Internals](modules.md): generic extension registration, artifact
+  families, module construction, semantic reconciliation, validation, and
+  reference mechanics.
+- [D&D Module Internals](dnd.md): shared D&D extractor conventions, generated
+  reference projections, and lane-specific exceptions. Durable D&D and
+  Seriatim data shapes remain in the [integration contracts](../integrations/).
+
+Use this map to find an owner, then read the focused document and its tests
+before changing behavior.

docs/internal/pipeline.md

@@ -1,451 +1,194 @@
 # Pipeline Internals
 
-The implemented resolver and runner live in `internal/framework/pipeline`.
+This document describes the framework-owned pipeline mechanics in
-Their fixed workflow and ownership boundaries are defined by
+**internal/framework/pipeline**. [Configuration](../config.md) owns selectable
-[Architecture](../policy/architecture.md#system-shape). Configuration fields,
+profiles, bindings, and retry settings; [Operations](../operations.md) owns
-defaults, and selectable keys are defined in
+state lifecycle and recovery; and the [integration contracts](../integrations/)
-[Configuration](../config.md#pipelines).
+own durable output shapes. Concrete production extensions are covered by
+[Module Internals](modules.md).
 
-Resolution fixes the ordered steps, selected lanes, and all stage bindings;
+## Boundary
-preparation constructs every selected implementation before the runner begins
-source work. After serial input parsing and plan selection or generation, the
-runner materializes chunks and executes one step at a time. Within a step,
-extract work uses one bounded run-wide worker pool in chunk-first, lane-second
-order. Each lane's merge and normalize operations remain serial, and lanes in
-the same step may overlap once their extracts are terminal. A later step cannot
-start across its barrier until every earlier lane is terminal and each required
-generated artifact has been accepted and handed off.
 
-## Resolution
+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
+CLI owns process arguments, configuration discovery, session resolution,
+physical roots, and placement of returned output files.
 
-`internal/core/config.Config.Resolve` validates the loaded configuration,
+The framework has one fixed shape:
-selects the named profile, applies the runtime inputs supplied by the CLI, and
-calls `pipeline.ResolvePipeline`.
 
-`ResolvePipeline`:
+~~~
+input -> chunk -> extract -> merge -> normalize -> output
+~~~
 
-1. selects the explicit ordered steps, or creates the implicit `default` step
+Input and chunking are pipeline-wide. A selected artifact lane owns extract,
-   from the legacy top-level `artifacts` map;
+merge, and normalize; output aggregates the terminal lane outcomes. A pipeline
-2. selects and sorts artifact lanes within each step while enforcing global lane
+is an ordered list of steps, not an arbitrary workflow graph.
-   identity;
-3. completes omitted bindings using the documented configuration defaults;
-4. looks up each module and validator spec without constructing it;
-5. for a typed extractor, derives its artifact kind, requires the codec, and
-   selects exact-type merger, normalizer, and validator variants;
-6. checks required and provided capabilities in workflow order;
-7. resolves external and generated target-aware reference bindings and
-   validates producer order, consumer slot declarations, and artifact-kind
-   compatibility;
-8. validates each selected module and validator option set through its registry
-   entry; and
-9. calculates a digest over the resolved structure, including step order, step
-   IDs, lane membership, generated topology, producer and consumer identities,
-   typed artifact kind and schema identity, and the effective validator policy
-   in its resolved execution order.
 
-Resolution returns a `ResolvedPipeline` containing ordered steps, lanes,
+## Resolve, Materialize, Prepare
-concrete bindings, validator chains, reference targets, and the digest. It does
-not read external reference bytes or construct runtime modules. CLI lane and
-reference selector syntax is defined in the [CLI reference](../cli.md#run).
 
-The digest includes each resolved step's ID and lane membership, generated
+Resolution turns a configured pipeline profile into a **ResolvedPipeline**.
-producer/consumer topology, and each validator chain's stage, lane, owning
+It normalizes the pipeline and lane identities, applies stage defaults, selects
-module, ordered validator bindings, execution classes, targets, and artifact
+requested lanes where that is supported, resolves validator chains, checks
-kinds. Changing step order, a dependency, a default chain, or an explicit
+module capabilities and typed artifact compatibility, validates options, and
-override therefore changes pipeline identity whenever it changes effective
+assigns a deterministic resolved-composition digest. The resolved pipeline
-execution policy.
+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.
+Configuration resolution supplies the selected profile and catalog; see
+[Configuration Internals](configuration.md).
 
-## Reference Materialization
+External reference materialization happens before preparation. The materializer
+checks that each slot is declared by the selected module, resolves a file path
+relative to the correct configuration or working-directory origin, reads
+UTF-8 text, verifies media type and size limits, and retains bounded
+provenance. For a positive slot limit, it reads at most the limit plus one byte
+and rejects overflow before retaining content. A generated-artifact selector
+remains declared but has no bytes until its producing step completes.
 
-The CLI calls `MaterializeReferences` after resolution and before constructing
+Preparation is the construction boundary. It validates the resolved shape and
-the LLM client or running the pipeline. For external bindings, the materializer
+registry set, clones the resolved data, then constructs the input adapter,
-checks each binding against its resolved target declaration, reads and validates
+chunker, stage-local validators, every typed lane, and output encoder. Each
-the file, and builds both a `contracts.ReferenceSet` and provenance-only
+registered builder receives its own cloned build request immediately before its
-metadata on the corresponding `ResolvedReferenceTarget`. A structured
+module-owned code runs. Preparation also collects stable checkpoint
-generated binding is declaration-only at this point: its producer bytes do not
+fingerprints. Missing registrations, incompatible typed entries, nil
-exist until the producer lane reaches an accepted normalized result.
+implementations, and constructor failures are reported before source parsing
+or any stage operation begins.
 
-Preparation delivers the materialized external set for each target through
+An output encoder can opt into source-evidence publication through its output
-`pipeline.BuildRequest`: chunkers and chunk validators receive the chunk target;
+policy. Preparation keeps the configured lane allowlist and active lanes
-extractors and extract validators receive the lane extract target; mergers and
+separate, then verifies an exact typed evidence projector and registered codec
-merge validators receive the lane merge target; and normalizers and normalize
+for each active lane. The resulting private plan is immutable; lanes excluded
-validators receive the lane normalize target. Input and output builders receive
+by invocation filtering remain configured but do not acquire a projector for
-an empty set because those stages cannot declare references. Every builder gets
+that run.
-an isolated deep clone of its target set, so construction-time mutation cannot
-change another builder, the resolved pipeline, or later runtime requests.
 
-Prepared consumers do not need to be reconstructed when generated content is
+## Typed Lanes And References
-available. At the step boundary, the runner encodes the accepted producer value
-through its registered canonical codec, validates the generated bytes against
-each target slot's kind, schema, media type, and size, and clones one immutable
-reference item into the operation request. The item includes canonical digest,
-size, and bounded producer provenance but no filesystem URI. A handoff failure
-is a framework dependency error and prevents every consumer in that step from
-starting.
 
-The runner continues to clone the resulting set into the chunk, extract, merge,
+Each resolved lane has one artifact kind, codec, and exact Go type. The
-or normalize request that owns the target. LLM-backed extensions may convert
+framework uses private type erasure only around those typed operations; every
-those items into named prompt inputs. Reference content remains separate from
+handoff checks exact type and codec identity and reports incompatibility as an
-source evidence and source digests, whether the item came from a file or a
+error rather than panicking. Encoding through the registered codec is the
-generated handoff.
+boundary for output, checkpoints, debug records, and generated references.
 
-Binding precedence, path resolution, accepted content, and media-type behavior
+Reference targets are stage- and lane-specific. External reference bytes are
-are configuration contracts; see [Configuration](../config.md#pipelines).
+cloned into the operation request. Generated references are built at the next
-Durable provenance is defined in the
+step boundary from exactly one accepted normalized producer output. The
-[JSON output contract](../integrations/json-output.md#manifestjson), while
+framework decodes and re-encodes that output with the registered producer
-runtime sensitive-data handling belongs in [Operations](../operations.md).
+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.
 
-## Registries And Specs
+## Execution And Ordering
 
-`pipeline.Registries` holds option validators and run-local builders used during
+The runner validates its input, installs no-op state collaborators when none
-resolution and preparation.
+were supplied, and serially performs source parsing and chunk-plan selection.
-`pipeline.ModuleCatalog` exposes their specs during configuration validation and
+It transports the supplied session unchanged to prompt-facing operations and
-resolution. Separate registries exist for every stage and for validators;
+run-manifest metadata; it neither derives a session nor substitutes a parsed
-`ValidatorChainRegistry` stores production default-chain mappings. Both
+source document identifier. The public session contract is owned by the
-containers also carry an `ArtifactCodecRegistry`. Generic registration records
+[CLI reference](../cli.md#run).
-one codec per stable artifact kind, validates its schema metadata and JSON
+An accepted plan is materialized into source-addressed chunks and passes the
-Schema, retains the exact schema digest and Go type, and safely encodes or
+configured chunk validators before any lane runs. A chunk rejection is a
-decodes framework-erased values with typed errors on incompatibility.
+recorded pipeline outcome: lanes do not start, but the output stage can encode
+the terminal result.
 
-Typed extractor entries are keyed by module key and declare one artifact kind.
+For each ordered step, the runner first builds generated reference sets from
-Merger, normalizer, and typed-validator variants are keyed by module or
+the accepted normalized outputs of earlier steps. It then executes the step's
-validator key plus artifact kind. Chunk and serialized validators occupy
+lanes. Later steps do not begin until the current step is terminal and its
-separate target namespaces; serialized registrations declare whether they
+generated handoffs have succeeded.
-support chunks, artifacts, or both. Duplicate variants and exact Go-type
-mismatches are rejected deterministically.
 
-Lane-sensitive merger and normalizer spec discovery always supplies the
+Within a step, the lane engine dispatches extraction jobs in deterministic
-extractor's artifact kind, so variants under one reusable key may declare
+chunk-first, lane-second order to a bounded worker group. When all extraction
-different capabilities and reference slots. Kind-neutral registry inspection
+jobs for one lane are terminal, a bounded continuation group can run that
-selects the first registered artifact kind in sorted order.
+lane's merge and normalize work while extraction for other lanes continues.
+The framework does not create an unbounded goroutine per chunk or lane.
 
-Production composition registers the D&D spell-list codec and typed extractor,
+Completion timing does not determine public results. The coordinator restores
-matching typed merge, normalize, and semantic-validator variants, and
+lane and chunk order before merging results, and selects a framework error by
-serialized JSON validators. Every artifact lane resolves through the typed
+stable stage, lane, and chunk position. A validator rejection records a lane
-registries and a matching codec.
+outcome without cancelling unrelated work. A framework error or parent
+cancellation cancels derived work, prevents queued work from starting, waits
+for started workers, and prevents output encoding.
 
-A `ModuleSpec` declares its stage plus required and provided capabilities.
+## Validation, Retries, And Output
-Chunk, extract, merge, and normalize specs may also declare reference slots.
-Registry implementations defensively copy spec metadata, reject duplicate keys,
-and verify that a constructed implementation reports the registered key.
-Builder registrations accept `ModuleDependencies` and cloned configuration
-options through one `BuildRequest`. Builders decode those options and retain
-typed values or injected dependencies in the constructed implementation.
-Extractors declare their artifact kind, and merger, normalizer, and validator
-resolution selects the matching typed variant.
 
-A `ValidatorSpec` declares a validator key and execution class. Resolution uses
+Every chunk, extract, merge, and normalize candidate passes its resolved
-the execution class to reject incompatible profile bindings before execution.
+validator chain. Validators receive immutable canonical input appropriate to
-The current production catalog and default chain are listed only in
+their target: chunks, codec-decoded typed candidates, or serialized codec
-[Configuration](../config.md#implemented-production-validators).
+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.
 
-## Preparation And Runner Boundary
+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.
 
-`pipeline.Prepare` receives a resolved pipeline, the registries, and shared
+After terminal lane work, the runner assembles manifest provenance, normalized
-module dependencies. It constructs input; chunk and its validators; every
+artifacts, rejections, warnings, and an optional accepted chunk map. When an
-step's lane extract, merge, and normalize modules and validator chains in
+output policy selected evidence lanes, it decodes accepted serialized normalize
-resolved order; then output. It stops at the first error with pipeline, step,
+outputs through their registered codecs and invokes the prepared typed
-stage, lane, module, and validator context as applicable. It never invokes an
+projectors. Rejected or absent lanes contribute nothing. This reconstruction is
-operation method. Generated references are not available during preparation;
+also used after normalized-checkpoint reuse, so no second typed output channel
-the operation request is the handoff boundary.
+is retained. The runner passes the resulting owned artifact to the output
-
+encoder, which returns logical files and does not choose a physical directory.
-`PreparedPipeline` keeps private constructed executors and exposes cloned
+The CLI publishes those files only after the runner returns without a framework
-resolved input, chunk, lane, and output identities. Prepared components may
+error. Logical file names and schemas are defined by the [output integration
-implement `pipeline.CheckpointFingerprintProvider` to contribute explicit
+contracts](../integrations/).
-semantic identities to checkpoint reuse. Preparation trims and validates each
-non-secret name and value, prefixes it with the component's stage, lane,
-module, and validator scope, rejects duplicates, and retains the resulting
-sorted collection behind a defensive-copy accessor. Fingerprints must be
-stable and must not contain source content, credentials, local paths,
-timestamps, or other invocation-specific values.
-
-`pipeline.RunInput` carries that prepared pipeline, raw source input, run identity and timing, optional
-session and profile metadata, a chunk-plan store and mode, a checkpoint
-execution policy, and checkpoint/debug collaborators. The runner
-parses source bytes through the already constructed input adapter. Later stage
-requests receive the generic source model; extract requests receive
-chunk-scoped input material, while chunk, merge, and normalize requests retain
-access to the original source material. Input, chunk, and output operation
-requests do not carry raw module options. The chunk request also does not carry
-an LLM client; an LLM-backed chunker receives the shared client during
-preparation. Their operation requests retain run-specific source, reference,
-profile, session, metadata, and step-handoff context as applicable. A generated
-reference is cloned into each compatible consumer request and is never exposed
-as a path.
-
-Prepared lanes retain exact-type-checked erased operation closures. The runner
-uses those closures to keep each value typed through extraction, validation,
-merge, and normalization.
-
-Source validation requires every unit to carry a canonical self-reference to
-its containing document and its own unit ID. Explicit clone, checkpoint, and
-debug boundaries retain that reference, and the canonical source digest covers
-it deterministically. Chunks use the same source model and carry one canonical
-reference spanning the first selected unit through the last.
-
-`pipeline.RunOutput` carries the run manifest, accepted normalized serialized
-artifacts with lane and normalizer provenance,
-rejected results, warnings, checkpoint events, and logical files returned by the
-output encoder. The CLI owns debug-summary and durable filesystem writes after
-the runner returns.
-
-## Execution Flow
-
-The pipeline-wide coordinator owns the ordered step loop, generated-reference
-sets at each barrier, and deterministic merging of step outcomes. For one step,
-one run-local lane engine owns worker lifecycle, cancellation, dispatch,
-continuation queues, and result collection. It initializes checkpoint state in
-lane order, dispatches bounded extract work, advances terminal lanes through
-serial merge and normalize work, selects failures by stable pipeline scope, and
-merges lane-local outcomes back in resolved order. Completion timing never
-becomes public ordering.
-
-The runner:
-
-1. validates its prepared input;
-2. parses the raw input with the prepared adapter and validates the generic
-   source document;
-3. selects a stored plan or executes the configured chunker's `Plan` operation;
-4. canonicalizes and materializes the plan, then validates the resulting
-   chunks;
-5. executes each resolved step in configuration order. For one step, it
-   dispatches extract jobs in source-chunk then resolved-lane order, starts a
-   bounded lane continuation when all extracts for that lane are terminal, and
-   waits for every lane to become terminal;
-6. encodes and validates each accepted normalized producer artifact, then
-   builds the immutable generated reference sets for the next step;
-7. invokes the prepared output encoder only after every step succeeds and
-   validates its logical file results;
-8. returns the assembled manifest, outcomes, warnings, and files.
-
-Within each artifact lane, it reuses the prepared extractor, merger, normalizer,
-and validators while performing these transitions:
-
-1. extract once per accepted chunk and add runner-owned lane, source, and chunk
-   provenance;
-2. validate each extract result and omit rejected results from merge input;
-3. skip the rest of the lane when no extract result is accepted;
-4. merge accepted extract results in their existing order;
-5. validate the merge result and skip normalization on rejection;
-6. normalize the accepted merge result;
-7. validate and append the accepted normalized result.
-
-At a step barrier, a lane with no accepted normalized output is still a regular
-rejection unless a later generated binding names that lane as a required
-producer. In that case the runner raises a deterministic dependency error and
-does not start the consumer step. One accepted typed artifact may fan out to
-multiple compatible target slots. Consumers in the same step may run
-concurrently after the handoff; no work crosses the barrier early.
-
-Module-provided warnings and payload warnings are promoted only from attempts
-whose results are accepted and used.
-
-## Chunk Plans And Reuse
-
-`Chunker.Plan` returns a `source.ChunkPlan`: the canonical source digest,
-ordered unit-ID ranges, and optional plan or range annotations. The framework
-owns plan canonicalization and materialization. It creates the generic chunks
-and therefore owns their IDs, indexes, source references, JSON content, units,
-media type, and generic metadata. Plan and range annotations are independently
-owned raw JSON and become `Chunk.PlanAnnotations` and `Chunk.Annotations`.
-
-In `auto`, the runner looks up the source digest before invoking the chunker. A
-valid hit is materialized and sent through the current run's configured chunk
-validators; it does not invoke the chunk module, consume its retry budget, or
-make a chunk-stage LLM call. A missing, invalid, or unmaterializable record
-generates a candidate. `refresh` generates without lookup; `bypass` generates
-without cache access. Generated plans are published only after the full chunk
-validator chain approves them. A validator rejection is a regular rejected
-pipeline outcome and never replaces a cached plan.
-
-The store is source-addressed, not pipeline-addressed. Changes to pipeline
-configuration, requested chunker, options, references, lanes, validators, or
-LLM profile do not prevent a source-digest hit. The manifest records both the
-currently requested chunker and the effective plan producer. Cache state and
-paths are configured and operated outside the runner; see
-[Configuration](../config.md#state-surfaces) and [Operations](../operations.md).
-
-The extract job channel has the same capacity as the effective extract worker
-count, so dispatch applies backpressure. A fixed continuation executor prevents
-ready or checkpoint-reused lanes from creating one goroutine each. Workers and
-continuations publish lane-local results; the coordinator is the only writer of
-aggregate output and merges those results in resolved lane and source-chunk
-order.
-
-## Plan Canonicalization And Chunk Materialization
-
-Plan canonicalization requires canonical JSON annotations, a matching source
-digest, at least one range, existing ordered boundaries, and increasing range
-starts. Ranges may overlap or leave gaps; a chunker may impose stricter policy.
-Materialization deterministically reconstructs each range from the current
-source document, deep-clones JSON-shaped source-unit metadata, and copies
-annotations without interpreting their namespaces. Materialized chunks and
-separate materializations do not share mutable unit metadata; unsupported or
-cyclic metadata fails materialization with context.
-
-Before lane execution, generic chunk validation checks the materialized chunks'
-identities, order, source references, content, media type, units, and metadata.
-No chunk checkpoint participates in plan selection: plan storage is the only
-chunk-reuse mechanism. Extract, merge, and normalize checkpoints continue to
-use materialized chunk digests as their dependencies.
-
-## Validation And Retries
-
-Chunk, extract, merge, and normalize results pass through the resolved validator
-chain for their stage and module. Chunk validators receive canonical chunks;
-typed validators receive the domain value; and serialized validators receive
-canonical chunk JSON or artifact codec bytes. Validators execute in resolved
-order and stop at the first error or rejection. An empty chain approves the
-result.
-
-Production D&D artifact chains keep generic JSON syntax validation first, then
-run every rejecting domain validator before generic JSON Schema validation. The
-domain validator therefore owns expected semantic diagnostics; the generic
-schema validator remains the final rejecting representation backstop, before
-warning-only relatedness validation. Explicitly configured validator chains
-retain their configured order.
-
-`runWithRetry` applies the effective retry policy around module execution and
-its complete validation chain. A module or validator error becomes a framework
-error when attempts are exhausted. A rejection becomes a recorded
-`RejectedOutput` when attempts are exhausted. Cancellation stops retry
-processing immediately.
-
-Rejected output is a non-fatal pipeline outcome and does not advance. Warnings
-from discarded attempts are not promoted. Configuration owns retry counts and
-validator overrides; see [Module Bindings](../config.md#module-bindings).
 
 ## Checkpoint And Debug Hooks
 
-The runner depends on recorder and loader interfaces, using no-op
+The runner receives checkpoint and debug interfaces rather than roots. It
-implementations when collaborators are absent. Each checkpointed workflow
+records workflow transitions and reuse decisions through the supplied
-boundary records a running, succeeded, or failed transition. Reuse decisions
+collaborators, and clones reusable artifacts before they re-enter normal typed
-are consulted in workflow order and accepted payloads are cloned before
+handoff. Generated-reference dependencies participate in checkpoint decisions.
-entering the normal handoff path. Typed extract, merge, and normalize
+Selective recomputation can require a canonical accepted normalized predecessor
-checkpoints store codec bytes with artifact kind, schema ID, name, version and
+before a dependent lane starts.
-exact digest, and media type. Reuse compares that identity with the prepared
-codec and decodes through the codec; missing identity, mismatches, corrupt
-bytes, and decode failures become explicit reuse misses and execute the lane
-normally. Dependency fingerprints and debug content digests use the same stable
-codec bytes that cross those boundaries.
 
-That progressive extract, merge, and normalize reuse is the ordinary resume
+Debug recording is attempt-scoped and application-owned. A failure to persist
-path. A lane marked as a required predecessor for selective recomputation takes
+required debug data is a framework error. State roots, persistence, reason-code
-a separate accepted-output path before extract scheduling. The loader reads the
+meanings, resume, and cleanup are intentionally owned by
-existing successful normalize manifest and payload by step, lane, and
+[Run State Internals](state.md) and [Operations](../operations.md).
-normalizer, without consulting extract or merge dependencies. It requires the
-current non-empty checkpoint identity to match, so the invocation identity
-still binds the input, resolved topology and configuration, references, runtime
-overrides, profiles, and component fingerprints.
 
-The runner decodes and canonically re-encodes each reusable artifact once with
+## Invariants To Preserve
-the prepared codec, requiring exact kind, schema identity and digest, media
-type, canonical bytes, content digest, and producer provenance. A valid accepted
-producer becomes a runner-owned cloned normalized output, restores only
-normalize-checkpoint warnings, and records one `accepted_artifact_reused`
-normalize decision. It does not invoke or record extract, merge, normalize, or
-their validators. Invalid or unavailable accepted state records its decision
-and fails the producer step; the dependent step never starts and the producer
-is not implicitly rerun. If a later required lane fails during initialization,
-already hydrated terminal lanes remain in the failed output in resolved order.
 
-Generated references add downstream dependencies containing the producer's
+- The six fixed stages remain explicit; a pipeline is not a general DAG.
-artifact kind, complete schema identity, media type, canonical content digest,
+- Resolution and preparation reject statically discoverable incompatibility
-and size. Compatible accepted producer outputs may therefore feed a later step
+  before parsing or execution.
-without re-executing the producer. Forced lanes bypass accepted-output
+- Every typed lane uses one compatible artifact kind, codec, and exact Go type.
-hydration and execute normally. A missing, rejected, corrupt, incompatible, or
+- Generated references come only from one earlier accepted normalized producer
-changed producer blocks its dependent while leaving independent work eligible
+  and carry canonical identity rather than an unverified value.
-for reuse. The runner records bounded decision
+- Rejections are recorded outcomes; framework errors cancel derived work and
-categories: `reused`, `executed`, `forced_recompute`, and
+  prevent output encoding.
-`dependency_invalidated`. Operator meanings for the stable reason codes belong
+- Public ordering and selected errors are independent of goroutine completion
-to [Operations](../operations.md#resume-and-selective-recompute).
+  order.
+- Pipeline modules receive collaborators and data, never CLI streams or
+  physical output, cache, or debug roots.
 
-The CLI includes prepared-component fingerprints in the run-wide checkpoint
+## Focused Tests
-identity alongside resolved configuration, raw input, reference provenance,
-runtime overrides, and LLM-profile fingerprints. Module metadata is not used
-implicitly for cache identity: components opt in only with stable semantic
-values that can change accepted output. Adding or changing a component
-fingerprint intentionally produces a cold cache miss. Existing checkpoint
-schemas and paths remain unchanged.
 
-The CLI's `--recompute-step` policy forces the selected step and all transitive
+- **internal/framework/pipeline/profile_test.go** and
-dependents, but requires accepted normalized artifacts for every unselected
+  **typed_resolution_test.go** cover resolution, defaults, ordered steps,
-producer on which that closure depends. It changes execution policy only; it
+  compatibility, validators, references, and resolved identity.
-does not alter persistent checkpoint identity.
+- **internal/framework/pipeline/preparation_test.go** covers complete
+  construction before execution and contextual construction failures.
+- **internal/framework/pipeline/references_test.go** and **handoff_test.go**
+  cover external materialization, generated references, provenance, and typed
+  producer checks.
+- **internal/framework/pipeline/runner_concurrency_test.go** covers bounded
+  execution, ordered steps, stable error selection, rejections, and
+  cancellation.
+- **internal/framework/pipeline/runner_chunk_plan_test.go**,
+  **runner_typed_checkpoint_test.go**, and
+  **runner_accepted_checkpoint_test.go** cover state hooks and reuse behavior.
+- **internal/framework/pipeline/runner_attempt_debug_test.go** and
+  **runner_terminal_debug_test.go** cover attempt and terminal debug behavior.
 
-Debug instrumentation wraps run, stage, attempt, validator, and structured LLM
+Run **go test ./internal/framework/pipeline ./internal/cli** after changing a
-boundaries. Every executed chunk, extract, merge, and normalize attempt writes
+pipeline boundary. Use the more focused tests above while iterating.
-one terminal envelope for acceptance, validator rejection, module or validator
-error, or applicable candidate or final serialization error. The envelope
-contains its attempt-local warnings, any available candidate and rejection,
-and terminal error text; failures before a candidate exists omit that payload.
-Only LLM calls made by the module operation belong to the module attempt.
-Validator calls retain independent scopes under `validate/` and are not
-duplicated into the module envelope. A failed terminal-envelope write is a
-non-retryable framework error and is joined with any primary attempt error.
-Debug data is never used as a checkpoint source. Typed artifact debug envelopes
-are domain-neutral, redact sensitive metadata and bytes through the common
-debug policy, and record codec identity plus schema and content digests.
-
-Merge and normalize attempts serialize their in-memory candidate with the
-codec's required candidate encoder before typed validation. Serialized
-validators and attempt debug use that candidate representation, which carries
-the codec media type and schema identity but is never checkpointed or passed
-downstream. Only a validator-approved value is encoded through the strict final
-codec and made eligible for a checkpoint or stage output.
-
-Checkpoint identity, physical layout, reuse behavior, and debug artifact
-handling are operator contracts in [Operations](../operations.md). Serialization
-and recorder implementation are inventoried in
-[Internal Overview](overview.md#run-state-components).
-
-## Results And Failures
-
-The runner owns manifest assembly and handoff summaries but not the durable JSON
-schema. It records resolved module and lane provenance, validator chains,
-source/reference identities, selected LLM profiles, normalized and rejected
-summaries, status, and timing. Serialized artifact content remains outside the manifest.
-Module metadata providers may add non-secret singleton or lane-scoped metadata.
-
-Execution errors include stage, module, lane, or validator context. Once a
-manifest exists, a failing run returns it with failed status and completion
-time. Successful status reflects whether any result was rejected. The
-durable manifest and logical file schemas are defined in the
-[JSON output contract](../integrations/json-output.md).
-
-On a framework failure, the runner cancels its derived context, stops submitting
-new extract work, drains started tasks, and skips the output encoder. Parent
-cancellation takes precedence. Otherwise context-cancellation fallout is
-discarded when a substantive error exists, and the primary error is selected by
-stage, resolved lane, and source chunk rather than completion time.
-
-## Tests To Inspect
-
-- `internal/core/config/effective_config_test.go`: config-to-resolution boundary.
-- `internal/framework/pipeline/profile_test.go`: selection, defaults,
-  capabilities, validator chains, and digest behavior.
-- `internal/framework/pipeline/artifact_codec_registry_test.go`: typed codec
-  metadata, registration, erasure safety, strict decoding, and cloning.
-- `internal/framework/pipeline/typed_resolution_test.go`: heterogeneous typed
-  lane resolution and preparation, target-specific validators,
-  incompatibilities, ordering, and schema-sensitive pipeline identity.
-- `internal/framework/pipeline/runner_concurrency_test.go`: bounded dispatch and
-  continuations, reverse completion, stable errors, rejection, cancellation,
-  retries, and independent provider-call limits.
-- `internal/framework/pipeline/preparation_test.go`: option validation,
-  construction order, dependency failures, and the before-source-work boundary.
-- `internal/framework/pipeline/references_test.go`: target resolution and
-  materialization.
-- `internal/cli/run_contract_test.go`: production run transitions, retries,
-  rejections, warnings, CLI recomputation controls, debug hooks, and manifests.
-- `internal/cli/recompute_execution_contract_test.go`: filesystem-backed
-  selective recomputation and accepted-producer recovery.
-- `internal/cli/production_contract_test.go`: production composition and
-  configuration-resolution smoke coverage.
-- `internal/cli/example_contract_test.go`: maintained example resolution and
-  execution ownership.
-- `internal/modules/integration/*_test.go` and
-  `internal/modules/seriatim/input/transcript/runner_test.go`: typed runner
-  composition across concrete module families.
-- `internal/framework/checkpoint/*_test.go`: checkpoint serialization and reuse
-  collaborators.

docs/internal/state.md

@@ -2,7 +2,8 @@
 
 This document describes the implementation collaborators behind output, cache,
 and debug state. User-visible fields belong in [Configuration](../config.md),
-and layouts and lifecycle belong in [Operations](../operations.md).
+and physical layout, retention, recovery, reason codes, and cleanup belong in
+[Operations](../operations.md).
 
 ## Composition
 
@@ -12,11 +13,22 @@ constructs cache collaborators, writes logical output files, and reports paths.
 Pipeline modules receive interfaces and request data, never output, cache, or
 debug roots.
 
+The CLI creates no chunk-plan store in bypass mode. It creates a checkpoint
+recorder only when recording is enabled and a checkpoint loader only for a
+resume invocation. It allocates debug state only after a safe run identity has
+been generated and only when debug capture was requested. These choices keep
+the three state families independently composable.
+
 ## Output And Cache
 
 The pipeline runner returns logical output files. After validating every
 logical name, the CLI exclusively creates the run directory beneath the
 selected output root and performs confined, atomic file writes within it.
+The runner supplies an accepted chunk map as an optional, defensively owned
+output-request artifact. The JSON encoder alone decides whether its explicit
+option writes the map and optional index descriptor; neither the map payload
+nor its annotations are copied into the run manifest. The durable fields are
+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
@@ -27,12 +39,17 @@ codecs, loader, and recorder. The CLI constructs a recorder whenever checkpoint
 recording is enabled and constructs a loader only for a `--resume` invocation.
 Identity incorporates explicit stable semantic fingerprints collected from
 prepared modules and validators in addition to configuration, input,
-references, runtime overrides, and LLM profiles.
+references, runtime overrides, observed LLM profiles, and the LLM runtime's
+non-secret effective profile-source identity. A profile source change therefore
+causes a cold miss even when the configured profile ID remains unchanged.
 The serialized
 `workspace_schema_version` identifiers are frozen wire-compatibility fields;
 they do not describe a current public state surface.
 
 Ordered-step lane checkpoints include the step identity in their storage scope.
+Accepted step and lane identities are encoded injectively before becoming
+filesystem path components, while ordinary safe identifiers retain their
+readable paths.
 When a later lane consumes a generated artifact, its dependency fingerprints
 include the producer's artifact kind, complete schema identity, media type,
 canonical content digest, and size. Ordinary resume compares those fingerprints
@@ -63,12 +80,19 @@ names only the step, lane, and stable reason code. Decision detail is selected
 from code-owned descriptions by reason code and then UTF-8 normalized and
 bounded; callers cannot supply arbitrary diagnostic prose. Typed categories and
 codes remain intact through pipeline events and become strings only in manifest
-and debug-summary JSON. [Operations](../operations.md#resume-and-selective-recompute)
+and debug-summary JSON.
-is the canonical operator-facing reason-code reference.
+[Operations](../operations.md#checkpoint-recording-resume-and-recompute) is the
+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.
 
+The CLI constructs selective-recomputation policy from resolved generated
+artifact dependencies. It forces the selected step and transitive consumers,
+while marking unforced producers as required reusable inputs. The runner owns
+the actual hydration and rejection decisions; the [Operations guide](../operations.md#checkpoint-recording-resume-and-recompute)
+owns the operator workflow and stable reason-code meanings.
+
 ## Debug Bundles
 
 `internal/core/debugbundle` allocates an explicitly requested per-run bundle
@@ -93,12 +117,27 @@ operation writes the success report, or makes one attempt each to write the
 failure report and error log. Terminal persistence failures are reported
 separately and never replace the command's primary error.
 
+## Invariants To Preserve
+
+- Modules receive state collaborators and request data, never physical roots.
+- Output logical paths are validated before a run directory is allocated, and
+  files are atomically written within that directory.
+- Chunk-plan publication occurs only for accepted plans; bypass does not
+  construct or touch a plan store.
+- Checkpoint recording and checkpoint loading remain separate collaborators.
+- Debug state is opt-in, is not cache input, and terminal reporting does not
+  obscure the command's primary failure.
+
 ## Tests To Inspect
 
 - `internal/cli/run_contract_test.go`: command-owned state allocation,
   terminalization, and output/report boundaries.
+- `internal/cli/cache_contract_test.go`: cache-mode precedence, root selection,
+  and resume collaborator construction.
 - `internal/cli/state_hardening_test.go`: independent roots, reuse, failures,
   permissions, cleanup, and redaction.
+- `internal/cli/recompute_policy_test.go`: forced dependents and required
+  reusable predecessors for selective recomputation.
 - `internal/cli/recompute_execution_contract_test.go`: selective recomputation,
   filesystem recovery, deterministic decisions, and failed predecessor state.
 - `internal/cli/production_contract_test.go`: production composition and

docs/operations.md

@@ -1,288 +1,301 @@
 # Operations
 
-This is the canonical guide to operating Notarius filesystem state. Command
+This is the canonical guide for operating Notarius runtime state. The
-syntax is in the [CLI reference](cli.md); field definitions and precedence are
+[CLI reference](cli.md) owns command syntax and exit statuses, while
-in [Configuration](config.md).
+[Configuration](config.md) owns fields, defaults, and precedence. Maintainers
+who need implementation mechanics should read [Run State Internals](internal/state.md).
 
-## State Model
+## State Surfaces
 
-Notarius uses three independent filesystem surfaces:
+Each run can use independent roots with different retention and access-control
+needs.
 
-- output is durable user data;
+| Surface | Purpose | Created when | Retention |
-- cache is reconstructible chunk-plan and checkpoint state; and
+| --- | --- | --- | --- |
-- debug is explicitly requested inspection data.
+| Output | Durable user-facing result bundle | A pipeline completes and returns logical output files | Keep until consumers no longer need it. |
+| Chunk-plan cache | Reconstructible source-addressed plan | The configured cache mode permits cache I/O | Keep while reuse is useful. |
+| Checkpoint cache | Reconstructible execution and recovery state | Checkpoint recording is enabled | Keep only while recovery or reuse is useful. |
+| Debug bundle | Explicit diagnostic record | A run requests debug collection | Keep only under an intentional sensitive-data retention policy. |
 
-Choose separate roots and access controls for each surface. A normal run writes
+Output, cache, and debug roots are never merged or cleaned automatically. Use
-durable output, may use the chunk-plan cache, and records checkpoints when
+separate locations and permissions for operators or services that must not
-`cache.checkpoints.enabled` is true. It does not create debug state unless its
+share application data.
-invocation includes `--debug`.
 
-## Output
+## Roots And Permissions
 
-Durable logical files are written under:
+The configured output and debug directories are exact roots. An empty cache
+directory selects a per-user root:
 
-```text
+~~~
+<os.UserCacheDir>/notarius/chunk-plans
+<os.UserCacheDir>/notarius/checkpoints
+~~~
+
+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
+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.
+
+## PromptKit Profile Deployment
+
+Profile deployment has four distinct layers:
+
+| Layer | Owner | Operational role |
+| --- | --- | --- |
+| Prompts and schemas | Notarius module families | Embedded request and structured-output definitions. They are not deployment profile files. |
+| Fallback profiles | Notarius module families | Embedded application defaults, including D&D's `dnd-extraction` profile. |
+| Built-in profiles | PromptKit | Upstream catalog entries available when no higher-precedence source defines an ID. |
+| Operator profiles | Deployment filesystem | Complete environment-specific definitions selected by `promptkit.profile_file` or `promptkit.profile_dir`. |
+
+The maintained D&D pipeline uses the workload ID `dnd-extraction`. The
+embedded fallback makes that ID usable without an operator file. Production,
+development, and local deployments can each install a different complete
+definition for the same ID, retaining the pipeline while choosing their own
+model, backend, timeout, or reasoning policy. An operator definition wins over
+the fallback; it is not merged with it. The configuration field and full
+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
+can be copied as a format starting point. Validate a deployment without a
+provider call or credentials:
+
+~~~sh
+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.
+
+## Run Lifecycle
+
+Use the [run command](cli.md#run) to start a pipeline. A valid invocation loads
+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
+their exit classes are defined in the [CLI reference](cli.md#output-streams-and-exit-statuses).
+
+## Output Bundles
+
+Each successful run receives a generated safe run identifier and writes beneath:
+
+~~~
 <output-root>/<run-id>/
-```
+~~~
 
-The CLI generates one run ID in the form
+The [JSON output contract](integrations/json-output.md) owns the logical files
-`run-<started-at-unix-nanoseconds>-<32-lowercase-hex-characters>` and uses it
+and their schemas. Before creating the run directory, Notarius validates every
-for output, manifests, and any requested debug bundle. It validates every
+logical output path. It refuses an existing run directory without changing it.
-logical output name before exclusively creating the run directory. If that
+Files are written atomically; if a later write fails, the newly created partial
-directory already exists, the invocation fails without changing it.
+run directory remains for inspection and is never removed automatically.
 
-Each output file is written atomically. A later file-write failure leaves the
+Treat an output bundle as durable user data. Do not use cache-cleanup policy to
-newly allocated partial run directory in place for inspection; Notarius never
+remove it. An optional accepted chunk map is also durable output and can carry
-automatically removes output. The
+source- or model-derived annotations; its content and compatibility contract
-[JSON output contract](integrations/json-output.md) owns the logical file
+are defined in [Accepted Chunk Map](integrations/chunk-map.md). An optional
-names, schemas, and media types inside a run directory.
+[evidence context](integrations/evidence-context.md) contains source-unit text
-
+and metadata. It is not a cache or debug artifact: retain it with the output
-Remove an output run directory only after its consumer data is no longer
+bundle only for as long as consumers need it, and apply source-content access
-needed. This is data deletion, not cache cleanup.
+controls to the entire bundle. Its selected source-unit excerpt may include
-
+every source unit once when coverage is broad or its configured window is
-## Ordered D&D Workflow
+large, so do not assume a byte or token reduction or reduced sensitivity.
-
-The maintained [NPC-grounded configuration](../examples/dnd-npc-grounded.config.yml)
-contains one pipeline with two ordered steps. The first step extracts and
-normalizes NPCs. Only after that lane reaches an accepted terminal result does
-the second step begin; its generated NPC reference is supplied in memory to
-spell extraction, combat extraction, and combat normalization.
-
-```sh
-go run ./cmd/notarius run dnd-npc-grounded \
-  --config examples/dnd-npc-grounded.config.yml \
-  --input examples/seriatim-minimal-transcript.json \
-  --output-dir ./npc-grounded-output
-```
-
-The NPC artifact grounds canonical names through a names-only prompt projection, not spell or combat
-evidence. Current-transcript source ranges remain the only event evidence. The
-manifest records generated-reference identity and bounded producer provenance;
-it does not record generated payload content, and no generated content is
-exposed through a filesystem path. The same producer artifact may fan out to
-compatible consumers, while a missing or rejected producer prevents the later
-step from starting.
-
-Standalone module configurations continue to support external NPC files when a
-workflow intentionally crosses a process or session boundary. Those files are
-validated against the consumer slot and must be protected as sensitive
-campaign data. They are not part of the maintained ordered handoff workflow.
 
 ## Chunk-Plan Cache
 
-Chunk plans are stored at:
+Chunk plans live beneath the selected chunk-plan root:
 
-```text
+~~~
 <chunk-plan-root>/<source-sha256-hex>/plan.json
-```
+~~~
 
-`auto` reuses a complete valid plan or regenerates missing or invalid state.
+One validated canonical plan is active for each source digest. The plan stores
-`refresh` regenerates and atomically replaces a plan after chunk validation.
+boundaries and provenance, not a second copy of the entire source. This
-`bypass` performs no plan-cache I/O and does not resolve or create the root.
+source-addressed policy is recorded in [ADR-0005](adr/0005-cache-canonical-chunk-plans-by-source.md).
-Plan selection is source-addressed and independent of checkpoint and debug
-roots.
 
-When its directory is empty in configuration, the root is
+The configured cache mode controls one invocation:
-`<os.UserCacheDir>/notarius/chunk-plans`. A configured directory is the exact
-root; no suffix is appended. Directories and files created by the store use
-`0700` and `0600` permissions on supported Unix systems. The configured root
-is a trust boundary: do not share it among mutually untrusted users.
 
-Remove an exact digest directory or the configured root only when accepting the
+- **auto** looks for a valid active plan. Missing or invalid state causes a new
-cost of recomputing plans and any chunk-stage work. Cache publication is atomic;
+  plan to be generated; an accepted new plan is atomically published.
-there is no history, locking, garbage collection, or rollback facility.
+- **refresh** skips lookup, generates a plan with the configured chunker, and
+  atomically replaces the active plan after it is accepted.
+- **bypass** performs no chunk-plan cache I/O. It does not resolve or create a
+  chunk-plan root.
 
-For a Linux service account, provision a dedicated restrictive root such as:
+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.
 
-```yaml
+## Checkpoint Recording, Resume, And Recompute
-cache:
-  chunk_plans:
-    directory: /var/cache/notarius/chunk-plans
-```
 
-## Checkpoint Cache
+Checkpoint recording is an explicit configuration choice and is disabled by
+default. When enabled, each run records stage transitions and the state needed
+for compatible recovery. A run records checkpoints even when it does not ask
+to reuse them. Checkpoint payloads can contain source-derived and intermediate
+application data, so treat the entire root as sensitive.
 
-Checkpoint recording is controlled by `cache.checkpoints.enabled`, which
+Checkpoint loading is separate: [**--resume**](cli.md#run) asks a run to reuse
-defaults to `false`. When enabled, every run records running, succeeded, and
+compatible recorded work. A resume request fails when checkpoint recording is
-failed transitions and reusable validator-approved results. Successful,
+disabled. Without **--resume**, a recording-enabled run executes normally and
-rejected, and failed runs may therefore all leave checkpoint state. The
+does not load checkpoint state. Compatibility includes the resolved pipeline,
-`--resume` flag additionally loads compatible completed work before executing
+input, selected lanes, runtime overrides, reference provenance, LLM-profile
-missing or incompatible stages. Without `--resume`, a recording-enabled run
+provenance, the effective PromptKit profile-source fingerprint, and
-never loads checkpoints. Using `--resume` while recording is disabled is an
+prepared-component fingerprints. When a local PromptKit backend is configured,
-error.
+compatibility also includes a non-secret fingerprint of its endpoint. Changing
+profile content or the local endpoint causes a cold miss; changing only the
+local concurrency limit does not. A changed identity produces a cold miss;
+Notarius does not migrate, rewrite, or delete older checkpoint directories.
+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.
 
-Checkpoints use the selected root and the existing identity hierarchy:
+Checkpoint state is confined below an identity-specific path:
 
-```text
+~~~
-<checkpoint-root>/<pipeline-id>/<input-key>-<source-or-input-digest>/<pipeline-digest>/<identity-digest>/...
+<checkpoint-root>/<pipeline-id>/<input-key>-<source-or-input-digest-prefix>/<pipeline-digest-prefix>/<identity-digest-prefix>/
-```
+~~~
 
-The final identity digest includes stable semantic fingerprints explicitly
+### Selective Recompute
-contributed by prepared modules and validators. Adding or changing one of
-these fingerprints intentionally causes a cold cache miss; old checkpoint
-directories are left in place and are never migrated or deleted automatically.
 
-An empty configured directory selects
+[**--recompute-step**](cli.md#run) requires both **--resume** and enabled
-`<os.UserCacheDir>/notarius/checkpoints`. The root is exact when configured.
+checkpoint recording. It forces the selected ordered step and every lane that
-Created directories and files use `0700` and `0600` permissions on supported
+depends on it through generated artifact references. Unrelated lanes remain
-Unix systems.
+eligible for reuse.
 
-Checkpoint payloads can contain source text, intermediate artifacts, metadata,
+For an earlier producer required by a forced consumer, Notarius requires a
-warnings, and content digests. Treat them as sensitive derived application
+compatible accepted normalized artifact. It validates that artifact before
-data. Compatible files from a former checkpoint root remain reusable when
+hydrating it and does not silently rerun the producer. If that state is
-`cache.checkpoints.directory` names that exact existing root. They are not
+missing, rejected, corrupt, non-canonical, or incompatible, the run stops
-moved, migrated, or deleted automatically. The frozen serialized identifier
+before its dependent starts. Rerun the required producer deliberately instead
-`workspace_schema_version` remains part of checkpoint compatibility; it is not
+of copying or editing checkpoint files.
-a configuration setting.
 
-For a Linux service account, independently provision:
+## Checkpoint Decisions And Recovery
 
-```yaml
+Checkpoint events classify work as **executed**, **reused**,
-cache:
+**forced_recompute**, or **dependency_invalidated**. Their stable reason codes
-  checkpoints:
+are written to run diagnostics and provenance. Use the code, not a copied
-    enabled: true
+error message, to decide what to repair.
-    directory: /var/cache/notarius/checkpoints
-```
 
-Remove an exact checkpoint identity directory or the configured root only when
+| Reason code | Recovery meaning |
-recomputation is acceptable.
-
-### Resume And Selective Recompute
-
-`--resume` loads compatible accepted work only when checkpoint recording is
-enabled. A normal resumed run may reuse source, extract, merge, and normalize
-checkpoints independently and may recompute a stage after a cache miss.
-Generated references add a dependency fingerprint
-for the producer's artifact kind, schema identity, media type, canonical
-content digest, and size. If that fingerprint changes or the producer is
-missing, dependent checkpoints are invalidated; unrelated work remains eligible
-for reuse.
-
-`--recompute-step <step-id>` requires both `--resume` and
-`cache.checkpoints.enabled: true`. It forces the named step and all transitive
-dependents to execute, while compatible predecessors and unrelated lanes remain
-reusable. The ID may be an explicit configured step or `default` for an
-implicit single-step pipeline. It cannot be combined with `--only`, and it does
-not change the persistent identity of otherwise identical checkpoints.
-Decisions are bounded and categorized as `reused`, `executed`,
-`forced_recompute`, or `dependency_invalidated`.
-
-For an unselected producer required by a recomputed step, Notarius loads the
-accepted normalized artifact directly. Valid normalize state is sufficient even
-when that producer's extract or merge checkpoint is missing or corrupt. The
-normalize manifest must be successful and match workspace schema v3, the exact
-current invocation identity, step, lane, and normalizer; its payload digest and
-canonical codec representation must also validate. A forced producer bypasses
-this lookup and executes.
-
-If a required predecessor's accepted normalized artifact is missing, rejected,
-corrupt, non-canonical, or incompatible, the run fails before the dependent
-step starts. It does not fall back to rerunning that predecessor. The failure
-manifest retains completed upstream outcomes and dependency context but not
-generated reference content. For diagnosis, first check the producer step and
-lane in the manifest, then inspect checkpoint decision categories and reason
-codes. Rerun the producer explicitly rather than copying an artifact into the
-checkpoint root.
-
-The decision that caused a required-predecessor failure is retained before the
-run returns, and the CLI error identifies its step, lane, and reason code.
-
-Checkpoint reason codes are stable diagnostic identifiers:
-
-| Reason code | Operator meaning |
 | --- | --- |
-| `loading_disabled` | This invocation did not enable checkpoint loading. |
+| **loading_disabled** | This invocation did not permit checkpoint loading. |
-| `checkpoint_missing` | The requested checkpoint file does not exist. |
+| **checkpoint_missing**, **checkpoint_path_invalid**, **checkpoint_read_failed**, **checkpoint_decode_failed** | The stored checkpoint could not be located or read safely; normal resume work can execute again. |
-| `checkpoint_path_invalid` | The requested checkpoint location failed confinement validation. |
+| **workspace_schema_incompatible**, **identity_mismatch**, **stage_mismatch**, **step_mismatch**, **lane_mismatch**, **module_mismatch** | Stored state belongs to a different compatible scope or identity; allow a fresh run to create new state. |
-| `checkpoint_read_failed` | An existing checkpoint could not be read. |
+| **status_not_reusable** | The recorded operation did not end in reusable state. |
-| `checkpoint_decode_failed` | Checkpoint JSON could not be decoded. |
+| **dependency_mismatch** | A dependency changed; dependent work is invalidated rather than reused. |
-| `workspace_schema_incompatible` | The stored workspace schema is not supported by this build. |
+| **artifact_payload_invalid**, **artifact_digest_mismatch**, **artifact_codec_incompatible**, **artifact_not_canonical** | A stored artifact cannot safely be hydrated; rerun the producer instead of modifying the cache. |
-| `identity_mismatch` | The stored invocation identity differs from the current invocation. |
+| **checkpoint_reused** | A normal checkpoint passed compatibility checks. |
-| `stage_mismatch`, `step_mismatch`, `lane_mismatch`, `module_mismatch` | Stored scope does not match the requested pipeline scope. |
+| **accepted_artifact_reused** | A required predecessor's accepted normalized artifact was safely hydrated. |
-| `status_not_reusable` | The stored operation did not finish in a reusable status. |
+| **recompute_step** | Selective recomputation deliberately forced this work. |
-| `dependency_mismatch` | Stored dependencies differ; the category is `dependency_invalidated`. |
-| `artifact_payload_invalid` | Stored artifact payload structure or encoding is invalid. |
-| `artifact_digest_mismatch` | Stored artifact bytes do not match their recorded digest. |
-| `artifact_codec_incompatible` | Stored artifact identity is incomplete or incompatible with the codec contract. |
-| `artifact_not_canonical` | The codec can decode the artifact, but its bytes are not canonical. |
-| `checkpoint_reused` | The stored checkpoint passed validation and was reused. |
-| `accepted_artifact_reused` | A required producer's accepted normalized artifact was canonically validated and hydrated. |
-| `recompute_step` | Selective recomputation forced execution of this lane. |
 
-Decision detail is bounded explanatory text derived from the stable reason code,
+Reason detail is bounded code-owned text. It is diagnostic information, not a
-not caller-supplied prose or a data-recovery channel. It never contains
+path-discovery or data-recovery mechanism, and does not contain checkpoint,
-checkpoint paths, artifact or reference content, source content, credentials,
+source, reference, credential, or environment content.
-or environment values.
 
 ## Debug Bundles
 
-Only `notarius run --debug` enables debug collection. The selected root contains
+Only a [debug-enabled run](cli.md#run) creates a bundle:
-one retained bundle per invocation:
 
-```text
+~~~
 <debug-root>/<run-id>/
   summary/
   trace/
-```
+~~~
 
-`summary/` contains redacted invocation, effective-configuration, resolved
+The summary contains redacted invocation and resolution information plus run,
-pipeline and reference provenance, checkpoint and chunk-plan decisions, run
+warning, checkpoint, chunk-plan, and terminal reporting artifacts. The trace
-manifest, warnings, report, and any available error text. It excludes raw
+contains allowlisted application diagnostic records and can include source or
-source, references, annotations, prompts, model responses, credentials, and
+derived application data. Neither surface is a cache input. Do not treat a
-malformed cache bytes.
+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.
 
-`trace/` contains application-owned execution detail, including source and
+Notarius never creates debug state without an explicit request and never
-stage material, plans, chunks, validator attempts, prompts, model responses,
+automatically deletes a requested bundle. If allocation succeeds, the command
-timing, and serialized artifacts. It may retain application data omitted from
+reports its path on both success and later failure. A summary, trace, or
-output. Credentials, credential-shaped values, sensitive metadata, unrelated
+terminal-report persistence failure fails the command while preserving any
-environment values, and unrelated filesystem content are not captured.
+already-written diagnostic data for inspection.
-
-Bundles inherit the sensitivity of the application data they capture. Their
-additional risk comes from copying and aggregating that data, so restrict
-access, avoid shared roots between untrusted users, and define retention outside
-Notarius. Created bundle directories use `0700` and files use `0600` on
-supported Unix systems.
-
-Notarius never automatically deletes a requested bundle. If allocation
-succeeds, its path is reported on success and failure. A requested summary or
-trace write failure makes the command fail, preserving whatever bundle data was
-already written for inspection. Every allocated bundle makes one best-effort
-attempt to record a terminal `run-report.json`.
-
-## Failures And Warnings
-
-Failures before debug allocation are reported on stderr without a bundle.
-Failures after allocation report the bundle path on stderr and make independent
-attempts to write a failure `run-report.json` and `error.log`. The report retains
-the paths and pipeline outcome fields known at the failure point. If either
-terminal write fails, the original command error remains first on stderr,
-followed by the persistence error and bundle path. An output-write failure
-leaves the allocated bundle in place. A successful run with warnings exits `0`,
-reports a warning count on stderr, and records warnings in durable output and
-any requested debug summary.
 
 ## Cleanup
 
-Use exact paths for manual cleanup. Examples:
+Cleanup is manual and destructive. First inspect the exact leaf directory,
+then remove only that leaf; do not use a glob or a parent root as the target.
 
-```sh
+~~~
-rm -rf ./notarius-output/run-1721300000000000000-0123456789abcdef0123456789abcdef
+rm -rf -- /srv/notarius/output/run-1721300000000000000-0123456789abcdef0123456789abcdef
-rm -rf /var/cache/notarius/chunk-plans/0123abcd
+rm -rf -- /srv/notarius/chunk-plans/0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef
-rm -rf /var/cache/notarius/checkpoints/pipeline/input-0123/pipeline-4567/identity-89ab
+rm -rf -- /srv/notarius/checkpoints/example/seriatim-0123456789abcdef/0123456789abcdef/0123456789abcdef
-rm -rf ./notarius-debug/run-1721300000000000000-0123456789abcdef0123456789abcdef
+rm -rf -- /srv/notarius/debug/run-1721300000000000000-0123456789abcdef0123456789abcdef
-```
+~~~
 
-Avoid broad recursive cleanup against a parent root unless it is an explicit
+Deleting output permanently removes user data. Deleting chunk plans or
-operator policy. Output deletion is permanent user-data loss. Cache deletion is
+checkpoints is recoverable but may repeat expensive provider or pipeline work.
-recoverable but can repeat expensive work. Debug deletion removes troubleshooting
+Deleting a debug bundle removes troubleshooting evidence and a retained copy of
-evidence and any retained application-data copy.
+application data. Notarius has no cache garbage collector, rollback operation,
+or automatic cleanup command.
 
 ## Operational Limits
 
-Provider retries and timeouts are handled by Scriptorium according to the
+Provider execution settings and the generation timeout come from the selected
-selected execution profile. Pipeline module retry settings are defined in
+PromptKit profile. The invocation-only **--reasoning-effort** and
-[Configuration](config.md#module-bindings). Extract worker concurrency and
+**--clear-reasoning-effort** controls may replace or clear that profile setting
-actual provider-call concurrency are separate limits; their fields and
+for all LLM-backed calls in one run without changing the profile. PromptKit
-validation are defined in [Configuration](config.md#concurrency). Notarius
+v0.5.0 does not add a provider retry loop. Notarius binding retries rerun the
-writes local files only; remote storage and archive management are outside the
+complete module operation and validation chain as defined by
-implemented CLI.
+[module bindings](config.md#module-bindings-and-validators).
+
+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.
+The pinned upstream boundary and profile-format links are in
+[PromptKit Integration](integrations/pkg-promptkit.md).
+
+Concurrency has two independent layers. Notarius **total_llm** defaults to 16
+and is the application-wide provider-call limit shared by all backends,
+modules, retries, and validators. PromptKit may impose a narrower admission
+limit for the selected backend. The effective active-generation bound is the
+intersection of the Notarius limit, any PromptKit backend limit, and work made
+available by the pipeline. Built-in OpenRouter profiles use PromptKit's
+upstream backend limit; endpoint-only profiles have no PromptKit backend limit
+and remain bounded by Notarius. For the configured local backend, a zero
+**concurrency_limit** leaves only the Notarius scheduler as a call limit. A
+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.
+
+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
+maps that failure to Notarius's existing provider-neutral capacity error and
+does not retry it. The calling stage's configured retry policy applies
+normally, and the run fails if those attempts are exhausted. Caller
+cancellation remains authoritative. Configuration contracts are documented
+under [PromptKit profiles](config.md#promptkit-profiles) and
+[concurrency](config.md#concurrency-output-cache-and-debug). Extract-worker
+limits and actual provider-call limits are independent. Notarius writes local
+filesystem state only; remote storage, archival, and retention automation are
+outside the implemented CLI.
+
+Every run has an effective prompt session used for provider routing and run
+provenance. The generated default is stable for the same input module and raw
+input bytes; use [**--session-id**](cli.md#run) only when intentionally grouping
+different invocations. Both generated and explicit values can be visible to
+providers, manifests, checkpoints, and requested debug bundles. Do not put
+credentials or other secrets in an explicit session identifier; command-line
+values are not a credential mechanism.

docs/policy/architecture.md

@@ -24,6 +24,12 @@ DAGs or a general workflow language. Every stage remains explicit; general
 chunking, merging, or normalization behavior must not be hidden inside an
 extractor.
 
+A stage module is one configured implementation of one pipeline stage. An
+artifact family is the cohesive domain feature that owns an artifact across
+the explicit stages and supporting codecs, validators, prompts, identity
+rules, and reference projections. Artifact-family ownership does not combine
+stages or alter the fixed pipeline.
+
 Input and chunking are pipeline-wide. Each selected artifact lane owns its
 extract, merge, and normalize stages, and the output stage aggregates the run's
 lane outcomes.
@@ -39,11 +45,24 @@ implementations. Domain-neutral model and framework layers provide reusable
 policy, contracts, and orchestration. Concrete input, pipeline, output, and
 validation extensions depend inward on those generic layers.
 
+Semantic reconciliation is one such domain-neutral framework mechanism. It
+prepares bounded source context, invokes a shared model-judgment protocol,
+validates proposals, and applies safe plans through typed policies supplied by
+the consuming artifact family. It does not own domain identity, durable IDs,
+warning semantics, or artifact construction rules.
+
 Generic layers must not depend on production extensions. Concrete extensions
 must not compose the application or take ownership of process behavior. The
 current packages implementing these layers are inventoried in
 [Internal Overview](../internal/overview.md).
 
+The root `assets` package is a content-only dependency leaf. It may expose a
+read-only embedded filesystem, but it must contain no business logic and must
+not depend on `internal` packages or PromptKit. Consumers scope that filesystem
+to the content they own; the root package is not a behavioral registry or a
+public extension contract. The rationale and compatibility consequence are
+recorded in [ADR-0011](../adr/0011-centralize-llm-assets.md).
+
 The following dependency boundaries are mandatory:
 
 - extractors and validators do not depend on concrete input adapters;
@@ -73,12 +92,26 @@ Extract modules own artifact semantics, prompt use, response schemas, and
 domain interpretation. Domain-specific concepts remain in the relevant module,
 validator, shared domain helper, and artifact contract.
 
+Physical centralization of LLM-facing content does not transfer semantic
+ownership from those modules. Modules retain their manifests, response-schema
+identities, prompt ordering, and registration, while reading only their scoped
+content subtree. Generic framework code remains domain-neutral when it reads
+its own scoped generic assets from the shared content container.
+
 Typed artifact registrations declare one stable artifact kind and exact Go
 type from extraction through merge, normalization, and semantic validation.
 Pipeline resolution requires a compatible codec and matching kind-specific
 variants before a typed lane can be accepted. Framework-owned erasure remains
 private and must report type incompatibility as an error rather than a panic.
 
+An artifact kind may additionally provide a typed evidence projection that
+copies its direct generic source references. Preparation proves that projection
+matches the artifact codec's exact Go type before retaining it for an output
+policy. The runner reconstructs evidence only from accepted serialized
+normalized artifacts, and the output boundary owns any resulting publication.
+Generic framework code never infers evidence by inspecting domain JSON or
+depends on domain artifact types.
+
 Auxiliary references provide context or disambiguation. They are not source
 evidence and must not be converted into source references.
 
@@ -162,6 +195,17 @@ 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.
 
+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
+application identifiers. Semantic reconciliation may instead expose
+contiguous, one-based candidate handles that exist only for one request;
+deterministic code resolves them before typed application, and they never
+become durable identity. This is the approved request-local-label application
+of [ADR-0012](../adr/0012-resolve-opaque-entity-identifiers-deterministically.md)
+recorded by
+[ADR-0013](../adr/0013-use-request-local-candidate-handles-for-semantic-reconciliation.md).
+
 LLM calls and other external operations accept cancellation and respect
 timeouts. Concurrency control belongs in shared runtime plumbing rather than in
 individual modules.
@@ -169,7 +213,8 @@ individual modules.
 The application-wide LLM scheduler bounds actual provider calls independently
 of framework worker limits. Every LLM-backed module, retry, and validator uses
 the single injected scheduled client, including work performed by overlapping
-lanes.
+lanes. Provider runtime adapters may enforce a narrower backend-specific limit
+beneath this mandatory application-wide scheduler.
 
 ## Configuration And Provenance