Improve D&D registry caching and retire the completed roadmap

README.md

@@ -2,8 +2,9 @@
 
 Notarius is a Go CLI for turning source material into structured artifacts with
 configured extraction pipelines. The implemented D&D workflow reads Seriatim
-transcript JSON and can produce scene descriptions, item and currency events,
-NPC identities, combat turns, NPC interactions, and spell casts.
+transcript JSON and can produce location registries and occurrences, scene
+descriptions, item and currency events, NPC identities, combat turns, NPC
+interactions, enemy events, and spell casts.
 
 ## Quickstart
 

docs/cli.md

@@ -40,7 +40,7 @@ pipeline ID and **--input** are required.
 | **--debug-dir path** | Override the debug-bundle root. Requires **--debug**. |
 | **--only lane-a,lane-b** | Run only the selected comma-separated artifact lanes when that selection is valid for the configured pipeline. |
 | **--llm-profile id** | Highest-precedence configured profile for selected LLM-backed bindings and validators; it replaces binding and [pipeline](config.md#pipelines) defaults. |
-| **--session-id id** | Supply a non-empty prompt session identifier to LLM-backed module calls. |
+| **--session-id id** | Override the generated prompt session identifier with a non-empty value for LLM-backed module calls. |
 | **--reasoning-effort value** | Replace the selected PromptKit profile's reasoning effort for every LLM-backed call in this run. The value must be non-empty and the flag may be specified only once. |
 | **--clear-reasoning-effort** | Clear reasoning effort inherited from the selected PromptKit profile for every LLM-backed call in this run. |
 | **--reference selector=path** | Add or replace a file reference binding. Repeatable. |
@@ -57,6 +57,16 @@ Persistent reasoning settings remain a PromptKit profile concern.
 **--recompute-step** requires **--resume**; checkpoint requirements and reuse
 behavior are documented in [Operations](operations.md).
 
+Every run uses one effective prompt session. Without **--session-id**, Notarius
+generates a stable `notarius:v1:` identifier from the trimmed resolved input
+module key and the input file's exact raw bytes. The same module and bytes
+therefore produce the same identifier, regardless of pipeline, references,
+profile, retries, or run settings. An explicit non-empty value replaces that
+default. Session identifiers are visible to providers; they are non-secret
+correlation identifiers, not credential storage. See
+[Operations](operations.md#operational-limits) for privacy and workflow
+guidance.
+
 ### Reference selectors
 
 Use **--reference** only for a reference slot declared by the selected

docs/config.md

@@ -58,7 +58,7 @@ Built-in defaults are:
 
 | Field | Default |
 | --- | --- |
-| **concurrency.total_llm** | 1 |
+| **concurrency.total_llm** | 16 |
 | **concurrency.stage_workers.extract** | Effective **total_llm** |
 | **output.directory** | **./notarius-output** |
 | **cache.chunk_plans.mode** | **auto** |
@@ -370,14 +370,33 @@ selected target declares them:
 | **players** | Optional text player context. |
 | **glossary** | Optional text campaign glossary. |
 | **spell_catalog** | Optional JSON spell-catalog overlay for spell extraction and normalization. See [spell-catalog overlays](integrations/dnd-spell-catalog-overlays.md). |
-| **npcs** | Normalized NPC registry. Optional for spells and combat turns; required for NPC interactions. |
-| **scene_descriptions** | Required normalized scene-description artifact for combat-turn extraction. |
+| **locations** | Required normalized location registry for location-occurrence extraction and normalization. |
+| **npcs** | Normalized NPC registry. Optional for spells and combat turns; required for NPC interactions and enemy-event extraction and normalization. |
+| **scene_descriptions** | Required normalized scene-description artifact for combat-turn and enemy-event extraction. |
+| **combat_turns** | Required normalized combat-turn artifact for enemy-event extraction. |
+| **npc_interactions** | Required normalized NPC-interaction artifact for enemy-event extraction. |
+
+Location-occurrence and enemy-event artifact slots have the following exact
+binding contracts. Durable semantics and wire shapes remain in their
+[location-occurrence](integrations/dnd-location-occurrence-artifacts.md) and
+[enemy-event](integrations/dnd-enemy-event-artifacts.md) contracts.
+
+| Slot | Accepted artifact kind | Media type | Maximum size | Required stage |
+| --- | --- | --- | --- | --- |
+| `npcs` | `dnd/npc-list` | `application/json` | 1,048,576 bytes | extract and normalize |
+| `scene_descriptions` | `dnd/scene-description-list` | `application/json` | 1,048,576 bytes | extract only |
+| `combat_turns` | `dnd/combat-turn-list` | `application/json` | 1,048,576 bytes | extract only |
+| `npc_interactions` | `dnd/npc-interaction-list` | `application/json` | 1,048,576 bytes | extract only |
+| `locations` | `dnd/location-list` | `application/json` | 1,048,576 bytes | location-occurrence extract and normalize |
 
 Scene descriptions accept **party**, **players**, and **glossary**, but not
 **roster**. NPC interactions require **npcs** for both extraction and
 normalization. Combat turns require **scene_descriptions** for extraction; the
-normalized combat-turn module may use optional **npcs**. The complete example
-shows generated **npcs** and **scene_descriptions** bindings.
+normalized combat-turn module may use optional **npcs**. Location occurrences
+require **locations** for extraction and normalization. Enemy-event extraction
+requires all four of its JSON artifact slots; its normalizer requires **npcs**.
+The [complete example](../examples/dnd-complete.config.yml) shows the ordered
+generated bindings.
 
 ## Production Module Keys
 
@@ -385,18 +404,27 @@ shows generated **npcs** and **scene_descriptions** bindings.
 | --- | --- |
 | Input | **seriatim** |
 | Chunk | **generic**, **dnd/scenes** |
-| Extract | **dnd/spells**, **dnd/npcs**, **dnd/combat-turns**, **dnd/item-events**, **dnd/npc-interactions**, **dnd/scene-descriptions** |
+| Extract | **dnd/spells**, **dnd/npcs**, **dnd/combat-turns**, **dnd/item-events**, **dnd/npc-interactions**, **dnd/scene-descriptions**, **dnd/enemy-events**, **dnd/locations**, **dnd/location-occurrences** |
 | Merge | **appendorder** |
-| Normalize | **noop**, **dnd/spells**, **dnd/npcs**, **dnd/combat-turns**, **dnd/item-events**, **dnd/npc-interactions**, **dnd/scene-descriptions** |
+| Normalize | **noop**, **dnd/spells**, **dnd/npcs**, **dnd/combat-turns**, **dnd/item-events**, **dnd/npc-interactions**, **dnd/scene-descriptions**, **dnd/enemy-events**, **dnd/locations**, **dnd/location-occurrences** |
 | Output | **json** |
 
+`dnd/locations` extraction and normalization are `llm_backed`; location
+normalization may use the pipeline's selected LLM profile for bounded duplicate
+proposals. `dnd/location-occurrences` extraction is `llm_backed`, while its
+normalizer is `deterministic`. The complete example binds the registry in one
+step and the occurrence lane in the next.
+
 The D&D artifact contracts define each emitted schema:
 [spells](integrations/dnd-spell-artifacts.md),
 [NPCs](integrations/dnd-npc-artifacts.md),
 [NPC interactions](integrations/dnd-npc-interaction-artifacts.md),
 [combat turns](integrations/dnd-combat-turn-artifacts.md),
-[item events](integrations/dnd-item-event-artifacts.md), and
-[scene descriptions](integrations/dnd-scene-description-artifacts.md).
+[item events](integrations/dnd-item-event-artifacts.md),
+[scene descriptions](integrations/dnd-scene-description-artifacts.md), and
+[enemy events](integrations/dnd-enemy-event-artifacts.md),
+[locations](integrations/dnd-location-artifacts.md), and
+[location occurrences](integrations/dnd-location-occurrence-artifacts.md).
 
 ## Production Validator Keys And Default Chains
 
@@ -411,6 +439,9 @@ Available validator keys are:
 | Item events | **extract/dnd/item-events/shape**, **extract/dnd/item-events/source_refs**, **extract/dnd/item-events/source_relatedness**, **normalize/dnd/item-events/invariants** |
 | NPC interactions | **extract/dnd/npc-interactions/shape**, **extract/dnd/npc-interactions/registry**, **extract/dnd/npc-interactions/source_refs**, **extract/dnd/npc-interactions/source_relatedness**, **normalize/dnd/npc-interactions/invariants** |
 | Scene descriptions | **extract/dnd/scene-descriptions/shape**, **extract/dnd/scene-descriptions/source_refs**, **extract/dnd/scene-descriptions/source_relatedness**, **normalize/dnd/scene-descriptions/invariants** |
+| Enemy events | **extract/dnd/enemy-events/shape**, **extract/dnd/enemy-events/engagements**, **extract/dnd/enemy-events/source_refs**, **extract/dnd/enemy-events/source_relatedness**, **normalize/dnd/enemy-events/invariants** |
+| Locations | **extract/dnd/locations/shape**, **extract/dnd/locations/source_refs**, **extract/dnd/locations/source_relatedness**, **normalize/dnd/locations/identity** |
+| Location occurrences | **extract/dnd/location-occurrences/shape**, **extract/dnd/location-occurrences/registry**, **extract/dnd/location-occurrences/source_refs**, **extract/dnd/location-occurrences/source_relatedness**, **normalize/dnd/location-occurrences/invariants** |
 
 When no override is configured, production D&D bindings use the following
 ordered chains. Each row lists extract then normalize; spell chains are the
@@ -424,6 +455,9 @@ same at both stages.
 | Item events | generic/valid_json, extract/dnd/item-events/shape, extract/dnd/item-events/source_refs, generic/valid_json_schema, extract/dnd/item-events/source_relatedness | generic/valid_json, extract/dnd/item-events/shape, normalize/dnd/item-events/invariants, extract/dnd/item-events/source_refs, generic/valid_json_schema, extract/dnd/item-events/source_relatedness |
 | NPC interactions | generic/valid_json, extract/dnd/npc-interactions/shape, extract/dnd/npc-interactions/registry, extract/dnd/npc-interactions/source_refs, generic/valid_json_schema, extract/dnd/npc-interactions/source_relatedness | generic/valid_json, extract/dnd/npc-interactions/shape, extract/dnd/npc-interactions/registry, normalize/dnd/npc-interactions/invariants, extract/dnd/npc-interactions/source_refs, generic/valid_json_schema, extract/dnd/npc-interactions/source_relatedness |
 | Scene descriptions | generic/valid_json, extract/dnd/scene-descriptions/shape, extract/dnd/scene-descriptions/source_refs, generic/valid_json_schema, extract/dnd/scene-descriptions/source_relatedness | generic/valid_json, extract/dnd/scene-descriptions/shape, normalize/dnd/scene-descriptions/invariants, extract/dnd/scene-descriptions/source_refs, generic/valid_json_schema, extract/dnd/scene-descriptions/source_relatedness |
+| Enemy events | generic/valid_json, extract/dnd/enemy-events/shape, extract/dnd/enemy-events/engagements, extract/dnd/enemy-events/source_refs, generic/valid_json_schema, extract/dnd/enemy-events/source_relatedness | generic/valid_json, extract/dnd/enemy-events/shape, normalize/dnd/enemy-events/invariants, extract/dnd/enemy-events/source_refs, generic/valid_json_schema, extract/dnd/enemy-events/source_relatedness |
+| Locations | generic/valid_json, extract/dnd/locations/shape, extract/dnd/locations/source_refs, generic/valid_json_schema, extract/dnd/locations/source_relatedness | generic/valid_json, extract/dnd/locations/shape, normalize/dnd/locations/identity, extract/dnd/locations/source_refs, generic/valid_json_schema, extract/dnd/locations/source_relatedness |
+| Location occurrences | generic/valid_json, extract/dnd/location-occurrences/shape, extract/dnd/location-occurrences/registry, extract/dnd/location-occurrences/source_refs, generic/valid_json_schema, extract/dnd/location-occurrences/source_relatedness | generic/valid_json, extract/dnd/location-occurrences/shape, extract/dnd/location-occurrences/registry, normalize/dnd/location-occurrences/invariants, extract/dnd/location-occurrences/source_refs, generic/valid_json_schema, extract/dnd/location-occurrences/source_relatedness |
 
 Chains are only registered for the D&D extract and normalize modules shown
 above; select an explicit override when a different compatible chain is

docs/consumers/subprocess.md

@@ -27,10 +27,13 @@ notarius run pipeline-id \
 ```
 
 Use absolute paths for supplied input, configuration, output-root, and
-reference files. When a stable prompt session identifier or references are
-needed, pass the supported CLI flags. Supply credentials through Notarius's
-documented configuration and environment mechanisms, never as command-line
-arguments or generated secret-bearing configuration.
+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

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

@@ -64,6 +64,8 @@ kind, and complete valid evidence. It does not infer turns, initiative, or
 actions from registry or scene data.
 
 The [NPC-interaction artifact](dnd-npc-interaction-artifacts.md) records
-broader NPC occurrences. The [JSON output contract](json-output.md) defines
-publication, and [D&D module internals](../internal/dnd.md) describes routing
-and validation mechanics.
+broader NPC occurrences. The [enemy-event artifact](dnd-enemy-event-artifacts.md)
+uses combat turns as grounding only; turns do not establish an enemy event or
+its outcome. The [JSON output contract](json-output.md) defines publication,
+and [D&D module internals](../internal/dnd.md) describes routing and validation
+mechanics.

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-interaction 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-location-artifacts.md

@@ -0,0 +1,89 @@
+# D&D Location Artifact
+
+This contract defines the durable, source-grounded location registry produced
+by `dnd/locations`. 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-list` |
+| Schema ID | `notarius.dnd.locations` |
+| Schema name | `notarius_dnd_locations_v1` |
+| Schema version | `v1` |
+| Media type | `application/json` |
+| Identity policy | `dnd.locations.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:sha256:5c1a91f15729df0b8c257093865fdf2452b43c215375e8cf2341aa9c37bb99aa",
+      "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.locations.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, such
+as planes, regions, settlements, districts, buildings, rooms, landmarks,
+routes, and geographic features. A generic label is permitted only when it
+identifies a specific place in the transcript. 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
+`locations` reference slot. Its prompt receives an ordered source-free `{id,
+name}` projection and must not treat registry references 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-location-occurrence-artifacts.md

@@ -0,0 +1,84 @@
+# 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-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:sha256:5c1a91f15729df0b8c257093865fdf2452b43c215375e8cf2341aa9c37bb99aa",
+      "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 `locations` reference of
+kind `dnd/location-list`, media type `application/json`, and at most 1 MiB. The
+registry provides identity grounding only: unknown IDs and mismatched ID/name
+pairs 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-npc-artifacts.md

@@ -55,15 +55,26 @@ 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. Groups, generic roles, invented labels, and descriptive
 enrichment are excluded. Its source references prove registry provenance; they
-do not become evidence for a spell, interaction, or combat occurrence.
+do not become evidence for a spell, interaction, combat, or enemy-event
+occurrence.
 
 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 interaction](dnd-npc-interaction-artifacts.md).
+The [enemy-event artifact](dnd-enemy-event-artifacts.md) also uses it only for
+subject grounding and canonical display names.
 The [JSON output contract](json-output.md) defines publication, and
 [D&D module internals](../internal/dnd.md) owns pipeline mechanics.

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

@@ -74,5 +74,8 @@ Only entries with the same 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 and the [JSON output contract](json-output.md) for publication.
-Pipeline mechanics are described in [D&D module internals](../internal/dnd.md).
+occurrences. The [enemy-event artifact](dnd-enemy-event-artifacts.md) consumes
+only `combat_opponent` interactions 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-scene-description-artifacts.md

@@ -62,8 +62,9 @@ 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) uses an exact matching
+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 combat evidence. Publication is defined by the
+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/json-output.md

@@ -74,8 +74,11 @@ than infer a lane schema from its name. The current D&D payload contracts are
 [spells](dnd-spell-artifacts.md), [NPCs](dnd-npc-artifacts.md),
 [NPC interactions](dnd-npc-interaction-artifacts.md),
 [combat turns](dnd-combat-turn-artifacts.md),
-[item events](dnd-item-event-artifacts.md), and
-[scene descriptions](dnd-scene-description-artifacts.md).
+[item events](dnd-item-event-artifacts.md),
+[scene descriptions](dnd-scene-description-artifacts.md), and
+[enemy events](dnd-enemy-event-artifacts.md),
+[locations](dnd-location-artifacts.md), and
+[location occurrences](dnd-location-occurrence-artifacts.md).
 
 ## `manifest.json`
 
@@ -98,6 +101,11 @@ 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.
 
+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).
+
 Each `llm_profiles` entry identifies effective, non-secret LLM execution
 provenance:
 

docs/integrations/pkg-promptkit.md

@@ -48,11 +48,13 @@ 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 its trimmed run session through PromptKit's direct session
+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. Session IDs are stable, non-secret correlation identifiers and
-may be exposed to providers and provider observability.
+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

docs/internal/cli.md

@@ -89,9 +89,11 @@ handoff:
    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, prepared modules, and the
-   requested cache/checkpoint collaborators;
-6. read the source input and invoke the framework runner; and
+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.
 
@@ -102,6 +104,13 @@ 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

docs/internal/dnd.md

@@ -7,7 +7,7 @@ selectable keys, bindings, reference syntax, and default validator chains.
 
 ## Durable Artifact Contracts
 
-The six lanes have separate durable wire contracts. This guide deliberately
+The nine lanes have separate durable wire contracts. This guide deliberately
 does not repeat their JSON shapes or schemas.
 
 | Lane | Durable contract |
@@ -18,6 +18,9 @@ does not repeat their JSON shapes or schemas.
 | Item events | [item-event artifacts](../integrations/dnd-item-event-artifacts.md) |
 | NPC interactions | [NPC-interaction artifacts](../integrations/dnd-npc-interaction-artifacts.md) |
 | Scene descriptions | [scene-description artifacts](../integrations/dnd-scene-description-artifacts.md) |
+| Enemy events | [enemy-event artifacts](../integrations/dnd-enemy-event-artifacts.md) |
+| Locations | [location artifacts](../integrations/dnd-location-artifacts.md) |
+| Location occurrences | [location-occurrence artifacts](../integrations/dnd-location-occurrence-artifacts.md) |
 
 ## Family Composition
 
@@ -25,9 +28,9 @@ 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 normalization are
-registered as `llm_backed`; the remaining current D&D mergers and normalizers
-are `deterministic`. The metadata is available to catalog inspection and
+typed builder. Scene chunking, every extractor, NPC normalization, and location
+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.
@@ -41,10 +44,10 @@ the contracts above define durable data.
 ## Prompt Construction
 
 D&D extractors assemble prompts from an ordered manifest of shared and
-module-owned assets. Reuse the shared D&D system, evidence, identity,
-reference, and transcript assets instead of copying their text into individual
-modules. A manifest’s declared sequence, including cache-control placement, is
-part of the prompt behavior.
+module-owned assets. The location extractor and occurrence extractor reuse the
+shared D&D system, evidence, identity, reference, and transcript assets instead
+of copying their text into individual modules. A manifest’s declared sequence,
+including cache-control placement, is part of the prompt behavior.
 
 Every maintained D&D LLM prompt selects `dnd-extraction` as its default
 profile. The D&D registrar embeds that fallback profile with the maintained
@@ -71,7 +74,9 @@ remains stable.
 The other D&D LLM prompts intentionally follow different patterns. Scene
 chunking has no sibling extraction lane with which to share its full transcript,
 so it renders campaign references before its task and instructions, then places
-the cacheable full transcript last. NPC normalization keeps its task and
+the cacheable full transcript last. NPC and location normalization share the
+entity-reconciliation response schema and safety boundary while retaining their
+own task and identity rules. NPC normalization keeps its task and
 cacheable instructions before the candidate collection, followed by the
 cacheable transcript windows: candidates must be available before their
 supporting evidence is evaluated, and those windows are not a cross-lane
@@ -101,14 +106,20 @@ 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 events, NPC
-interactions, and scene descriptions. They canonicalize display values and
-evidence, use source-document order for stable output, and issue bounded
-warnings for changes or collapsed duplicates. The NPC normalizer is the
-intentional exception: it first produces a deterministic candidate set, then
-uses a bounded structured-LLM proposal to reconcile identity groups. Invalid
-or unusable proposals retain the deterministic result and surface retry or
-fallback diagnostics; the model does not directly replace durable records.
+interactions, 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. The NPC and location
+normalizers are intentional exceptions: each first produces a deterministic
+candidate set, then may use a bounded structured-LLM proposal to reconcile
+identity groups. Invalid or unusable proposals retain the deterministic result
+and surface retry or fallback diagnostics; the model does not directly replace
+durable records.
 
 ## Generated References And Grounding
 
@@ -122,7 +133,13 @@ NPC registries are names-only grounding projections: they may canonicalize
 actors for spells and combat turns and are required for NPC interactions, but
 they do not supply evidence. Scene-description registries are eligibility-only
 projections: they retain the current chunk’s classification data, not scene
-prose or evidence, and exist to route combat extraction.
+prose or evidence, and exist to route combat extraction. Enemy-event extraction
+also projects combat turns to `actor` and `turn_kind` and filters NPC
+interactions to `combat_opponent` names and kinds. Location registries project
+ordered `{id, name}` pairs to location-occurrence extraction and normalization;
+exact ID/name matching keeps same-name locations distinguishable. These compact
+projections, like NPC grounding, are source-free guidance and never event
+evidence.
 
 ## Lane-Specific Rules
 
@@ -137,11 +154,17 @@ shared helper changes.
 | Item events | Uses campaign context for disambiguation but has no NPC-registry or scene-description dependency. |
 | NPC interactions | Requires the normalized NPC registry at extraction and normalization, using it for canonical actor grounding only. |
 | 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-interaction 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. |
+| Locations | Produces a source-anchored, session-scoped registry. 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
 

docs/internal/llm.md

@@ -26,9 +26,10 @@ durable schemas. Those responsibilities remain with the module and its
 
 `PromptKitClient` validates the request target and prompt identity, maps each
 named material to a PromptKit inline artifact while preserving its origin URI,
-maps the trimmed request session 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 then creates one
+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

docs/internal/pipeline.md

@@ -10,11 +10,11 @@ own durable output shapes. Concrete production extensions are covered by
 ## Boundary
 
 The pipeline framework accepts a resolved composition, registries, shared
-dependencies, input bytes, 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, physical roots, and placement of returned output
-files.
+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.
 
 The framework has one fixed shape:
 
@@ -84,6 +84,10 @@ incompatible producer prevents the consumer step from starting.
 
 The runner validates its input, installs no-op state collaborators when none
 were supplied, and serially performs source parsing and chunk-plan selection.
+It transports the supplied session unchanged to prompt-facing operations and
+run-manifest metadata; it neither derives a session nor substitutes a parsed
+source document identifier. The public session contract is owned by the
+[CLI reference](../cli.md#run).
 An accepted plan is materialized into source-addressed chunks and passes the
 configured chunk validators before any lane runs. A chunk rejection is a
 recorded pipeline outcome: lanes do not start, but the output stage can encode

docs/operations.md

@@ -266,16 +266,18 @@ 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** 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
-both limits and can therefore be lower than **total_llm**. 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.
+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
@@ -289,3 +291,11 @@ under [PromptKit profiles](config.md#promptkit-profiles) and
 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/roadmap/future.md

@@ -7,37 +7,6 @@ not as committed release dates.
 
 ## Near-Term D&D Pipeline
 
-### Combat Enemy Ledger
-
-- Add a D&D artifact that identifies enemies faced during combat and supports
-  an end-of-session encounter ledger.
-- Track each enemy's observed state using a small controlled vocabulary such as
-  `active`, `killed`, `fled`, `captured`, or `incapacitated`, while preserving
-  an explicit unresolved state when the transcript does not establish an
-  outcome.
-- Preserve the evidence for enemy participation and state changes rather than
-  inferring a terminal outcome from combat ending or an enemy disappearing
-  from the conversation.
-- Define how repeated mentions, groups of unnamed enemies, summoned or allied
-  creatures, and the same enemy appearing in multiple combats affect identity
-  and ledger entries.
-- Evaluate whether the ledger should be extracted directly, derived from
-  combat-turn artifacts, or use a sequential pipeline that consumes combat
-  turns and the normalized NPC registry as grounding references.
-
-### Location Extraction
-
-- Add a D&D artifact for locations visited by the party or otherwise mentioned
-  in the transcript.
-- Distinguish observed visits from references, plans, recalled places, and
-  uncertain or inferred locations so a mention alone is not reported as a
-  visit.
-- Preserve transcript evidence for each visit or mention and reconcile aliases,
-  nested places, and repeated appearances without collapsing distinct
-  locations that share a generic name.
-- Define how the location artifact should ground later narrative reports and
-  whether future event artifacts should retain canonical location identities.
-
 ### Evaluate Spell Extraction And Normalization
 
 - Evaluate ordinary extraction retries and the completed normalization path
@@ -53,71 +22,6 @@ not as committed release dates.
   spell, combat, interaction, and scene-description lanes after real-world use.
   Add more complex chunking only in response to demonstrated failures.
 
-## Cross-Cutting LLM Runtime
-
-### Deterministic Prompt Session Identity
-
-- Replace the source-document-ID default for prompt sessions with one
-  predictable, procedurally generated session ID for the complete
-  source-processing workload.
-- Preserve an explicit non-empty `--session-id` as the highest-precedence
-  override. Otherwise, derive the default only from the effective input module
-  identity and the exact raw input bytes.
-- Use a versioned, bounded representation such as
-  `notarius:v1:<sha256(input-module + NUL + raw-input)>`. The exact encoding
-  must fit PromptKit's session length contract and must not embed source
-  content.
-- Keep the derived session stable across runs, pipelines, selected lanes,
-  ordered steps, retries, resume, recomputation, LLM profiles, reasoning
-  overrides, and output, debug, or cache settings.
-- Do not include file-backed references, generated references, reference
-  contents, or the composition of a reference bundle in session derivation.
-  References may change between prompt calls within one pipeline without
-  changing routing affinity.
-- Resolve the authoritative session before checkpoint construction and use the
-  same value for checkpoint runtime identity, every prompt-facing module,
-  PromptKit's direct session field, the compatibility `session_id` prompt
-  variable, run-manifest metadata, and debug metadata.
-- Keep routing identity separate from cache and checkpoint content identity.
-  Exact prompt prefixes, reference contents, model settings, and other
-  generation-affecting inputs must continue to participate in their existing
-  hashes and checkpoint fingerprints even though they do not change the
-  session.
-- Treat the generated value as a provider-visible, stable pseudonymous
-  correlation identifier. Do not introduce an installation-specific HMAC or
-  secret unless a concrete multi-tenant or privacy requirement justifies
-  sacrificing deterministic identity across installations.
-
-### Raise The Default Application-Wide LLM Limit
-
-- Raise the default `concurrency.total_llm` value from 1 to 16 so ordinary
-  single-backend runs can use PromptKit's expected OpenRouter capacity and
-  lower-capacity local backends without an unnecessarily narrower Notarius
-  limit.
-- Keep the Notarius application-wide scheduler mandatory and require
-  `total_llm` to remain a positive integer. Do not make the default unlimited:
-  endpoint-only profiles, an unrestricted local backend, injected clients, and
-  aggregate work across several backends may have no narrower PromptKit limit.
-- Continue defaulting `concurrency.stage_workers.extract` to the effective
-  `total_llm`, making its default 16 as part of the same change. Preserve an
-  explicit lower extract-worker setting when an operator wants less queued or
-  concurrent extraction work.
-- Define effective provider concurrency as the intersection of the Notarius
-  application-wide limit, the selected PromptKit backend limit when present,
-  and the work made available by stage execution. A Notarius limit of 16 does
-  not narrow a backend already limited to 16, while a local backend limited to
-  4 remains bounded at 4.
-- Treat the default as an application-wide safety ceiling across profiles,
-  backends, modules, retries, and validators. A run that intentionally needs
-  the combined capacity of several backends may configure a higher
-  `total_llm` and an appropriate extract-worker count explicitly.
-- Retain the existing configuration and environment override surfaces. Update
-  canonical configuration, operations, and internal documentation together
-  when the default changes.
-- Reconsider decoupling the extract-worker default from `total_llm` only after
-  mixed-backend workloads demonstrate a need for a high global emergency
-  ceiling with a lower default work-production rate.
-
 ## Shared Normalization And Quality Work
 
 ### Generic LLM-Assisted Deduplication

docs/roadmap/implementation.md

@@ -1,586 +0,0 @@
-# PromptKit v0.5 Implementation Plan
-
-## Objective
-
-Implement the target state in
-[PromptKit v0.5 Integration And LLM Profile Policy](promptkit.md). Each numbered
-stage is intended to be one implementation prompt for a GPT-5.6-Terra coding
-agent. Complete stages in order and leave the repository buildable, tested, and
-internally coherent after every stage.
-
-Follow [Architecture](../policy/architecture.md),
-[Testing Policy](../policy/testing.md), and
-[Documentation Policy](../policy/documentation.md) throughout. Preserve
-unrelated user changes. Use `apply_patch` for source and documentation edits,
-run `gofmt` on changed Go files, and add only tests that protect the behaviors
-and risks assigned to that stage.
-
-Do not implement the separate deterministic session-ID or default-concurrency
-roadmap items as part of this plan. Do not perform paid or credentialed LLM
-calls.
-
-## Background Summary
-
-Notarius currently pins PromptKit v0.3.0, calls `Prepare` and then `Run` for one
-completion, validates profiles through a synthetic prompt, has no application
-fallback profile source, and accepts LLM profiles only at individual bindings
-or through the run-wide CLI override. PromptKit v0.5.0 is source-compatible
-with the current tree; a temporary v0.5.0 module override has already passed
-`go test ./...`.
-
-The implementation must nevertheless treat the upstream optional-parameter
-change as intentional: unset `temperature`, `max_tokens`, and `top_p` remain
-unset and are omitted from compatible provider requests. Do not restore the old
-implicit `top_p: 1` default.
-
-## Stage 1: Upgrade The PromptKit Dependency
-
-### Goal
-
-Establish a clean PromptKit v0.5.0 baseline before adopting its new APIs.
-
-### Work
-
-- Update `go.mod` and `go.sum` from PromptKit v0.3.0 to v0.5.0 and run
-  `go mod tidy`.
-- Change the PromptKit built-in profile-catalog marker in
-  `internal/framework/llm/promptkit_profile_fingerprint.go` to identify
-  v0.5.0. This deliberately invalidates LLM checkpoints tied to the prior
-  catalog identity.
-- Review PromptKit-facing compile errors or test failures against the v0.4.0
-  and v0.5.0 release guides. Do not adopt prepared execution, inspection, or
-  fallback profiles in this stage.
-- Replace the existing test assertion for one exact built-in fingerprint hash
-  with durable assertions that the fingerprint is deterministic, non-empty,
-  non-secret, and changes when a semantic profile source changes. Do not add a
-  new version-constant or exact-hash change detector.
-- Update `docs/integrations/pkg-promptkit.md` to pin and link v0.5.0 and state
-  the implemented dependency-level behavior: unset optional sampling controls
-  are provider defaults. Do not document later stages as implemented.
-- Update any other canonical text that explicitly claims the dependency is
-  v0.3.0, but defer descriptions of unimplemented v0.5 APIs.
-
-### Tests And Validation
-
-- `go test ./internal/framework/llm ./internal/cli`
-- `go test ./...`
-- `go vet ./...`
-- `go build ./cmd/notarius`
-- `rg -n 'promptkit v0\.3\.0|promptkit@v0\.3\.0|PromptKit v0\.3\.0' .`
-- `git diff --check`
-
-### Completion Criteria
-
-- The repository directly pins v0.5.0 and all default offline checks pass.
-- The profile-source fingerprint identifies the new upstream catalog without a
-  brittle literal-hash test.
-- Current documentation no longer identifies v0.3.0 as the supported version.
-
-## Stage 2: Execute One Frozen Prepared Snapshot
-
-### Goal
-
-Make Notarius debug details and generation use one exact PromptKit preparation.
-
-### Work
-
-- Refactor `PromptKitClient.CompleteStructured` to call
-  `PrepareExecution`, immediately defer `Discard`, obtain a caller-owned
-  `Details` value, and execute with `RunPrepared`.
-- Preserve the existing Notarius request mapping, cancellation precedence,
-  validation classification, raw structured bytes, response decoding,
-  profile recording, usage reporting, and credential redaction.
-- Ensure every preparation, execution, validation, empty-result, and decode
-  error retains useful Notarius prompt context without exposing prepared handle
-  state or secrets.
-- Use `errors.As` to obtain `*promptkit.CapacityError` on admission rejection.
-  Preserve `contracts.ErrLLMCapacityExceeded` as the stable classification and
-  add a nonblank backend ID only to safe application-owned diagnostic context.
-  Do not expose `promptkit.CapacityError` outside the LLM adapter.
-- Update `docs/internal/llm.md` and the implemented-mechanics portion of
-  `docs/integrations/pkg-promptkit.md` to describe the single frozen execution
-  snapshot and structured capacity adaptation.
-
-### Tests And Validation
-
-- Adapt existing PromptKit client tests to the prepared-execution path.
-- Retain or add one behavioral test proving that the debug prompt details match
-  the request actually passed to generation when a backing prompt source could
-  otherwise change between independent preparations. Test the resulting
-  snapshot consistency, not a private helper call count.
-- Retain capacity tests proving `errors.Is` reaches
-  `contracts.ErrLLMCapacityExceeded`, the selected backend can appear in safe
-  diagnostic context, and provider calls are not made after rejected
-  admission.
-- Run `go test ./internal/framework/llm` and
-  `go test -race ./internal/framework/llm`.
-- Run `go test ./...` and `git diff --check`.
-
-### Completion Criteria
-
-- `CompleteStructured` no longer calls independent `Prepare` and `Run`
-  operations for one request.
-- Debug prompt material and generation result originate from the same frozen
-  PromptKit snapshot.
-- Capacity remains a provider-neutral Notarius error classification.
-
-## Stage 3: Replace Synthetic Profile Validation With Inspection
-
-### Goal
-
-Validate profiles through PromptKit's exact profile-inspection boundary and
-centralize engine profile-source construction.
-
-### Work
-
-- Introduce a small provider-adapter-owned profile inspection or validation
-  function in `internal/framework/llm`. Its public internal signature must use
-  Notarius-owned configuration and result/error types rather than returning
-  PromptKit types to the CLI.
-- Share the code that applies `profile_dir`, `profile_file`, and registered
-  backend options between the production PromptKit engine and the inspection
-  engine. Preserve the mutual-exclusion and local-backend rules.
-- Change CLI explicit-profile preflight to use `Engine.InspectProfile` through
-  that LLM boundary.
-- Remove `profileCheckPromptID`, `profileCheckPromptFS`, the `testing/fstest`
-  production dependency, and the synthetic `Prepare` request.
-- Preserve distinct, useful errors for an absent profile, invalid profile,
-  unknown backend registration, cancellation, and invalid profile source.
-- Do not require `api_key_env` to be populated during configuration validation.
-  Inspection may report credential requirements internally, but actual
-  preparation remains responsible for credential availability before a model
-  call.
-- Update current-behavior sections in `docs/internal/cli.md` and
-  `docs/internal/llm.md`. Keep field definitions in `docs/config.md`.
-
-### Tests And Validation
-
-- Replace synthetic-prompt tests with profile inspection tests covering:
-  configured local backend success; missing local backend failure; absent
-  profile; malformed profile; and an otherwise valid profile whose credential
-  environment variable is intentionally unset.
-- Prove validation performs no provider HTTP call and remains offline.
-- Run `go test ./internal/framework/llm ./internal/cli` and `go test ./...`.
-- Run `git diff --check`.
-
-### Completion Criteria
-
-- No production synthetic profile-check prompt remains.
-- Profile validation uses the same ordinary profile source and backend
-  registrations as execution.
-- Configuration validation succeeds for structurally valid profiles without
-  reading credential values.
-
-## Stage 4: Add Application Fallback Profile Asset Plumbing
-
-### Goal
-
-Allow module families to register application-owned fallback profile YAML
-without placing domain policy in generic LLM code.
-
-### Work
-
-- Extend `internal/framework/llm.AssetRegistry` with a separate fallback
-  profile source collection, registration method, flattened filesystem, and
-  safe content digest.
-- Reuse the existing asset-source path validation and flattening behavior where
-  appropriate. Reject invalid roots, unreadable assets, and duplicate flattened
-  paths. Do not parse PromptKit profile YAML in Notarius.
-- Add `promptkit.WithFallbackProfileFS` to production engine options only when
-  at least one fallback profile source is registered.
-- Supply the identical assembled fallback source to the profile-inspection
-  engine. Adjust CLI composition so pipeline-aware profile validation can use
-  the production LLM asset registry without exposing PromptKit types.
-- Extend profile-source checkpoint identity to include the exact fallback
-  profile asset digest in addition to the PromptKit catalog marker and operator
-  source. Keep the resulting fingerprint hash-only and path/content/credential
-  free.
-- Keep operator source precedence owned by PromptKit. Do not implement profile
-  merging or duplicate PromptKit source resolution in Notarius.
-- Update `docs/internal/llm.md` only for the new implemented generic asset and
-  fingerprint mechanics. No domain fallback exists until Stage 5.
-
-### Tests And Validation
-
-- Add focused AssetRegistry tests for successful flattening, invalid roots,
-  duplicate paths, and hash changes when fallback bytes change.
-- Add adapter-level tests showing that the fallback filesystem reaches both
-  execution construction and inspection construction.
-- Extend checkpoint tests to prove fallback content changes profile-source
-  identity without exposing raw YAML or paths. Use relational comparisons, not
-  a fixed hash literal.
-- Run `go test ./internal/framework/llm ./internal/cli` and `go test ./...`.
-- Run `git diff --check`.
-
-### Completion Criteria
-
-- Generic plumbing can carry application fallback profiles while remaining
-  unaware of D&D IDs or model settings.
-- Inspection, execution, and checkpoint identity use the same fallback asset
-  source.
-
-## Stage 5: Adopt The D&D `dnd-extraction` Fallback
-
-### Goal
-
-Give the D&D module family one stable embedded workload profile that operators
-can replace.
-
-### Work
-
-- Add a D&D-owned embedded PromptKit profile asset with ID `dnd-extraction`
-  under `internal/modules/dnd`. Use the exact baseline defined in
-  `promptkit.md`: OpenRouter, `openai/gpt-5.6-luna`, no explicit reasoning
-  effort, a 240-second timeout, flex service tier, and no selected temperature,
-  token limit, or `top_p`. The omitted reasoning value intentionally allows
-  OpenAI's backend to apply its `medium` default.
-- Register the profile filesystem from the D&D registrar through the generic
-  fallback profile asset boundary. Keep D&D policy out of
-  `internal/framework/llm` and the CLI composition root.
-- Change every maintained D&D LLM prompt definition—including scene chunking,
-  all D&D extractors, and NPC normalization—from the model-named default to
-  `default_profile: dnd-extraction`.
-- Add an integration-level profile-resolution test proving that:
-  - the fallback resolves when no operator source defines the ID;
-  - a valid operator profile with the same ID wins completely; and
-  - an invalid matching operator profile fails rather than falling through.
-- Test through Notarius's assembled production assets and PromptKit boundary;
-  do not duplicate every upstream source-precedence case.
-- Update the implemented profile ownership and prompt-default behavior in
-  `docs/internal/dnd.md`, `docs/internal/llm.md`, and
-  `docs/integrations/pkg-promptkit.md`. Defer the complete operator walkthrough
-  and examples to Stage 10.
-
-### Tests And Validation
-
-- Run focused D&D prompt preparation tests and the production composition
-  tests.
-- Run `go test ./internal/modules/dnd/... ./internal/framework/llm
-  ./internal/cli`.
-- Run `go test ./...`.
-- Verify `rg -n 'default_profile: gemini-2-flash' internal/modules/dnd`
-  returns no matches.
-- Run `git diff --check`.
-
-### Completion Criteria
-
-- All maintained D&D prompts use the application-owned logical profile ID.
-- The fallback works without an operator profile and remains authoritatively
-  overridable by a matching valid operator definition.
-
-## Stage 6: Introduce Module Execution-Class Metadata
-
-### Goal
-
-Make each production module's ability to use an LLM statically discoverable
-without yet changing profile inheritance.
-
-### Work
-
-- Add `ExecutionClass contracts.ExecutionClass` to `pipeline.ModuleSpec` and
-  preserve it through normalization, cloning, catalogs, registries, JSON/debug
-  views, and lookup helpers.
-- In this transitional stage only, allow an omitted execution class to
-  normalize to deterministic so existing test-only fixtures can be migrated in
-  Stage 7 without breaking the repository midway.
-- Explicitly classify every production module:
-  - D&D scene chunking, every D&D extractor, and D&D NPC normalization as
-    `llm_backed`;
-  - all other current production input, chunk, merge, normalize, and output
-    modules as `deterministic`.
-- Update production module specification tests and production catalog tests to
-  assert the semantic class alongside stage, artifact kind, and capabilities.
-- Add catalog lookup support needed by later resolution to retrieve a selected
-  module's execution class by stage and key without constructing it.
-- Do not implement pipeline-level profile inheritance or reject deterministic
-  profiles yet.
-- Update `docs/internal/modules.md` and `docs/internal/dnd.md` to identify
-  execution class as registered module metadata, while noting only implemented
-  uses.
-
-### Tests And Validation
-
-- Run module registration/spec tests across generic, Seriatim, and D&D
-  families.
-- Run `go test ./internal/framework/pipeline ./internal/modules/...`.
-- Run `go test ./...` and `git diff --check`.
-
-### Completion Criteria
-
-- Every production module has an explicit correct execution class.
-- Catalog consumers can retrieve that class without a concrete module
-  instance.
-- Test-only omitted classes remain the only temporary compatibility behavior.
-
-## Stage 7: Enforce Execution Metadata And Remove Runtime Probing
-
-### Goal
-
-Finish the execution-class contract so missing metadata cannot cause future
-profile drift.
-
-### Work
-
-- Update every framework, CLI, and integration test module specification to
-  declare an explicit execution class appropriate to the fake behavior.
-- Change module-spec validation so an empty or unsupported execution class is a
-  registration error. Remove the transitional deterministic default from
-  Stage 6.
-- Replace the chunk runner's special `ChunkExecutionClassProvider` probe with
-  specification-derived behavior. Remove the now-redundant provider interface,
-  implementation methods, and tests when they have no remaining consumer.
-- Ensure chunk producer provenance remains unchanged: it records a non-empty
-  effective binding profile for an LLM-backed chunker, while a deterministic
-  chunker records no profile. A profile selected only through the prompt
-  default remains represented by PromptKit's actual-profile manifest rather
-  than being invented as an explicit chunk binding.
-- Review helper constructors and fixtures for opportunities to set execution
-  class once without obscuring the class under test. Do not introduce an
-  elaborate test-spec framework.
-- Update internal documentation if the removal changes any described runtime
-  mechanics.
-
-### Tests And Validation
-
-- Add or retain focused registration tests for missing and invalid execution
-  classes.
-- Retain chunk-plan provenance tests for LLM-backed and deterministic
-  chunkers.
-- Run `go test ./internal/framework/pipeline ./internal/modules/...`.
-- Run `go test ./...`, `go vet ./...`, and `git diff --check`.
-
-### Completion Criteria
-
-- No registered module specification relies on an implicit execution class.
-- Pipeline metadata, not a concrete runtime type assertion, owns module
-  execution classification.
-
-## Stage 8: Resolve Programmatic Pipeline Profile Defaults
-
-### Goal
-
-Implement profile inheritance and precedence inside the pipeline resolver
-before exposing the field through YAML configuration.
-
-### Work
-
-- Add an optional trimmed `LLMProfile` field to
-  `pipeline.PipelineProfile`. Add a non-empty runtime override field to
-  `pipeline.ResolveOptions` so all precedence decisions occur in the resolver
-  rather than through pre-resolution mutation.
-- After module selection, `--only` filtering, default validator-chain
-  selection, and validator compatibility resolution, apply effective profiles
-  to every selected input, chunk, extract, merge, normalize, output, and
-  validator binding according to the precedence in `promptkit.md`.
-- Apply profiles only when the selected module or validator execution class is
-  `llm_backed`.
-- Reject a binding-specific `llm_profile` on any deterministic module or
-  validator. Do not reject or inspect an unused pipeline default when no
-  selected LLM-backed binding consumes it.
-- Leave an LLM-backed binding empty when no CLI, binding, or pipeline profile is
-  selected so PromptKit can use the prompt's `default_profile`.
-- Store the effective values on resolved bindings before digest construction.
-  Do not add a second inheritance decision to execution.
-- Ensure semantically equivalent repeated binding profiles and one inherited
-  default produce the same resolved pipeline digest. Ensure any changed
-  effective profile changes the digest.
-- Do not modify file configuration or CLI parsing in this stage.
-
-### Tests And Validation
-
-- Add pipeline package tests for the complete precedence matrix:
-  runtime override; binding-specific exception; pipeline default; prompt
-  fallback; and deterministic bindings.
-- Cover default and explicitly configured validator chains, all relevant stage
-  categories, `--only` lane selection, unused defaults, deterministic-profile
-  rejection, and semantic digest equivalence.
-- Prefer table-driven package-level tests over assertions on private traversal
-  helpers.
-- Run `go test ./internal/framework/pipeline` and `go test ./...`.
-- Run `git diff --check`.
-
-### Completion Criteria
-
-- Programmatic pipelines resolve one canonical effective profile policy.
-- Only LLM-backed resolved bindings can contain a profile.
-- Runtime override, binding, pipeline, and prompt precedence is unambiguous and
-  digest-stable.
-
-## Stage 9: Expose Pipeline Defaults Through Configuration And CLI
-
-### Goal
-
-Make the profile-default workflow available to operators while preserving
-validation and override behavior.
-
-### Work
-
-- Add optional `pipelines.<id>.llm_profile` support to the version 4 file
-  configuration model. Use presence-aware decoding so an explicitly set blank
-  value is rejected, while omission remains valid.
-- Preserve the field through file application, configuration cloning,
-  effective configuration, and programmatic profile copies without aliasing or
-  trimming drift.
-- Remove `applyLLMProfileOverride`. Pass the CLI override through the resolver's
-  runtime-override input so deterministic bindings are never populated.
-- Update effective profile-ID collection to cover every selected LLM-backed
-  module stage and LLM-backed validator, including future LLM-backed input and
-  output modules. Do not inspect deterministic or unselected profiles.
-- Ensure `run`, `config validate --pipeline`, resume/checkpoint identity, and
-  relevant dry preflight paths all use the same resolved effective profiles.
-- Preserve `--llm-profile` as the highest-precedence non-empty run-wide
-  override and preserve binding-specific profiles as exceptions when no CLI
-  override is present.
-- Do not increment the configuration version.
-- Update current configuration and CLI contracts in `docs/config.md` and
-  `docs/cli.md` in the same stage. Link to operations for the deployment
-  workflow rather than duplicating it prematurely.
-
-### Tests And Validation
-
-- Add file-config tests for omission, trimming, explicit blank rejection,
-  unknown-key behavior, cloning, and round-trip application.
-- Add effective-config and CLI contract tests for precedence, LLM-only
-  application, inherited-profile inspection failure before factory execution,
-  `--only`, and digest changes.
-- Retain offline operation and do not require credentials for
-  `config validate --pipeline`.
-- Run `go test ./internal/core/config ./internal/framework/pipeline
-  ./internal/cli`.
-- Run `go test ./...`, `go vet ./...`, and `git diff --check`.
-
-### Completion Criteria
-
-- Operators can select `dnd-extraction` once per pipeline.
-- Configuration and CLI paths share the resolver's precedence policy.
-- Unknown effective profiles fail preflight, while deterministic and unused
-  profiles do not cause spurious inspection.
-
-## Stage 10: Complete Operator Documentation, Examples, And Decision Record
-
-### Goal
-
-Make the implemented workflow understandable, copyable, and maintainable
-without duplicating canonical facts.
-
-### Work
-
-- Create an ADR using the next sequential number for the durable decision to
-  use workload-oriented pipeline defaults with operator-overridable application
-  fallback profiles. Record context, decision, alternatives, and consequences;
-  do not turn the ADR into a field reference or implementation log.
-- Complete `docs/config.md` as the canonical owner of profile-source fields,
-  `pipelines.<id>.llm_profile`, validation, and precedence.
-- Complete `docs/operations.md` with an operator workflow that distinguishes
-  Notarius embedded prompts, Notarius fallback profiles, PromptKit built-ins,
-  and deployment filesystem profiles. Include production/development/local use
-  of the same `dnd-extraction` ID, credential handling, absolute-path guidance,
-  and the fact that current relative profile paths use the process working
-  directory rather than the configuration file's directory.
-- Complete `docs/integrations/pkg-promptkit.md` with the v0.5.0 boundary,
-  prepared execution, inspection, fallback and ordinary source precedence,
-  optional provider controls, capacity adaptation, and compatibility policy.
-- Update `docs/internal/configuration.md`, `docs/internal/pipeline.md`,
-  `docs/internal/cli.md`, `docs/internal/llm.md`, `docs/internal/modules.md`, and
-  `docs/internal/dnd.md` only for their owned implementation details. Link to
-  canonical configuration, operations, and upstream format contracts rather
-  than restating them.
-- Keep exactly the existing two D&D configuration examples. Add
-  `llm_profile: dnd-extraction` to the minimal and complete pipelines and remove
-  the now-redundant model-named binding override from the complete example.
-- Add one secret-free maintained operator profile at
-  `examples/profiles/dnd-extraction.yml`. It should be a complete valid profile
-  for the same logical ID and may mirror the embedded baseline; its purpose is
-  to demonstrate file ownership and format, not claim automatic environment
-  detection. Link it from the configuration and operations documentation.
-- If the complete example selects the external profile file, use a path that
-  is valid for the documented repository-root invocation and explicitly note
-  the working-directory rule. Keep the minimal example dependent only on the
-  embedded fallback.
-- Add or extend maintained-example validation so both configuration examples
-  and the profile YAML are checked without generation or credentials.
-- Remove the now-implemented `Pipeline-Level LLM Profile Defaults` section from
-  `docs/roadmap/future.md`. Preserve the unrelated deterministic session and
-  concurrency items.
-- Do not delete `promptkit.md` or this implementation plan during the feature
-  implementation; retire them only after post-implementation review.
-
-### Tests And Validation
-
-- Run maintained example/configuration tests and relevant CLI help/parser
-  tests.
-- Run `go test ./...`.
-- Run `rg -n 'gemini-2-flash' examples docs` and review every remaining match
-  for intentional model-policy or historical context.
-- Run `rg -n 'v0\.3\.0|profileCheckPrompt|applyLLMProfileOverride' .` and resolve
-  stale production or current-documentation matches.
-- Verify all new links and `git diff --check`.
-
-### Completion Criteria
-
-- Every current fact has one canonical documentation owner.
-- Operators can distinguish and deploy all profile layers without reading Go
-  source.
-- Both maintained configurations and the maintained external profile are valid,
-  secret-free, and tested offline.
-- Implemented profile work no longer remains in `future.md`.
-
-## Stage 11: Final Verification And Quality Review
-
-### Goal
-
-Verify the complete migration as one integrated change and correct only defects
-or omissions found during that review.
-
-### Work
-
-- Review the final diff against every acceptance criterion in `promptkit.md`.
-- Confirm provider-specific PromptKit types remain inside the LLM integration
-  boundary and D&D policy remains inside the D&D module family.
-- Confirm execution and inspection receive identical ordinary, fallback, and
-  backend configuration.
-- Confirm no paths, profile YAML, endpoints, credentials, or prepared handle
-  state leak into fingerprints or ordinary diagnostics.
-- Confirm all production module specs have explicit correct execution classes
-  and every resolved deterministic binding is profile-free.
-- Confirm prompt default, pipeline default, binding override, and CLI override
-  behavior through representative assembled configurations.
-- Review tests for redundancy and remove obsolete synthetic-prompt,
-  runtime-probe, exact-hash, or duplicated upstream-behavior tests superseded by
-  stronger contract tests.
-- Perform an optional manual D&D quality comparison if credentials and an
-  evaluation transcript are deliberately supplied. Record no private input or
-  credential material, and do not make this comparison a completion gate.
-
-### Validation Commands
-
-```sh
-gofmt -w <changed-go-files>
-go test ./...
-go test -race ./internal/framework/llm ./internal/core/config ./internal/framework/pipeline ./internal/cli
-go vet ./...
-go build ./cmd/notarius
-git diff --check
-```
-
-Also run focused stale-contract searches:
-
-```sh
-rg -n 'gitea.maximumdirect.net/eric/promptkit v0\.3\.0|PromptKit v0\.3\.0' .
-rg -n 'default_profile: gemini-2-flash|profileCheckPrompt|applyLLMProfileOverride' internal docs examples
-```
-
-Review any matches rather than deleting intentional historical references
-blindly.
-
-### Completion Criteria
-
-- All automated checks pass offline and without real credentials.
-- The implemented behavior matches `promptkit.md` with no known architecture,
-  provenance, checkpoint, profile-precedence, or documentation gap.
-- Any optional live evaluation is clearly separate from correctness testing.
-
-## Open Questions
-
-None. The roadmap decisions are sufficient to implement every stage without an
-additional product or architecture choice.

docs/roadmap/promptkit.md

@@ -1,318 +0,0 @@
-# PromptKit v0.5 Integration And LLM Profile Policy
-
-## Purpose
-
-This roadmap defines the target state for upgrading Notarius from PromptKit
-v0.3.0 to v0.5.0 and adopting the upstream runtime and profile facilities that
-directly improve Notarius. It also defines the application policy for stable,
-domain-oriented LLM profile names, operator overrides, pipeline inheritance,
-profile validation, provider defaults, checkpoint identity, and documentation.
-
-The ordered work needed to reach this state belongs in
-[the implementation plan](implementation.md). Current behavior remains defined
-by the canonical documentation outside `docs/roadmap/` until the corresponding
-work is implemented.
-
-## Background
-
-Notarius currently pins PromptKit v0.3.0. Its adapter prepares a request once
-for debug material and then independently runs the original request, causing
-PromptKit to prepare the same logical call a second time. The CLI validates an
-explicit profile by preparing a synthetic prompt. PromptKit profile selection
-can be repeated on individual module bindings or replaced for one invocation
-with `--llm-profile`, but a configured pipeline cannot yet declare one inherited
-profile policy.
-
-PromptKit v0.4.0 and v0.5.0 add the upstream boundaries needed to improve these
-areas:
-
-- [v0.4.0](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/releases/v0.4.0.md)
-  adds opaque prepared executions, exact profile and prompt inspection, and a
-  typed backend-capacity error;
-- [v0.5.0](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/releases/v0.5.0.md)
-  adds application fallback profile filesystems and stops sending unset
-  optional sampling controls as framework-selected provider values; and
-- the [v0.5.0 format contract](https://gitea.maximumdirect.net/eric/promptkit/src/tag/v0.5.0/docs/formats.md)
-  defines the resulting profile-source and execution-setting precedence.
-
-A source-compatibility test of the current Notarius repository against
-PromptKit v0.5.0 completed successfully. The work is therefore primarily an
-intentional runtime and configuration migration rather than a repair for a
-breaking Go API change.
-
-## Goals
-
-- Pin and document PromptKit v0.5.0 as Notarius's supported upstream contract.
-- Execute the exact prepared request snapshot whose safe details are recorded
-  in Notarius debug material.
-- Validate configured PromptKit profiles through the upstream inspection API
-  without synthetic prompts, provider calls, or credential-value access.
-- Give Notarius an application-owned, operator-overridable
-  `dnd-extraction` profile fallback.
-- Let a pipeline choose one default LLM profile without repeating that ID on
-  every LLM-backed binding.
-- Apply profile inheritance and run-wide overrides only where the resolved
-  module or validator can use an LLM.
-- Preserve accurate checkpoint invalidation, effective profile provenance,
-  redaction, cancellation, concurrency, and provider-neutral module contracts.
-- Provide operators with one clear deployment pattern for production,
-  development, and local profile definitions.
-
-## Target End State
-
-### PromptKit Runtime Boundary
-
-Notarius depends on PromptKit v0.5.0 and uses its public APIs rather than
-reimplementing source or execution resolution.
-
-For each structured completion, the adapter:
-
-1. builds one PromptKit run request from the provider-neutral Notarius request;
-2. calls `PrepareExecution` once;
-3. immediately arranges an idempotent `Discard` for every unexecuted handle;
-4. obtains credential-redacted `Details` for debug and response metadata; and
-5. calls `RunPrepared` so generation uses that exact frozen snapshot.
-
-The debug prompt and successful result therefore describe the same selected
-profile, rendered messages, input bytes, session, output contract, and effective
-settings even when a filesystem-backed source changes concurrently. PromptKit
-handle types remain private to `internal/framework/llm`.
-
-PromptKit admission failures continue to match Notarius's provider-neutral
-`ErrLLMCapacityExceeded` contract. When PromptKit supplies a `CapacityError`,
-the adapter obtains the normalized backend ID through `errors.As` and may add it
-to safe application-owned diagnostics without parsing upstream error wording.
-The backend ID does not become a provider-specific module contract.
-
-### Optional Provider Controls
-
-Notarius accepts PromptKit v0.5.0's new behavior for `temperature`,
-`max_tokens`, and `top_p`: an unset setting is omitted from compatible provider
-requests and the provider chooses its own default. Notarius does not restore
-PromptKit's former implicit `top_p: 1` value globally.
-
-An operator who requires a particular value specifies it in the selected
-PromptKit profile. The application fallback described below intentionally
-leaves these controls unset. A human-reviewed D&D extraction comparison should
-be performed after the upgrade, but paid or nondeterministic model output is
-not part of the default automated test suite.
-
-### Profile Inspection
-
-Pipeline-aware configuration validation uses `Engine.InspectProfile` for every
-effective explicit profile ID. It verifies that the profile exists, parses and
-validates, resolves its backend and target, and is compatible with the engine's
-registered backends. It does not create a synthetic prompt, load prompt inputs,
-contact a provider, or require credential values to exist in the validation
-process environment.
-
-Credential availability is execution-time state. PromptKit preparation still
-enforces the selected profile's credential contract before generation. This
-keeps `notarius config validate` useful in build and deployment validation
-environments where secrets are deliberately absent.
-
-PromptKit construction for inspection and execution uses one shared internal
-profile-source and backend-option path. The CLI does not expose PromptKit public
-types across the Notarius LLM boundary merely to perform inspection.
-
-`InspectPrompt` is not adopted merely because it exists. It remains available
-for a later, separately defined module-to-prompt interface preflight if a
-concrete validation requirement justifies that additional contract.
-
-### Application And Operator Profile Sources
-
-Notarius embeds one ordinary PromptKit YAML profile with the stable ID
-`dnd-extraction`. It is an application fallback registered through
-`WithFallbackProfileFS`, is owned by the D&D module family, and initially
-preserves the current effective D&D baseline:
-
-- backend: PromptKit's built-in `openrouter` backend;
-- model: `openai/gpt-5.6-luna`;
-- reasoning effort: unset, allowing OpenAI's backend to apply its default of
-  `medium`;
-- generation timeout: 240 seconds;
-- service tier: `flex`; and
-- no application-selected `temperature`, `max_tokens`, or `top_p`.
-
-All maintained D&D LLM prompt definitions use `dnd-extraction` as their
-`default_profile`. The ID communicates workload intent rather than a provider,
-model, or environment. Changing the embedded fallback is an intentional
-Notarius execution-policy change and participates in checkpoint identity.
-
-Effective profile definitions resolve in PromptKit's order:
-
-1. programmatic in-memory profiles used by tests or explicit consumers;
-2. the operator source configured by `promptkit.profile_file` or
-   `promptkit.profile_dir`;
-3. the Notarius application fallback source; and
-4. PromptKit's embedded built-in catalog.
-
-Only an absent ID falls through to the next source. A matching profile is a
-complete definition: fields are not merged with a lower-precedence definition,
-and a malformed matching operator profile fails rather than silently selecting
-the application fallback.
-
-Production, development, and local deployments should normally provide
-different complete definitions for the same `dnd-extraction` ID. An operator
-source is optional because the application fallback keeps the maintained D&D
-workflow usable, but a deployment that needs an intentional model or backend
-policy should configure its own definition.
-
-### Domain Ownership And Asset Assembly
-
-The D&D fallback profile remains under `internal/modules/dnd` and is registered
-by the D&D registrar, consistent with ADR-0004. Generic LLM plumbing knows how
-to collect and flatten application fallback profile filesystems but contains no
-D&D model or policy knowledge.
-
-The shared asset registry detects invalid roots, unreadable sources, and
-duplicate flattened paths. PromptKit remains responsible for strict profile
-YAML parsing, duplicate profile-ID detection, source precedence, and effective
-target resolution. The same assembled fallback source is supplied to runtime
-execution and CLI profile inspection.
-
-### Explicit Module Execution Metadata
-
-Every registered input, chunk, extract, merge, normalize, and output module
-declares one required execution class: `deterministic` or `llm_backed`.
-Validator registrations continue to declare the same distinction through their
-validator specifications.
-
-The registered specification is authoritative for configuration resolution.
-Current production classifications are:
-
-- the D&D scene chunker, all D&D extractors, and the D&D NPC normalizer are
-  LLM-backed;
-- the Seriatim input adapter, generic chunker, all current mergers, all other
-  current normalizers, and the JSON output encoder are deterministic; and
-- current validators retain their declared classifications.
-
-Missing or unsupported execution metadata is a registration error. Explicitly
-assigning `llm_profile` to a deterministic module or validator is a pipeline
-resolution error. The framework does not infer execution class by inspecting
-domain package names or concrete implementation types at runtime.
-
-The module specification replaces the chunk runner's special runtime
-execution-class probe. Effective resolved bindings already express the result:
-only LLM-backed bindings may retain a non-empty profile.
-
-### Pipeline-Level Profile Default
-
-Configuration version 4 gains one optional non-empty pipeline field:
-
-```yaml
-pipelines:
-  dnd-session:
-    llm_profile: dnd-extraction
-```
-
-No configuration-version increment is required because the field is additive
-and existing files remain valid. An explicitly present blank value is invalid.
-
-For every selected LLM-backed module and validator, the effective profile uses
-this precedence:
-
-1. non-empty run-wide `--llm-profile` override;
-2. binding-specific `llm_profile`;
-3. pipeline-level `llm_profile`; and
-4. the prompt definition's `default_profile`, represented by an empty effective
-   Notarius binding profile.
-
-The run-wide override and inherited pipeline default never attach to a
-deterministic binding. Binding-specific exceptions remain available when one
-operation needs a different cost, latency, quality, backend, or reasoning
-policy.
-
-Inheritance is resolved after module and validator selection, including
-`--only` lane filtering, but before effective-pipeline validation, digest
-construction, explicit-profile inspection, checkpoint construction,
-preparation, execution, or provenance capture. Only profiles used by selected
-LLM-backed bindings are inspected. An unused pipeline default in a pipeline
-with no selected LLM-backed work does not require an otherwise unused profile
-to exist.
-
-The resolved pipeline contains effective binding profiles rather than a second
-runtime inheritance mechanism. Two pipelines that differ only by spelling the
-same effective policy once as a pipeline default and once on every LLM-backed
-binding have the same semantic resolved digest. Changing an effective profile
-changes the digest and applicable checkpoint identity.
-
-### Provenance And Checkpoints
-
-The PromptKit profile-source checkpoint fingerprint covers:
-
-- the PromptKit v0.5.0 built-in profile catalog identity;
-- exact application fallback profile asset content; and
-- exact configured operator profile YAML content, when present.
-
-The existing local-backend target fingerprint remains separate and continues
-to exclude scheduling-only concurrency limits. Fingerprints contain hashes and
-stable markers, not profile contents, filesystem paths, endpoints, credentials,
-or other secrets.
-
-Changing the PromptKit version, application fallback, operator profile, or
-effective pipeline profile makes incompatible LLM checkpoints ineligible for
-reuse. The dependency upgrade is expected to invalidate checkpoints produced
-under v0.3.0.
-
-Successful run manifests continue to record only profiles actually selected by
-PromptKit, including their effective model, backend, and reasoning metadata.
-Debug output reports the same effective execution snapshot used for generation.
-
-### Operator Documentation And Examples
-
-Canonical documentation clearly distinguishes:
-
-- Notarius prompt and schema assets embedded in the application;
-- Notarius application fallback profiles embedded in the application;
-- PromptKit's own embedded built-in profiles; and
-- operator profile files on the deployment filesystem.
-
-The configuration reference owns the pipeline field, profile-source fields,
-validation rules, and precedence. Operations owns deployment layout, working
-directory behavior, credentials, and environment-specific profile management.
-The PromptKit integration document owns the pinned upstream contract and
-source-precedence boundary. Internal documents describe asset registration,
-resolution, inspection, prepared execution, fingerprinting, and tests without
-duplicating user-facing field definitions.
-
-The maintained examples continue to include only the minimal and complete D&D
-configurations. They use the stable `dnd-extraction` policy, and one maintained
-PromptKit profile file under `examples/` demonstrates an operator override.
-Examples remain secret-free and are validated without live provider calls.
-
-## Out Of Scope
-
-- Implementing the separate deterministic prompt-session identity roadmap
-  item.
-- Changing the default `concurrency.total_llm` value; PromptKit's retained
-  OpenRouter capacity of 16 remains relevant to that separate item.
-- Adding model evaluation as a deterministic or CI correctness gate.
-- Automatically selecting production, development, or local environments.
-  Deployment configuration chooses the operator profile source.
-- Profile inheritance, partial profile merging, or cross-profile aliases.
-- Exposing PromptKit types to modules, validators, durable output contracts, or
-  public configuration structures.
-- Adopting `InspectPrompt` without a separately justified prompt-interface
-  validation contract.
-
-## Acceptance Criteria
-
-- Notarius builds and its offline test suite passes with PromptKit v0.5.0.
-- Every structured completion executes the exact snapshot used for safe debug
-  prompt details.
-- Profile preflight uses profile inspection and no synthetic prompt.
-- The embedded `dnd-extraction` fallback resolves without an operator source,
-  and a matching valid operator profile replaces it completely.
-- Every production module has explicit, correct execution metadata.
-- Pipeline, binding, CLI, and prompt-default precedence behaves as defined for
-  modules and validators, while deterministic bindings remain profile-free.
-- Effective profiles participate in pipeline digests, profile inspection,
-  checkpoint identity, debug records, and run provenance at the appropriate
-  boundaries.
-- The dependency and application fallback changes invalidate incompatible old
-  checkpoints without exposing profile or credential content.
-- Canonical documentation and maintained examples accurately describe and
-  exercise the implemented operator workflow.
-- Default tests remain deterministic, offline, credential-free, and focused on
-  Notarius-owned behavior rather than duplicating PromptKit's upstream suite.

examples/dnd-complete.config.yml

@@ -36,10 +36,13 @@ pipelines:
           window_units: 3
           lanes:
             - item-events
+            - locations
+            - location-occurrences
             - npcs
             - spells
             - combat-turns
             - npc-interactions
+            - enemy-events
     steps:
       # Establish session-wide reference artifacts alongside independent item events.
       - id: describe-session
@@ -58,6 +61,14 @@ pipelines:
             normalize:
               module: dnd/npcs
               retries: 2
+          locations:
+            extract:
+              module: dnd/locations
+              retries: 2
+            merge: appendorder
+            normalize:
+              module: dnd/locations
+              retries: 2
           scene-descriptions:
             extract:
               module: dnd/scene-descriptions
@@ -65,9 +76,13 @@ pipelines:
             merge: appendorder
             normalize: dnd/scene-descriptions
       - id: extract-events
-        # Accepted NPC grounding and scene-description eligibility artifacts are
-        # supplied in memory to their compatible consumers in this step.
+        # Accepted registry artifacts and scene-description eligibility artifacts
+        # are supplied in memory to their compatible consumers in this step.
         references:
+          locations:
+            artifact:
+              step: describe-session
+              lane: locations
           npcs:
             artifact:
               step: describe-session
@@ -101,3 +116,34 @@ pipelines:
               retries: 2
             merge: appendorder
             normalize: dnd/npc-interactions
+          location-occurrences:
+            extract:
+              module: dnd/location-occurrences
+              retries: 2
+            merge: appendorder
+            normalize: dnd/location-occurrences
+      - id: track-enemies
+        references:
+          npcs:
+            artifact:
+              step: describe-session
+              lane: npcs
+          scene_descriptions:
+            artifact:
+              step: describe-session
+              lane: scene-descriptions
+          combat_turns:
+            artifact:
+              step: extract-events
+              lane: combat-turns
+          npc_interactions:
+            artifact:
+              step: extract-events
+              lane: npc-interactions
+        artifacts:
+          enemy-events:
+            extract:
+              module: dnd/enemy-events
+              retries: 2
+            merge: appendorder
+            normalize: dnd/enemy-events

internal/cli/cache_contract_test.go

@@ -249,10 +249,24 @@ func TestRunAutoReusesPlanWhenRunInputsChange(t *testing.T) {
 	}
 	harness.mu.Lock()
 	chunkCalls := harness.chunkCalls
+	sessions := append([]string(nil), harness.sessionIDs...)
 	harness.mu.Unlock()
 	if chunkCalls != 1 {
 		t.Fatalf("chunk calls across changed run inputs = %d, want 1", chunkCalls)
 	}
+	rawInput, err := os.ReadFile(roots.input)
+	if err != nil {
+		t.Fatal(err)
+	}
+	wantSessionID, err := resolvePromptSessionID("", "test/input", rawInput)
+	if err != nil {
+		t.Fatal(err)
+	}
+	for _, sessionID := range sessions {
+		if sessionID != wantSessionID {
+			t.Fatalf("session IDs across reference changes = %#v, want %q", sessions, wantSessionID)
+		}
+	}
 	assertFile(t, filepath.Join(roots.plans, strings.TrimPrefix(stateTestDigest, "sha256:"), "plan.json"))
 	assertAnyFile(t, roots.output)
 }

internal/cli/dnd_enemy_events_contract_test.go

@@ -0,0 +1,340 @@
+package cli
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"os"
+	"path/filepath"
+	"reflect"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/artifacts"
+	"gitea.maximumdirect.net/eric/notarius/internal/core/config"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/evidencecontext"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/chunk/scenes"
+	locationoccurrencecodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/locationoccurrences"
+	locationcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/locations"
+	combat "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/combatturns"
+	enemyevents "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/enemyevents"
+	itemevents "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/itemevents"
+	locationoccurrences "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/locationoccurrences"
+	locations "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/locations"
+	npcinteractions "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/npcinteractions"
+	npcs "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/npcs"
+	scenedescriptions "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/scenedescriptions"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/spells"
+	enemyeventnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/enemyevents"
+	locationnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/locations"
+	npcnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/npcs"
+)
+
+func TestProductionEnemyEventConfigurationResolvesGeneratedHandoffs(t *testing.T) {
+	components := productionTestComponents(t)
+	cfg := loadMaintainedExample(t, repositoryPath("examples", "dnd-complete.config.yml"))
+	effective, err := cfg.Resolve(resolveInputForMaintainedExample(components, "dnd-session"))
+	if err != nil {
+		t.Fatalf("Resolve() error = %v, want nil", err)
+	}
+	materialized, _, err := pipeline.MaterializeReferences(effective.ResolvedPipeline, catalogFromRegistries(components.registries), pipeline.ReferenceMaterializationOptions{
+		ConfigPath: repositoryPath("examples", "dnd-complete.config.yml"),
+		WorkingDir: repositoryPath("examples"),
+	})
+	if err != nil {
+		t.Fatalf("MaterializeReferences() error = %v, want nil", err)
+	}
+	lane := referenceContractLane(t, materialized, "enemy-events")
+	if lane.ArtifactKind != dnd.EnemyEventListKind || lane.Extract.Module != enemyevents.Key || lane.Extract.Retries != 2 || lane.Merge.Module != pipeline.DefaultMergeModule || lane.Normalize.Module != enemyeventnormalize.Key {
+		t.Fatalf("enemy event lane = %#v, want typed production composition", lane)
+	}
+	for slot, want := range map[string]struct{ step, lane string }{
+		"npcs":               {step: "describe-session", lane: "npcs"},
+		"scene_descriptions": {step: "describe-session", lane: "scene-descriptions"},
+		"combat_turns":       {step: "extract-events", lane: "combat-turns"},
+		"npc_interactions":   {step: "extract-events", lane: "npc-interactions"},
+	} {
+		binding, found := generatedReferenceBinding(lane.ExtractReferences.Bindings, slot)
+		if !found || binding.Artifact.Step != want.step || binding.Artifact.Lane != want.lane {
+			t.Fatalf("enemy event %s reference = %#v, want generated %s/%s artifact", slot, binding, want.step, want.lane)
+		}
+	}
+	if binding, found := generatedReferenceBinding(lane.NormalizeReferences.Bindings, "npcs"); !found || binding.Artifact.Step != "describe-session" || binding.Artifact.Lane != "npcs" {
+		t.Fatalf("enemy event normalizer NPC reference = %#v, want generated NPC artifact", binding)
+	}
+
+	catalog := catalogFromRegistries(components.registries)
+	extractSpec, ok := catalog.Extractors.Spec(enemyevents.Key)
+	if !ok || !reflect.DeepEqual(extractSpec.Requires, []string{"chunks", "source.transcript"}) || !reflect.DeepEqual(extractSpec.Provides, []string{"dnd.enemy_events"}) {
+		t.Fatalf("enemy event extractor spec = %#v, want source and artifact capabilities", extractSpec)
+	}
+	normalizeSpec, ok := catalog.Normalizers.SpecForArtifact(enemyeventnormalize.Key, dnd.EnemyEventListKind)
+	if !ok || !reflect.DeepEqual(normalizeSpec.Requires, []string{"merged"}) || !reflect.DeepEqual(normalizeSpec.Provides, []string{"normalized"}) {
+		t.Fatalf("enemy event normalizer spec = %#v, want merged/normalized capabilities", normalizeSpec)
+	}
+	for _, slot := range []string{"npcs", "scene_descriptions", "combat_turns", "npc_interactions"} {
+		if !hasReferenceSlot(extractSpec.ReferenceSlots, slot) {
+			t.Fatalf("enemy event extractor slots = %#v, want %q", extractSpec.ReferenceSlots, slot)
+		}
+	}
+	if !hasReferenceSlot(normalizeSpec.ReferenceSlots, "npcs") {
+		t.Fatalf("enemy event normalizer slots = %#v, want NPC registry", normalizeSpec.ReferenceSlots)
+	}
+
+	profile := cfg.Pipelines["dnd-session"]
+	profile.Steps[2].References["npcs"] = pipeline.GeneratedReference("track-enemies", "enemy-events")
+	cfg.Pipelines["dnd-session"] = profile
+	if _, err := cfg.Resolve(resolveInputForMaintainedExample(components, "dnd-session")); err == nil || !strings.Contains(err.Error(), "earlier step") {
+		t.Fatalf("Resolve() error = %v, want future generated-reference rejection", err)
+	}
+}
+
+func TestMaintainedCompleteExampleProducesEnemyEventsThroughGeneratedHandoffs(t *testing.T) {
+	t.Chdir(repositoryPath())
+	outputRoot := filepath.Join(t.TempDir(), "output")
+	configPath := completeExampleConfigWithTemporaryCache(t)
+	client := &enemyEventLLMClient{}
+	options := productionCLIOptions(t)
+	options.Now = func() time.Time { return time.Unix(1700000000, 0).UTC() }
+	options.RunIDGenerator = func(time.Time) (string, error) { return productionRunID, nil }
+	options.UserCacheDir = func() (string, error) { return "", errors.New("user cache must not be used") }
+	options.LLMClientFactory = func(context.Context, config.Config, string, LLMRuntimeOverrides) (contracts.StructuredLLMClient, []artifacts.LLMProfileManifest, error) {
+		return client, nil, nil
+	}
+
+	var stdout, stderr strings.Builder
+	code := RunWithOptions([]string{
+		"run", "dnd-session",
+		"--config", configPath,
+		"--input", repositoryPath("examples", "dnd-complete-transcript.json"),
+		"--chunk_cache", "bypass", "--output-dir", outputRoot, "--session-id", "enemy-event-session",
+	}, &stdout, &stderr, options)
+	if code != 0 {
+		t.Fatalf("code=%d stdout=%q stderr=%q", code, stdout.String(), stderr.String())
+	}
+
+	runRoot := filepath.Join(outputRoot, productionRunID)
+	index := readProductionJSON[exampleOutputIndex](t, filepath.Join(runRoot, "index.json"))
+	var enemyOutput exampleOutputIndexEntry
+	var locationOutput, occurrenceOutput exampleOutputIndexEntry
+	for _, entry := range index.OutputFiles {
+		switch entry.LaneID {
+		case "enemy-events":
+			enemyOutput = entry
+		case "locations":
+			locationOutput = entry
+		case "location-occurrences":
+			occurrenceOutput = entry
+		}
+	}
+	if enemyOutput.File != "lanes/enemy-events.json" || enemyOutput.SchemaID != "notarius.dnd.enemy_events" || enemyOutput.SchemaVersion != "v1" {
+		t.Fatalf("enemy event output = %#v, want typed enemy-event JSON", enemyOutput)
+	}
+	value := readProductionJSON[dnd.EnemyEventList](t, filepath.Join(runRoot, enemyOutput.File))
+	if len(value.Events) != 1 || value.Events[0].Name != "Kesh" || value.Events[0].Kind != dnd.EnemyEventKindFled || len(value.Events[0].SourceRefs) != 1 || value.Events[0].SourceRefs[0].SourceID != "session-ravenfall" || value.Events[0].SourceRefs[0].StartUnitID != 10 {
+		t.Fatalf("enemy event artifact = %#v, want source-linked Kesh fleeing event", value)
+	}
+	if locationOutput.File != "lanes/locations.json" || locationOutput.SchemaID != locationcodec.SchemaID || locationOutput.SchemaVersion != locationcodec.SchemaVersion {
+		t.Fatalf("location output = %#v, want typed location registry JSON", locationOutput)
+	}
+	locationsValue := readProductionJSON[dnd.LocationList](t, filepath.Join(runRoot, locationOutput.File))
+	if len(locationsValue.Locations) != 2 || locationsValue.Locations[0].Name != "Moon Gate" || locationsValue.Locations[1].Name != "Moon Gate" || locationsValue.Locations[0].ID == locationsValue.Locations[1].ID {
+		t.Fatalf("location registry = %#v, want distinct source-grounded identities for same-name locations", locationsValue)
+	}
+	if occurrenceOutput.File != "lanes/location-occurrences.json" || occurrenceOutput.SchemaID != locationoccurrencecodec.SchemaID || occurrenceOutput.SchemaVersion != locationoccurrencecodec.SchemaVersion {
+		t.Fatalf("location occurrence output = %#v, want typed occurrence JSON", occurrenceOutput)
+	}
+	occurrencesValue := readProductionJSON[dnd.LocationOccurrenceList](t, filepath.Join(runRoot, occurrenceOutput.File))
+	if len(occurrencesValue.Occurrences) != 2 || occurrencesValue.Occurrences[0].LocationID == occurrencesValue.Occurrences[1].LocationID || occurrencesValue.Occurrences[0].Name != "Moon Gate" || occurrencesValue.Occurrences[1].Name != "Moon Gate" {
+		t.Fatalf("location occurrences = %#v, want source-grounded references to distinct registry identities", occurrencesValue)
+	}
+
+	evidence := readProductionJSON[evidencecontext.Document](t, filepath.Join(runRoot, "evidence-context.json"))
+	for _, laneID := range []string{"enemy-events", "locations", "location-occurrences"} {
+		if !containsString(evidence.SelectedLanes, laneID) || !evidenceHasLane(evidence, laneID) {
+			t.Fatalf("evidence context = %#v, want direct %s evidence", evidence, laneID)
+		}
+	}
+
+	requests := client.requestsFor(enemyevents.PromptID)
+	if len(requests) != 1 {
+		t.Fatalf("enemy event requests = %#v, want only the combat scene request", requests)
+	}
+	request := requests[0]
+	if request.SessionID != "enemy-event-session" {
+		t.Fatalf("enemy event session = %q, want shared session", request.SessionID)
+	}
+	for slot, required := range map[string]string{
+		"npcs":             "Kesh",
+		"combat_turns":     "Kesh",
+		"npc_interactions": "Kesh",
+	} {
+		input, ok := request.Inputs[slot]
+		if !ok || !strings.Contains(string(input.Content), required) || strings.Contains(string(input.Content), "source_refs") || strings.Contains(string(input.Content), "start_unit_id") {
+			t.Fatalf("enemy event %s prompt input = %q, want compact source-free grounding", slot, input.Content)
+		}
+	}
+	locationRequests := client.requestsFor(locationoccurrences.PromptID)
+	if len(locationRequests) != 2 {
+		t.Fatalf("location occurrence requests = %#v, want one request per scene", locationRequests)
+	}
+	for _, request := range locationRequests {
+		registryInput := request.Inputs["locations"]
+		if !strings.Contains(string(registryInput.Content), "Moon Gate") || !strings.Contains(string(registryInput.Content), `"id"`) || strings.Contains(string(registryInput.Content), "source_refs") {
+			t.Fatalf("location occurrence registry input = %q, want source-free ID grounding", registryInput.Content)
+		}
+	}
+}
+
+func completeExampleConfigWithTemporaryCache(t *testing.T) string {
+	t.Helper()
+	content, err := os.ReadFile(repositoryPath("examples", "dnd-complete.config.yml"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	cacheRoot := t.TempDir()
+	updated := strings.Replace(string(content), "directory: ./notarius-cache/chunk-plans", fmt.Sprintf("directory: %q", filepath.Join(cacheRoot, "chunk-plans")), 1)
+	updated = strings.Replace(updated, "directory: ./notarius-cache/checkpoints", fmt.Sprintf("directory: %q", filepath.Join(cacheRoot, "checkpoints")), 1)
+	for relative, absolute := range map[string]string{
+		"./dnd-party.txt":          repositoryPath("examples", "dnd-party.txt"),
+		"./dnd-glossary.txt":       repositoryPath("examples", "dnd-glossary.txt"),
+		"./dnd-spell-catalog.json": repositoryPath("examples", "dnd-spell-catalog.json"),
+	} {
+		updated = strings.ReplaceAll(updated, relative, fmt.Sprintf("%q", absolute))
+	}
+	path := filepath.Join(t.TempDir(), "dnd-complete.config.yml")
+	if err := os.WriteFile(path, []byte(updated), 0o600); err != nil {
+		t.Fatal(err)
+	}
+	return path
+}
+
+type enemyEventLLMClient struct {
+	mu       sync.Mutex
+	requests []contracts.StructuredCompletionRequest
+}
+
+func (client *enemyEventLLMClient) CompleteStructured(ctx context.Context, request contracts.StructuredCompletionRequest, output any) (contracts.StructuredCompletionResponse, error) {
+	if err := ctx.Err(); err != nil {
+		return contracts.StructuredCompletionResponse{}, err
+	}
+	combatScene := strings.Contains(string(request.Inputs["transcript"].Content), "Roll initiative")
+	var content []byte
+	switch request.PromptID {
+	case scenes.PromptID:
+		content = []byte(`{"scenes":[{"start_unit_id":1,"end_unit_id":6},{"start_unit_id":7,"end_unit_id":11}]}`)
+	case npcs.PromptID:
+		if combatScene {
+			content = []byte(`{"npcs":[{"name":"Kesh","source_refs":[{"start_unit_id":7,"end_unit_id":7}]}]}`)
+		} else {
+			content = []byte(`{"npcs":[]}`)
+		}
+	case npcnormalize.PromptID:
+		content = []byte(`{"duplicate_groups":[]}`)
+	case scenedescriptions.PromptID:
+		kind, title := "narrative", "Arrival"
+		if combatScene {
+			kind, title = "combat", "Raiders attack"
+		}
+		content = []byte(fmt.Sprintf(`{"kind":%q,"title":%q,"summary":"session scene"}`, kind, title))
+	case locations.PromptID:
+		unitID := 1
+		if combatScene {
+			unitID = 7
+		}
+		content = []byte(fmt.Sprintf(`{"locations":[{"name":"Moon Gate","source_refs":[{"start_unit_id":%d,"end_unit_id":%d}]}]}`, unitID, unitID))
+	case locationnormalize.PromptID:
+		content = []byte(`{"duplicate_groups":[]}`)
+	case spells.PromptID:
+		content = []byte(`{"spell_casts":[]}`)
+	case itemevents.PromptID:
+		content = []byte(`{"events":[]}`)
+	case combat.PromptID:
+		content = []byte(`{"combat_turns":[{"actor":"Kesh","turn_kind":"turn","source_refs":[{"start_unit_id":8,"end_unit_id":8}]}]}`)
+	case npcinteractions.PromptID:
+		if combatScene {
+			content = []byte(`{"interactions":[{"name":"Kesh","kind":"combat_opponent","source_refs":[{"start_unit_id":7,"end_unit_id":7}]}]}`)
+		} else {
+			content = []byte(`{"interactions":[]}`)
+		}
+	case locationoccurrences.PromptID:
+		var registry struct {
+			Locations []struct {
+				ID string `json:"id"`
+			} `json:"locations"`
+		}
+		if err := json.Unmarshal(request.Inputs["locations"].Content, &registry); err != nil {
+			return contracts.StructuredCompletionResponse{}, fmt.Errorf("decode generated location registry: %w", err)
+		}
+		if len(registry.Locations) == 0 {
+			return contracts.StructuredCompletionResponse{}, fmt.Errorf("generated location registry has no locations")
+		}
+		unitID := 1
+		locationID := registry.Locations[0].ID
+		if combatScene {
+			unitID = 7
+			if len(registry.Locations) > 1 {
+				locationID = registry.Locations[1].ID
+			}
+		}
+		content = []byte(fmt.Sprintf(`{"occurrences":[{"location_id":%q,"name":"Moon Gate","kind":"visited","source_refs":[{"start_unit_id":%d,"end_unit_id":%d}]}]}`, locationID, unitID, unitID))
+	case enemyevents.PromptID:
+		content = []byte(`{"events":[{"name":"Kesh","kind":"fled","source_refs":[{"start_unit_id":10,"end_unit_id":10}]}]}`)
+	default:
+		return contracts.StructuredCompletionResponse{}, fmt.Errorf("unexpected prompt %q", request.PromptID)
+	}
+	if err := json.Unmarshal(content, output); err != nil {
+		return contracts.StructuredCompletionResponse{}, fmt.Errorf("populate fake structured target: %w", err)
+	}
+	client.mu.Lock()
+	client.requests = append(client.requests, request)
+	client.mu.Unlock()
+	return contracts.StructuredCompletionResponse{Content: content, Provider: "test", Model: "deterministic", ProfileID: request.ProfileID}, nil
+}
+
+func (client *enemyEventLLMClient) requestsFor(promptID string) []contracts.StructuredCompletionRequest {
+	client.mu.Lock()
+	defer client.mu.Unlock()
+	var requests []contracts.StructuredCompletionRequest
+	for _, request := range client.requests {
+		if request.PromptID == promptID {
+			requests = append(requests, request)
+		}
+	}
+	return requests
+}
+
+func containsString(values []string, want string) bool {
+	for _, value := range values {
+		if value == want {
+			return true
+		}
+	}
+	return false
+}
+
+func evidenceHasLane(value evidencecontext.Document, laneID string) bool {
+	for _, context := range value.Contexts {
+		for _, reference := range context.EvidenceRefs {
+			if reference.LaneID == laneID {
+				return true
+			}
+		}
+	}
+	return false
+}
+
+func generatedReferenceBinding(bindings []pipeline.ReferenceBinding, slotName string) (pipeline.ReferenceBinding, bool) {
+	for _, binding := range bindings {
+		if binding.SlotName == slotName && binding.Artifact != nil {
+			return binding, true
+		}
+	}
+	return pipeline.ReferenceBinding{}, false
+}

internal/cli/example_contract_test.go

@@ -15,6 +15,10 @@ import (
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	locationoccurrenceextract "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/locationoccurrences"
+	locationextract "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/locations"
+	locationoccurrencenormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/locationoccurrences"
+	locationnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/locations"
 	spellnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/spells"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/seriatim/input/transcript"
 )
@@ -24,6 +28,9 @@ func TestMaintainedExamplesLoadResolveAndList(t *testing.T) {
 	for _, example := range maintainedExampleFiles(t) {
 		t.Run(example.name, func(t *testing.T) {
 			cfg := loadMaintainedExample(t, example.path)
+			if example.name == "complete" && (cfg.Concurrency.TotalLLM != 2 || cfg.Concurrency.StageWorkers["extract"] != 2) {
+				t.Fatalf("complete example concurrency = %#v, want explicit limits of 2", cfg.Concurrency)
+			}
 			raw, err := os.ReadFile(example.transcriptPath)
 			if err != nil {
 				t.Fatalf("read maintained transcript %q: %v", example.transcriptPath, err)
@@ -51,8 +58,30 @@ func TestMaintainedExamplesLoadResolveAndList(t *testing.T) {
 					t.Fatalf("materialize maintained example references for %q: %v", pipelineID, err)
 				}
 				if example.name == "complete" {
-					if got := exampleStepLaneIDs(materialized); strings.Join(got, "|") != "describe-session:item-events,npcs,scene-descriptions|extract-events:combat-turns,npc-interactions,spells" {
-						t.Fatalf("complete example steps and lanes = %v, want every D&D extractor in the documented two-step composition", got)
+					if got := exampleStepLaneIDs(materialized); strings.Join(got, "|") != "describe-session:item-events,locations,npcs,scene-descriptions|extract-events:combat-turns,location-occurrences,npc-interactions,spells|track-enemies:enemy-events" {
+						t.Fatalf("complete example steps and lanes = %v, want the documented D&D extractor composition", got)
+					}
+					locationLane := referenceContractLane(t, materialized, "locations")
+					if locationLane.ArtifactKind != dnd.LocationListKind || locationLane.Extract.Module != locationextract.Key || locationLane.Extract.Retries != 2 || locationLane.Merge.Module != pipeline.DefaultMergeModule || locationLane.Normalize.Module != locationnormalize.Key || locationLane.Normalize.Retries != 2 {
+						t.Fatalf("location lane = %#v, want typed registry composition", locationLane)
+					}
+					occurrenceLane := referenceContractLane(t, materialized, "location-occurrences")
+					if occurrenceLane.ArtifactKind != dnd.LocationOccurrenceListKind || occurrenceLane.Extract.Module != locationoccurrenceextract.Key || occurrenceLane.Extract.Retries != 2 || occurrenceLane.Merge.Module != pipeline.DefaultMergeModule || occurrenceLane.Normalize.Module != locationoccurrencenormalize.Key {
+						t.Fatalf("location occurrence lane = %#v, want typed occurrence composition", occurrenceLane)
+					}
+					for _, target := range []pipeline.ResolvedReferenceTarget{occurrenceLane.ExtractReferences, occurrenceLane.NormalizeReferences} {
+						binding, found := generatedReferenceBinding(target.Bindings, "locations")
+						if !found || binding.Artifact.Step != "describe-session" || binding.Artifact.Lane != "locations" {
+							t.Fatalf("location occurrence %s reference = %#v, want generated location registry", target.Stage, binding)
+						}
+					}
+					for _, slot := range []string{"party", "glossary"} {
+						if len(occurrenceLane.ExtractReferences.ReferenceSet.Slots[slot].Items) != 1 {
+							t.Fatalf("location occurrence extractor %s reference was not materialized: %#v", slot, occurrenceLane.ExtractReferences)
+						}
+						if _, found := occurrenceLane.NormalizeReferences.ReferenceSet.Slots[slot]; found {
+							t.Fatalf("location occurrence normalizer unexpectedly consumes %s: %#v", slot, occurrenceLane.NormalizeReferences)
+						}
 					}
 					spellLane := referenceContractLane(t, materialized, "spells")
 					if len(spellLane.ExtractReferences.ReferenceSet.Slots["spell_catalog"].Items) != 1 ||
@@ -68,6 +97,18 @@ func TestMaintainedExamplesLoadResolveAndList(t *testing.T) {
 							t.Fatalf("item event lane unexpectedly depends on generated scene descriptions: %#v", itemEventLane)
 						}
 					}
+					enemyEventLane := referenceContractLane(t, materialized, "enemy-events")
+					for slot, want := range map[string]struct{ step, lane string }{
+						"npcs":               {step: "describe-session", lane: "npcs"},
+						"scene_descriptions": {step: "describe-session", lane: "scene-descriptions"},
+						"combat_turns":       {step: "extract-events", lane: "combat-turns"},
+						"npc_interactions":   {step: "extract-events", lane: "npc-interactions"},
+					} {
+						binding, found := generatedReferenceBinding(enemyEventLane.ExtractReferences.Bindings, slot)
+						if !found || binding.Artifact.Step != want.step || binding.Artifact.Lane != want.lane {
+							t.Fatalf("enemy event %s reference = %#v, want generated %s/%s artifact", slot, binding, want.step, want.lane)
+						}
+					}
 				}
 			}
 			var stdout, stderr strings.Builder
@@ -242,6 +283,12 @@ func TestMaintainedMalformedInputOnlyRecordsDebugFailureWhenRequested(t *testing
 			if report.Succeeded || report.PipelineID != "dnd-session" {
 				t.Fatalf("failure report = %#v, want failed dnd-session report", report)
 			}
+			invocation := readProductionJSON[debugbundle.Invocation](t, filepath.Join(bundle, "summary", "invocation.json"))
+			manifest := readProductionJSON[artifacts.RunManifest](t, filepath.Join(bundle, "summary", "run-manifest.json"))
+			manifestSession, found := manifest.Metadata["session_id"]
+			if invocation.SessionID == "" || !found || manifestSession != invocation.SessionID {
+				t.Fatalf("failed run sessions: invocation=%q manifest=%#v metadata=%#v", invocation.SessionID, manifestSession, manifest.Metadata)
+			}
 		})
 	}
 }

internal/cli/production_contract_test.go

@@ -29,12 +29,15 @@ import (
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/chunk/scenes"
 	combatcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/combatturns"
+	enemyeventcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/enemyevents"
 	itemeventcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/itemevents"
 	spellcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/spells"
 	combatextract "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/combatturns"
+	enemyeventextract "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/enemyevents"
 	itemeventextract "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/itemevents"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/spells"
 	combatnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/combatturns"
+	enemyeventnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/enemyevents"
 	itemeventnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/itemevents"
 	spellnormalize "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/normalize/spells"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/generic/normalize/noop"
@@ -47,9 +50,9 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 
 	assertProductionContains(t, "inputs", registries.Inputs.RegisteredKeys(), []string{"seriatim"})
 	assertProductionContains(t, "chunkers", registries.Chunkers.RegisteredKeys(), []string{"dnd/scenes", "generic"})
-	assertProductionContains(t, "extractors", registries.Extractors.RegisteredKeys(), []string{"dnd/spells", "dnd/npcs", combatextract.Key, itemeventextract.Key})
+	assertProductionContains(t, "extractors", registries.Extractors.RegisteredKeys(), []string{"dnd/spells", "dnd/npcs", combatextract.Key, itemeventextract.Key, enemyeventextract.Key})
 	assertProductionContains(t, "mergers", registries.Mergers.RegisteredKeys(), []string{"appendorder"})
-	assertProductionContains(t, "normalizers", registries.Normalizers.RegisteredKeys(), []string{"noop", spellnormalize.Key, "dnd/npcs", combatnormalize.Key, itemeventnormalize.Key})
+	assertProductionContains(t, "normalizers", registries.Normalizers.RegisteredKeys(), []string{"noop", spellnormalize.Key, "dnd/npcs", combatnormalize.Key, itemeventnormalize.Key, enemyeventnormalize.Key})
 	assertProductionContains(t, "outputs", registries.Outputs.RegisteredKeys(), []string{"json"})
 	assertProductionContains(t, "validators", registries.Validators.RegisteredKeys(), []string{
 		"extract/dnd/spells/catalog",
@@ -64,17 +67,23 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 		"extract/dnd/item-events/source_refs",
 		"extract/dnd/item-events/source_relatedness",
 		"normalize/dnd/item-events/invariants",
+		"extract/dnd/enemy-events/shape",
+		"extract/dnd/enemy-events/engagements",
+		"extract/dnd/enemy-events/source_refs",
+		"extract/dnd/enemy-events/source_relatedness",
+		"normalize/dnd/enemy-events/invariants",
 		"generic/always_accept",
 		"generic/always_reject",
 		"generic/valid_json",
 		"generic/valid_json_schema",
 	})
-	assertProductionContains(t, "artifact codec kinds", registries.ArtifactCodecs.RegisteredKinds(), []contracts.ArtifactKind{dnd.SpellListKind, dnd.NPCListKind, dnd.CombatTurnListKind, dnd.ItemEventListKind})
-	assertProductionContains(t, "merger variants", registries.Mergers.RegisteredArtifactKinds(pipeline.DefaultMergeModule), []contracts.ArtifactKind{dnd.SpellListKind, dnd.NPCListKind, dnd.CombatTurnListKind, dnd.ItemEventListKind})
-	assertProductionContains(t, "normalizer variants", registries.Normalizers.RegisteredArtifactKinds(pipeline.DefaultNormalizeModule), []contracts.ArtifactKind{dnd.SpellListKind, dnd.NPCListKind, dnd.CombatTurnListKind, dnd.ItemEventListKind})
+	assertProductionContains(t, "artifact codec kinds", registries.ArtifactCodecs.RegisteredKinds(), []contracts.ArtifactKind{dnd.SpellListKind, dnd.NPCListKind, dnd.CombatTurnListKind, dnd.ItemEventListKind, dnd.EnemyEventListKind})
+	assertProductionContains(t, "merger variants", registries.Mergers.RegisteredArtifactKinds(pipeline.DefaultMergeModule), []contracts.ArtifactKind{dnd.SpellListKind, dnd.NPCListKind, dnd.CombatTurnListKind, dnd.ItemEventListKind, dnd.EnemyEventListKind})
+	assertProductionContains(t, "normalizer variants", registries.Normalizers.RegisteredArtifactKinds(pipeline.DefaultNormalizeModule), []contracts.ArtifactKind{dnd.SpellListKind, dnd.NPCListKind, dnd.CombatTurnListKind, dnd.ItemEventListKind, dnd.EnemyEventListKind})
 	assertProductionContains(t, "spell normalizer variants", registries.Normalizers.RegisteredArtifactKinds(spellnormalize.Key), []contracts.ArtifactKind{dnd.SpellListKind})
 	assertProductionContains(t, "combat normalizer variants", registries.Normalizers.RegisteredArtifactKinds(combatnormalize.Key), []contracts.ArtifactKind{dnd.CombatTurnListKind})
 	assertProductionContains(t, "item event normalizer variants", registries.Normalizers.RegisteredArtifactKinds(itemeventnormalize.Key), []contracts.ArtifactKind{dnd.ItemEventListKind})
+	assertProductionContains(t, "enemy event normalizer variants", registries.Normalizers.RegisteredArtifactKinds(enemyeventnormalize.Key), []contracts.ArtifactKind{dnd.EnemyEventListKind})
 
 	wantChain := []pipeline.ModuleBinding{
 		pipeline.Binding("generic/valid_json"),
@@ -132,6 +141,17 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 	if got := registries.ValidatorChains.Validators(pipeline.StageNormalize, itemeventnormalize.Key); !reflect.DeepEqual(got, itemEventNormalizeChain) {
 		t.Fatalf("item event normalize validator chain = %#v, want %#v", got, itemEventNormalizeChain)
 	}
+	enemyEventExtractChain := []pipeline.ModuleBinding{
+		pipeline.Binding("generic/valid_json"),
+		pipeline.Binding("extract/dnd/enemy-events/shape"),
+		pipeline.Binding("extract/dnd/enemy-events/engagements"),
+		pipeline.Binding("extract/dnd/enemy-events/source_refs"),
+		pipeline.Binding("generic/valid_json_schema"),
+		pipeline.Binding("extract/dnd/enemy-events/source_relatedness"),
+	}
+	if got := registries.ValidatorChains.Validators(pipeline.StageExtract, enemyeventextract.Key); !reflect.DeepEqual(got, enemyEventExtractChain) {
+		t.Fatalf("enemy event extract validator chain = %#v, want %#v", got, enemyEventExtractChain)
+	}
 
 	assetNames := productionAssetNames(t, components.assets.PromptFS)
 	requiredAssets := []string{
@@ -162,6 +182,10 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 		"dnd.item_events/sharedassets/common-dnd-system.md",
 		"dnd.item_events/sharedassets/common-dnd-transcript.md",
 		"dnd.item_events/task.md",
+		"dnd.enemy_events/dnd.enemy_events.yaml",
+		"dnd.enemy_events/grounding.md",
+		"dnd.enemy_events/instructions.md",
+		"dnd.enemy_events/task.md",
 	}
 	assertProductionContains(t, "production prompt assets", assetNames, requiredAssets)
 
@@ -180,6 +204,7 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 		{stage: pipeline.StageExtract, key: "dnd/item-events", want: contracts.ExecutionClassLLMBacked},
 		{stage: pipeline.StageExtract, key: "dnd/npc-interactions", want: contracts.ExecutionClassLLMBacked},
 		{stage: pipeline.StageExtract, key: "dnd/scene-descriptions", want: contracts.ExecutionClassLLMBacked},
+		{stage: pipeline.StageExtract, key: enemyeventextract.Key, want: contracts.ExecutionClassLLMBacked},
 		{stage: pipeline.StageMerge, key: "appendorder", want: contracts.ExecutionClassDeterministic},
 		{stage: pipeline.StageNormalize, key: "noop", want: contracts.ExecutionClassDeterministic},
 		{stage: pipeline.StageNormalize, key: "dnd/spells", want: contracts.ExecutionClassDeterministic},
@@ -188,6 +213,7 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 		{stage: pipeline.StageNormalize, key: "dnd/item-events", want: contracts.ExecutionClassDeterministic},
 		{stage: pipeline.StageNormalize, key: "dnd/npc-interactions", want: contracts.ExecutionClassDeterministic},
 		{stage: pipeline.StageNormalize, key: "dnd/scene-descriptions", want: contracts.ExecutionClassDeterministic},
+		{stage: pipeline.StageNormalize, key: enemyeventnormalize.Key, want: contracts.ExecutionClassDeterministic},
 		{stage: pipeline.StageOutput, key: "json", want: contracts.ExecutionClassDeterministic},
 	} {
 		got, ok := catalog.ExecutionClass(test.stage, test.key)
@@ -211,6 +237,10 @@ func TestProductionCatalogCoversMaintainedConfigurations(t *testing.T) {
 	if !ok || itemEventCodecSpec.Kind != dnd.ItemEventListKind || itemEventCodecSpec.Schema.ID != itemeventcodec.SchemaID {
 		t.Fatalf("item event codec spec = %#v, ok=%t, want typed D&D item-event codec", itemEventCodecSpec, ok)
 	}
+	enemyEventCodecSpec, ok := catalog.ArtifactCodecs.Spec(dnd.EnemyEventListKind)
+	if !ok || enemyEventCodecSpec.Kind != dnd.EnemyEventListKind || enemyEventCodecSpec.Schema.ID != enemyeventcodec.SchemaID {
+		t.Fatalf("enemy event codec spec = %#v, ok=%t, want typed D&D enemy-event codec", enemyEventCodecSpec, ok)
+	}
 	if got := catalog.ValidatorChains.Validators(pipeline.StageExtract, spells.Key); !reflect.DeepEqual(got, wantChain) {
 		t.Fatalf("catalog validator chain = %#v, want %#v", got, wantChain)
 	}

internal/cli/run.go

@@ -153,10 +153,10 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 	resume := fs.Bool("resume", false, "reuse compatible recorded checkpoints")
 	recomputeStep := singleValueFlag{name: "--recompute-step"}
 	chunkCache := chunkCacheFlag{}
-	sessionID := sessionIDFlag{}
+	requestedSessionID := sessionIDFlag{}
 	referenceFlags := stringListFlag{}
 	withoutReferenceFlags := stringListFlag{}
-	fs.Var(&sessionID, "session-id", "prompt session identifier")
+	fs.Var(&requestedSessionID, "session-id", "prompt session identifier")
 	fs.Var(&reasoningEffort, "reasoning-effort", "reasoning effort override")
 	fs.Var(&chunkCache, "chunk_cache", "chunk plan cache mode: auto, bypass, or refresh")
 	fs.Var(&referenceFlags, "reference", "reference binding, as slot=path, chunk.slot=path, merge.slot=path, lane.slot=path, lane.extract.slot=path, lane.merge.slot=path, or lane.normalize.slot=path")
@@ -199,7 +199,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 		fmt.Fprintln(stderr, "notarius: --debug-dir must not be empty")
 		return 2
 	}
-	if sessionID.set && strings.TrimSpace(sessionID.value) == "" {
+	if requestedSessionID.set && strings.TrimSpace(requestedSessionID.value) == "" {
 		fmt.Fprintln(stderr, "notarius: --session-id must not be empty")
 		return 2
 	}
@@ -414,11 +414,19 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 	if err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("read input %q: %w", strings.TrimSpace(*inputPath), err))
 	}
+	effectiveSessionID, err := resolvePromptSessionID(requestedSessionID.value, effective.ResolvedPipeline.Input.Module, rawInput)
+	if err != nil {
+		return failPipelineCommand(stderr, commandState, terminalWriter, err)
+	}
+	invocation.SessionID = effectiveSessionID
+	if err := writeSummary(summary, func() error { return summary.WriteInvocation(invocation) }); err != nil {
+		return failPipelineCommand(stderr, commandState, terminalWriter, fmt.Errorf("write debug invocation metadata: %w", err))
+	}
 	chunkPlans, err := chunkPlanStoreForRun(effective.Config.Cache.ChunkPlans, opts)
 	if err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, err)
 	}
-	checkpointRecorder, checkpointLoader, err := checkpointHandlersForRun(effective.Config.Cache.Checkpoints, opts, effective.ResolvedPipeline, prepared.CheckpointFingerprints(), llmFingerprints, rawInput, only, llmProfiles, strings.TrimSpace(*llmProfile), strings.TrimSpace(sessionID.value), runtimeOverrides, *resume)
+	checkpointRecorder, checkpointLoader, err := checkpointHandlersForRun(effective.Config.Cache.Checkpoints, opts, effective.ResolvedPipeline, prepared.CheckpointFingerprints(), llmFingerprints, rawInput, only, llmProfiles, strings.TrimSpace(*llmProfile), effectiveSessionID, runtimeOverrides, *resume)
 	if err != nil {
 		return failPipelineCommand(stderr, commandState, terminalWriter, err)
 	}
@@ -427,7 +435,7 @@ func runPipelineCommand(args []string, stdout, stderr io.Writer, opts Options) i
 		Prepared:         prepared,
 		Path:             strings.TrimSpace(*inputPath),
 		RawInput:         rawInput,
-		SessionID:        strings.TrimSpace(sessionID.value),
+		SessionID:        effectiveSessionID,
 		RunID:            runID,
 		StartedAt:        startedAt,
 		LLMProfiles:      llmProfiles,

internal/cli/run_contract_test.go

@@ -492,19 +492,30 @@ func TestEffectiveLLMProfileIDsAreSortedDeduplicatedAndLLMOnly(t *testing.T) {
 	}
 }
 
-func TestRunSessionIDUsesExplicitValueOrSourceDocumentID(t *testing.T) {
+func TestRunSessionIDUsesEffectiveValueForPromptRequests(t *testing.T) {
 	for _, tt := range []struct {
 		name string
 		args []string
 		want string
 	}{
-		{name: "source document", want: "source"},
+		{name: "derived default"},
 		{name: "explicit trimmed value", args: []string{"--session-id", " explicit-session "}, want: "explicit-session"},
 	} {
 		t.Run(tt.name, func(t *testing.T) {
 			roots := newStateTestRoots(t)
+			want := tt.want
+			if want == "" {
+				rawInput, err := os.ReadFile(roots.input)
+				if err != nil {
+					t.Fatal(err)
+				}
+				want, err = resolvePromptSessionID("", "test/input", rawInput)
+				if err != nil {
+					t.Fatal(err)
+				}
+			}
 			harness := newStateTestHarness()
-			args := append([]string{"run", "sample", "--config", roots.config, "--input", roots.input, "--chunk_cache", "bypass"}, tt.args...)
+			args := append([]string{"run", "sample", "--config", roots.config, "--input", roots.input, "--chunk_cache", "bypass", "--debug"}, tt.args...)
 			var stdout, stderr bytes.Buffer
 			code := RunWithOptions(args, &stdout, &stderr, harness.options())
 			if code != 0 || stderr.Len() != 0 {
@@ -517,10 +528,15 @@ func TestRunSessionIDUsesExplicitValueOrSourceDocumentID(t *testing.T) {
 				t.Fatalf("session IDs = %#v, want all prompt-facing module requests", sessions)
 			}
 			for _, session := range sessions {
-				if session != tt.want {
-					t.Fatalf("session IDs = %#v, want %q", sessions, tt.want)
+				if session != want {
+					t.Fatalf("session IDs = %#v, want %q", sessions, want)
 				}
 			}
+			var manifest artifacts.RunManifest
+			readStateTestSummaryJSON(t, onlyChildDir(t, roots.debug), "run-manifest.json", &manifest)
+			if session, ok := manifest.Metadata["session_id"]; !ok || session != want {
+				t.Fatalf("manifest session = %#v, want %q; metadata = %#v", session, want, manifest.Metadata)
+			}
 		})
 	}
 }

internal/cli/session.go

@@ -0,0 +1,26 @@
+package cli
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"fmt"
+	"strings"
+)
+
+const generatedSessionIDPrefix = "notarius:v1:"
+
+func resolvePromptSessionID(explicitSessionID, inputModule string, rawInput []byte) (string, error) {
+	inputModule = strings.TrimSpace(inputModule)
+	if inputModule == "" {
+		return "", fmt.Errorf("resolve prompt session: input module key must not be empty")
+	}
+	if sessionID := strings.TrimSpace(explicitSessionID); sessionID != "" {
+		return sessionID, nil
+	}
+
+	hasher := sha256.New()
+	_, _ = hasher.Write([]byte(inputModule))
+	_, _ = hasher.Write([]byte{0})
+	_, _ = hasher.Write(rawInput)
+	return generatedSessionIDPrefix + hex.EncodeToString(hasher.Sum(nil)), nil
+}

internal/cli/session_test.go

@@ -0,0 +1,61 @@
+package cli
+
+import "testing"
+
+func TestResolvePromptSessionIDUsesVersionedInputIdentity(t *testing.T) {
+	got, err := resolvePromptSessionID("", " seriatim/input/transcript ", []byte("{\"entries\":[\"one\"]}\n"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	const want = "notarius:v1:e15fefdca48653e73248b4157900547be1fd34250c0138f8d3d1f2e89b43bb25"
+	if got != want {
+		t.Fatalf("resolved session = %q, want %q", got, want)
+	}
+}
+
+func TestResolvePromptSessionIDStabilityAndOverride(t *testing.T) {
+	rawInput := []byte("same input")
+	baseline, err := resolvePromptSessionID("", "input/transcript", rawInput)
+	if err != nil {
+		t.Fatal(err)
+	}
+	repeated, err := resolvePromptSessionID("", "input/transcript", rawInput)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if baseline != repeated {
+		t.Fatalf("resolved sessions = %q and %q, want stable value", baseline, repeated)
+	}
+
+	differentModule, err := resolvePromptSessionID("", "input/other", rawInput)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if baseline == differentModule {
+		t.Fatalf("resolved sessions = %q for distinct input modules", baseline)
+	}
+
+	differentInput, err := resolvePromptSessionID("", "input/transcript", []byte("same inpuu"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if baseline == differentInput {
+		t.Fatalf("resolved sessions = %q for distinct input bytes", baseline)
+	}
+
+	override, err := resolvePromptSessionID(" explicit-session ", "input/other", []byte("different input"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if override != "explicit-session" {
+		t.Fatalf("resolved override = %q, want %q", override, "explicit-session")
+	}
+}
+
+func TestResolvePromptSessionIDRejectsEmptyInputModule(t *testing.T) {
+	for _, explicitSessionID := range []string{"", "explicit-session"} {
+		if _, err := resolvePromptSessionID(explicitSessionID, " \t", []byte("input")); err == nil {
+			t.Fatalf("resolvePromptSessionID(%q) error = nil, want empty module failure", explicitSessionID)
+		}
+	}
+}

internal/cli/state_hardening_test.go

@@ -329,11 +329,25 @@ func TestRunUsesOneInjectedIdentityForDebugOutputAndManifest(t *testing.T) {
 	if manifest.StartedAt == nil || !manifest.StartedAt.Equal(wantStartedAt) {
 		t.Fatalf("manifest started at = %v, want %v", manifest.StartedAt, wantStartedAt)
 	}
+	rawInput, err := os.ReadFile(roots.input)
+	if err != nil {
+		t.Fatal(err)
+	}
+	wantSessionID, err := resolvePromptSessionID("", "test/input", rawInput)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if sessionID, ok := manifest.Metadata["session_id"]; !ok || sessionID != wantSessionID {
+		t.Fatalf("manifest session = %#v, want %q; metadata = %#v", sessionID, wantSessionID, manifest.Metadata)
+	}
 	var invocation debugbundle.Invocation
 	readStateTestSummaryJSON(t, debugPath, "invocation.json", &invocation)
 	if invocation.RunID != runID || !invocation.StartedAt.Equal(wantStartedAt) {
 		t.Fatalf("debug invocation identity = %#v, want run %q at %v", invocation, runID, wantStartedAt)
 	}
+	if invocation.SessionID != wantSessionID {
+		t.Fatalf("debug invocation session = %q, want %q", invocation.SessionID, wantSessionID)
+	}
 	report := readStateTestRunReport(t, debugPath)
 	if !report.Succeeded || report.RunID != runID || report.PipelineID != "sample" || report.OutputPath != outputPath || report.DebugPath != debugPath || report.OutputCount != 1 || report.RejectedCount != 0 || report.WarningCount != 0 || report.ValidationStatus != "approved" {
 		t.Fatalf("success report = %#v", report)
@@ -343,6 +357,45 @@ func TestRunUsesOneInjectedIdentityForDebugOutputAndManifest(t *testing.T) {
 	}
 }
 
+func TestRunCheckpointReuseRequiresSameEffectiveSession(t *testing.T) {
+	roots := newStateTestRoots(t)
+	harness := newStateTestHarness()
+	run := func(extra ...string) stateTestResult {
+		args := []string{"run", "sample", "--config", roots.config, "--input", roots.input, "--chunk_cache", "bypass"}
+		args = append(args, extra...)
+		var stdout, stderr bytes.Buffer
+		return stateTestResult{code: RunWithOptions(args, &stdout, &stderr, harness.options()), stdout: stdout.String(), stderr: stderr.String()}
+	}
+
+	if result := run(); result.code != 0 {
+		t.Fatalf("initial run code=%d stdout=%q stderr=%q", result.code, result.stdout, result.stderr)
+	}
+	harness.mu.Lock()
+	initialExtractCalls := harness.extractCalls
+	harness.mu.Unlock()
+	if initialExtractCalls != 1 {
+		t.Fatalf("initial extract calls = %d, want 1", initialExtractCalls)
+	}
+	if result := run("--resume"); result.code != 0 {
+		t.Fatalf("same-session resume code=%d stdout=%q stderr=%q", result.code, result.stdout, result.stderr)
+	}
+	harness.mu.Lock()
+	reusedExtractCalls := harness.extractCalls
+	harness.mu.Unlock()
+	if reusedExtractCalls != initialExtractCalls {
+		t.Fatalf("extract calls after same-session resume = %d, want %d", reusedExtractCalls, initialExtractCalls)
+	}
+	if result := run("--resume", "--session-id", "different-session"); result.code != 0 {
+		t.Fatalf("different-session resume code=%d stdout=%q stderr=%q", result.code, result.stdout, result.stderr)
+	}
+	harness.mu.Lock()
+	differentSessionExtractCalls := harness.extractCalls
+	harness.mu.Unlock()
+	if differentSessionExtractCalls != initialExtractCalls+1 {
+		t.Fatalf("extract calls after different-session resume = %d, want %d", differentSessionExtractCalls, initialExtractCalls+1)
+	}
+}
+
 func TestRunWritesTerminalArtifactsForResolutionPipelineAndOutputFailures(t *testing.T) {
 	for _, tc := range []struct {
 		name             string

internal/core/config/config.go

@@ -34,6 +34,8 @@ type ConcurrencyConfig struct {
 	defaultedExtractWorkers  int
 }
 
+const defaultLLMConcurrency = 16
+
 type OutputConfig struct {
 	Directory string `json:"directory"`
 }
@@ -60,9 +62,9 @@ func Default() Config {
 	return Config{
 		Pipelines: map[string]pipeline.PipelineProfile{},
 		Concurrency: ConcurrencyConfig{
-			TotalLLM:                1,
-			StageWorkers:            map[string]int{"extract": 1},
-			defaultedExtractWorkers: 1,
+			TotalLLM:                defaultLLMConcurrency,
+			StageWorkers:            map[string]int{"extract": defaultLLMConcurrency},
+			defaultedExtractWorkers: defaultLLMConcurrency,
 		},
 		Output: OutputConfig{Directory: "./notarius-output"},
 		Cache:  CacheConfig{ChunkPlans: ChunkPlanCacheConfig{Mode: pipeline.ChunkCacheAuto}},

internal/core/config/file_config_contract_test.go

@@ -14,7 +14,7 @@ import (
 
 func TestDefaultReturnsDocumentedValuesAndIndependentMaps(t *testing.T) {
 	first := Default()
-	if first.Concurrency.TotalLLM != 1 || first.Concurrency.StageWorkers["extract"] != 1 {
+	if first.Concurrency.TotalLLM != 16 || first.Concurrency.StageWorkers["extract"] != 16 {
 		t.Fatalf("concurrency defaults = %#v", first.Concurrency)
 	}
 	if first.Output.Directory != "./notarius-output" || first.Debug.Directory != "./notarius-debug" {
@@ -31,7 +31,7 @@ func TestDefaultReturnsDocumentedValuesAndIndependentMaps(t *testing.T) {
 	first.Concurrency.StageWorkers["other"] = 100
 	first.Pipelines["changed"] = pipeline.PipelineProfile{}
 	second := Default()
-	if second.Concurrency.StageWorkers["extract"] != 1 || len(second.Concurrency.StageWorkers) != 1 || len(second.Pipelines) != 0 {
+	if second.Concurrency.StageWorkers["extract"] != 16 || len(second.Concurrency.StageWorkers) != 1 || len(second.Pipelines) != 0 {
 		t.Fatalf("Default() returned state shared with an earlier result: %#v", second)
 	}
 }
@@ -45,7 +45,7 @@ func TestFileConfigMinimalVersion4AppliesOverDefaults(t *testing.T) {
 	if cfg.Output.Directory != "./notarius-output" || cfg.Debug.Directory != "./notarius-debug" || cfg.Cache.ChunkPlans.Mode != pipeline.ChunkCacheAuto {
 		t.Fatalf("minimal file changed unrelated defaults: %#v", cfg)
 	}
-	if cfg.Concurrency.TotalLLM != 1 || cfg.Concurrency.StageWorkers["extract"] != 1 || len(cfg.Pipelines) != 0 {
+	if cfg.Concurrency.TotalLLM != 16 || cfg.Concurrency.StageWorkers["extract"] != 16 || len(cfg.Pipelines) != 0 {
 		t.Fatalf("minimal file did not retain defaults: %#v", cfg)
 	}
 }

internal/core/debugbundle/bundle_test.go

@@ -179,6 +179,30 @@ func TestWriteInvocationPreservesReasoningEffortOverrideStates(t *testing.T) {
 		})
 	}
 }
+
+func TestWriteInvocationOmitsEmptySessionID(t *testing.T) {
+	for _, sessionID := range []string{"", "session-123"} {
+		bundle, err := Allocate(t.TempDir(), testBundleRunID, time.Unix(0, 42))
+		if err != nil {
+			t.Fatal(err)
+		}
+		if err := bundle.Summary().WriteInvocation(Invocation{Operation: "run", SessionID: sessionID}); err != nil {
+			t.Fatal(err)
+		}
+		data, err := os.ReadFile(filepath.Join(bundle.SummaryRoot(), ArtifactInvocationMetadata))
+		if err != nil {
+			t.Fatal(err)
+		}
+		var payload map[string]any
+		if err := json.Unmarshal(data, &payload); err != nil {
+			t.Fatal(err)
+		}
+		value, found := payload["session_id"]
+		if found != (sessionID != "") || (found && value != sessionID) {
+			t.Fatalf("session found=%t value=%#v, want found=%t value=%q; JSON=%s", found, value, sessionID != "", sessionID, data)
+		}
+	}
+}
 func TestSummaryWriterInternalWritesConfineArtifacts(t *testing.T) {
 	bundle, err := Allocate(t.TempDir(), testBundleRunID, time.Unix(0, 42))
 	if err != nil {

internal/core/debugbundle/summary.go

@@ -40,6 +40,7 @@ type Invocation struct {
 	OnlyLanes               []string  `json:"only_lanes,omitempty"`
 	ChunkCacheOverride      string    `json:"chunk_cache_override,omitempty"`
 	ReasoningEffortOverride *string   `json:"reasoning_effort_override,omitempty"`
+	SessionID               string    `json:"session_id,omitempty"`
 	RunID                   string    `json:"run_id"`
 	StartedAt               time.Time `json:"started_at"`
 }

internal/framework/pipeline/runner.go

@@ -91,6 +91,11 @@ func (r *Runner) Run(ctx context.Context, input RunInput) (output RunOutput, err
 	input.llmClient = input.Prepared.dependencies.LLM
 
 	output.Manifest = manifestFromPipeline(input)
+	sessionID := strings.TrimSpace(input.SessionID)
+	output.Manifest.Metadata, err = manifestMetadataWithSessionID(output.Manifest.Metadata, sessionID)
+	if err != nil {
+		return failOutput(output), err
+	}
 	output.ChunkPlan = &artifacts.ChunkPlanSummary{
 		Mode: string(effectiveChunkCacheMode(input.ChunkCacheMode)), RequestedModule: input.pipeline.Chunk.Module,
 		LookupStatus: "skipped", LookupReason: "chunk plan lookup skipped",
@@ -191,11 +196,6 @@ func (r *Runner) Run(ctx context.Context, input RunInput) (output RunOutput, err
 		return failOutput(output), fmt.Errorf("write source debug artifact: %w", err)
 	}
 	sourceInput := sourceInputMaterial(input.Path, input.RawInput)
-	sessionID := resolvedSessionID(input.SessionID, doc.ID)
-	output.Manifest.Metadata, err = manifestMetadataWithSessionID(output.Manifest.Metadata, sessionID)
-	if err != nil {
-		return failOutput(output), err
-	}
 	output.Manifest.SourceDigests = []string{doc.Digest}
 
 	chunker := input.Prepared.chunker
@@ -896,13 +896,6 @@ func sourceInputOriginURI(inputPath string) string {
 	return fileURI(inputPath)
 }
 
-func resolvedSessionID(explicit string, sourceDocumentID string) string {
-	if trimmed := strings.TrimSpace(explicit); trimmed != "" {
-		return trimmed
-	}
-	return strings.TrimSpace(sourceDocumentID)
-}
-
 func manifestMetadataWithSessionID(metadata map[string]any, sessionID string) (map[string]any, error) {
 	out, err := cloneMetadata(metadata)
 	if err != nil {

internal/framework/pipeline/runner_session_test.go

@@ -0,0 +1,37 @@
+package pipeline
+
+import (
+	"context"
+	"testing"
+)
+
+func TestRunnerUsesProvidedSessionWithoutSourceFallback(t *testing.T) {
+	for _, test := range []struct {
+		name    string
+		session string
+		wantSet bool
+		want    string
+	}{
+		{name: "provided", session: " supplied-session ", wantSet: true, want: "supplied-session"},
+		{name: "empty", wantSet: false},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			prepared, _ := preparedTerminalDebugPipeline(t)
+			output, err := New().Run(context.Background(), RunInput{
+				Prepared:  prepared,
+				RawInput:  []byte("input"),
+				SessionID: test.session,
+			})
+			if err != nil {
+				t.Fatal(err)
+			}
+			value, found := output.Manifest.Metadata["session_id"]
+			if found != test.wantSet {
+				t.Fatalf("manifest session presence = %t, want %t; metadata = %#v", found, test.wantSet, output.Manifest.Metadata)
+			}
+			if found && value != test.want {
+				t.Fatalf("manifest session = %#v, want %q", value, test.want)
+			}
+		})
+	}
+}

internal/modules/dnd/codec/enemyevents/assets/schemas/dnd_enemy_events.v1.json

@@ -0,0 +1,35 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.enemy_events",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["events"],
+  "properties": {
+    "events": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["name", "kind", "source_refs"],
+        "properties": {
+          "name": {"type": "string", "minLength": 1},
+          "kind": {"type": "string", "enum": ["engaged", "killed", "fled", "captured", "incapacitated"]},
+          "source_refs": {
+            "type": "array",
+            "minItems": 1,
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["source_id", "start_unit_id", "end_unit_id"],
+              "properties": {
+                "source_id": {"type": "string", "minLength": 1},
+                "start_unit_id": {"type": "integer", "minimum": 1},
+                "end_unit_id": {"type": "integer", "minimum": 1}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

internal/modules/dnd/codec/enemyevents/codec.go

@@ -0,0 +1,182 @@
+// Package enemyevents encodes durable D&D enemy-event artifacts.
+package enemyevents
+
+import (
+	"embed"
+	"encoding/json"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/candidatejson"
+)
+
+const (
+	SchemaID      = "notarius.dnd.enemy_events"
+	SchemaName    = "notarius_dnd_enemy_events_v1"
+	SchemaVersion = "v1"
+	MediaType     = "application/json"
+)
+
+//go:embed assets/schemas/dnd_enemy_events.v1.json
+var schemaAssets embed.FS
+
+var _ contracts.ArtifactCodec[dnd.EnemyEventList] = (*Codec)(nil)
+
+type Codec struct{}
+
+func New() *Codec { return &Codec{} }
+
+func (c *Codec) Kind() contracts.ArtifactKind { return dnd.EnemyEventListKind }
+
+func (c *Codec) Schema() contracts.ArtifactSchema {
+	raw, err := schemaAssets.ReadFile("assets/schemas/dnd_enemy_events.v1.json")
+	if err != nil {
+		return contracts.ArtifactSchema{}
+	}
+	return contracts.ArtifactSchema{
+		ID:         SchemaID,
+		Name:       SchemaName,
+		Version:    SchemaVersion,
+		JSONSchema: append([]byte(nil), raw...),
+	}
+}
+
+func (c *Codec) MediaType() string { return MediaType }
+
+func (c *Codec) Metadata(value dnd.EnemyEventList) map[string]any {
+	return map[string]any{"event_count": len(value.Events)}
+}
+
+func (c *Codec) Encode(value dnd.EnemyEventList) ([]byte, error) {
+	if err := validateRequiredValueFields(value); err != nil {
+		return nil, fmt.Errorf("encode dnd enemy event list: %w", err)
+	}
+	return c.EncodeCandidate(value)
+}
+
+// EncodeCandidate provides the durable representation before semantic
+// validators have approved a value.
+func (c *Codec) EncodeCandidate(value dnd.EnemyEventList) ([]byte, error) {
+	return candidatejson.EncodeCandidate("dnd enemy event list", cloneList(value))
+}
+
+func (c *Codec) Decode(content []byte) (dnd.EnemyEventList, error) {
+	value, err := c.DecodeCandidate(content)
+	if err != nil {
+		return dnd.EnemyEventList{}, err
+	}
+	if err := validateRequiredJSONFields(content); err != nil {
+		return dnd.EnemyEventList{}, fmt.Errorf("decode dnd enemy event list: %w", err)
+	}
+	return value, nil
+}
+
+// DecodeCandidate reads one strict durable JSON value before semantic
+// validators have approved it.
+func (c *Codec) DecodeCandidate(content []byte) (dnd.EnemyEventList, error) {
+	value, err := candidatejson.DecodeCandidate[dnd.EnemyEventList]("dnd enemy event list", content)
+	if err != nil {
+		return dnd.EnemyEventList{}, err
+	}
+	return cloneList(value), nil
+}
+
+func validateRequiredValueFields(value dnd.EnemyEventList) error {
+	if value.Events == nil {
+		return fmt.Errorf("events must be present")
+	}
+	for index, event := range value.Events {
+		if event.SourceRefs == nil {
+			return fmt.Errorf("events[%d].source_refs must be present", index)
+		}
+	}
+	return nil
+}
+
+func validateRequiredJSONFields(content []byte) error {
+	var root map[string]json.RawMessage
+	if err := json.Unmarshal(content, &root); err != nil || root == nil {
+		return fmt.Errorf("must be a JSON object")
+	}
+	events, err := requiredArray(root, "events", "")
+	if err != nil {
+		return err
+	}
+	for eventIndex, rawEvent := range events {
+		path := fmt.Sprintf("events[%d]", eventIndex)
+		var event map[string]json.RawMessage
+		if err := json.Unmarshal(rawEvent, &event); err != nil || event == nil {
+			return fmt.Errorf("%s must be an object", path)
+		}
+		for _, field := range []string{"name", "kind"} {
+			if err := requireField(event, field, path); err != nil {
+				return err
+			}
+		}
+		refs, err := requiredArray(event, "source_refs", path)
+		if err != nil {
+			return err
+		}
+		for refIndex, rawRef := range refs {
+			refPath := fmt.Sprintf("%s.source_refs[%d]", path, refIndex)
+			var ref map[string]json.RawMessage
+			if err := json.Unmarshal(rawRef, &ref); err != nil || ref == nil {
+				return fmt.Errorf("%s must be an object", refPath)
+			}
+			for _, field := range []string{"source_id", "start_unit_id", "end_unit_id"} {
+				if err := requireField(ref, field, refPath); err != nil {
+					return err
+				}
+			}
+		}
+	}
+	return nil
+}
+
+func requiredArray(object map[string]json.RawMessage, field, path string) ([]json.RawMessage, error) {
+	raw, err := requiredField(object, field, path)
+	if err != nil {
+		return nil, err
+	}
+	var values []json.RawMessage
+	if err := json.Unmarshal(raw, &values); err != nil {
+		return nil, fmt.Errorf("%s must be an array", fieldPath(path, field))
+	}
+	return values, nil
+}
+
+func requireField(object map[string]json.RawMessage, field, path string) error {
+	_, err := requiredField(object, field, path)
+	return err
+}
+
+func requiredField(object map[string]json.RawMessage, field, path string) (json.RawMessage, error) {
+	raw, ok := object[field]
+	if !ok || string(raw) == "null" {
+		return nil, fmt.Errorf("%s must be present", fieldPath(path, field))
+	}
+	return raw, nil
+}
+
+func fieldPath(path, field string) string {
+	if path == "" {
+		return field
+	}
+	return path + "." + field
+}
+
+func cloneList(value dnd.EnemyEventList) dnd.EnemyEventList {
+	if value.Events == nil {
+		return dnd.EnemyEventList{}
+	}
+	cloned := dnd.EnemyEventList{Events: make([]dnd.EnemyEvent, len(value.Events))}
+	for index, event := range value.Events {
+		cloned.Events[index] = event
+		if event.SourceRefs != nil {
+			cloned.Events[index].SourceRefs = append([]source.SourceRef(nil), event.SourceRefs...)
+		}
+	}
+	return cloned
+}

internal/modules/dnd/codec/enemyevents/codec_test.go

@@ -0,0 +1,99 @@
+package enemyevents
+
+import (
+	"encoding/json"
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+)
+
+func validList() dnd.EnemyEventList {
+	return dnd.EnemyEventList{Events: []dnd.EnemyEvent{
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindEngaged, SourceRefs: refs(1, 2)},
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindFled, SourceRefs: refs(5, 6)},
+	}}
+}
+
+func refs(start, end int) []source.SourceRef {
+	return []source.SourceRef{{SourceID: "session", StartUnitID: start, EndUnitID: end}}
+}
+
+func TestCodecRoundTripAndIdentity(t *testing.T) {
+	codec := New()
+	value := validList()
+	content, err := codec.Encode(value)
+	if err != nil {
+		t.Fatalf("Encode() error = %v", err)
+	}
+	decoded, err := codec.Decode(content)
+	if err != nil || !reflect.DeepEqual(decoded, value) {
+		t.Fatalf("Decode() = %#v, %v; want %#v", decoded, err, value)
+	}
+
+	schema := codec.Schema()
+	if codec.Kind() != dnd.EnemyEventListKind || codec.MediaType() != MediaType || schema.ID != SchemaID || schema.Name != SchemaName || schema.Version != SchemaVersion || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("codec identity/schema = %q/%q %#v", codec.Kind(), codec.MediaType(), schema)
+	}
+	registry := pipeline.NewArtifactCodecRegistry()
+	if err := pipeline.RegisterArtifactCodec(registry, codec); err != nil {
+		t.Fatal(err)
+	}
+	spec, ok := registry.Spec(dnd.EnemyEventListKind)
+	if !ok || spec.SchemaDigest != contracts.DigestArtifactSchema(schema) {
+		t.Fatalf("registered spec = %#v, %t", spec, ok)
+	}
+}
+
+func TestCodecRejectsStrictJSONAndMissingRequiredFields(t *testing.T) {
+	validJSON := `{"events":[{"name":"Ashfang","kind":"engaged","source_refs":[{"source_id":"session","start_unit_id":1,"end_unit_id":1}]}]}`
+	tests := []struct {
+		name, raw, want string
+	}{
+		{"invalid JSON", `{`, "decode dnd enemy event list"},
+		{"unknown root field", `{"events":[],"unexpected":true}`, "unknown field"},
+		{"unknown event field", strings.Replace(validJSON, `"kind":"engaged"`, `"kind":"engaged","unexpected":true`, 1), "unknown field"},
+		{"missing events", `{}`, "events must be present"},
+		{"missing name", strings.Replace(validJSON, `"name":"Ashfang",`, "", 1), "events[0].name must be present"},
+		{"missing kind", strings.Replace(validJSON, `"kind":"engaged",`, "", 1), "events[0].kind must be present"},
+		{"missing refs", strings.Replace(validJSON, `,"source_refs":[{"source_id":"session","start_unit_id":1,"end_unit_id":1}]`, "", 1), "events[0].source_refs must be present"},
+		{"missing source ID", strings.Replace(validJSON, `"source_id":"session",`, "", 1), "source_refs[0].source_id must be present"},
+		{"trailing JSON", `{"events":[]} {}`, "multiple JSON values"},
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := New().Decode([]byte(test.raw)); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Decode() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestCodecDefensivelyOwnsValuesAndDefersSemanticValidation(t *testing.T) {
+	codec := New()
+	candidate := dnd.EnemyEventList{Events: []dnd.EnemyEvent{{
+		Name: " ", Kind: "unsupported", SourceRefs: []source.SourceRef{{SourceID: "", StartUnitID: 0, EndUnitID: -1}},
+	}}}
+	content, err := codec.EncodeCandidate(candidate)
+	if err != nil {
+		t.Fatal(err)
+	}
+	decoded, err := codec.Decode(content)
+	if err != nil || !reflect.DeepEqual(decoded, candidate) {
+		t.Fatalf("Decode() = %#v, %v; want semantic candidate preservation", decoded, err)
+	}
+	decoded.Events[0].SourceRefs[0].SourceID = "changed"
+	if candidate.Events[0].SourceRefs[0].SourceID != "" {
+		t.Fatal("Decode() retained caller-owned source references")
+	}
+
+	first := codec.Schema()
+	first.JSONSchema[0] = '['
+	if second := codec.Schema(); !json.Valid(second.JSONSchema) || second.JSONSchema[0] == '[' {
+		t.Fatal("Schema() returned shared bytes")
+	}
+}

internal/modules/dnd/codec/locationoccurrences/assets/schemas/dnd_location_occurrences.v1.json

@@ -0,0 +1,54 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.location_occurrences",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["occurrences"],
+  "properties": {
+    "occurrences": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["location_id", "name", "kind", "source_refs"],
+        "properties": {
+          "location_id": {
+            "type": "string",
+            "pattern": "^location:sha256:[0-9a-f]{64}$"
+          },
+          "name": {
+            "type": "string",
+            "minLength": 1
+          },
+          "kind": {
+            "type": "string",
+            "enum": ["visited", "planned", "recalled", "mentioned"]
+          },
+          "source_refs": {
+            "type": "array",
+            "minItems": 1,
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["source_id", "start_unit_id", "end_unit_id"],
+              "properties": {
+                "source_id": {
+                  "type": "string",
+                  "minLength": 1
+                },
+                "start_unit_id": {
+                  "type": "integer",
+                  "minimum": 1
+                },
+                "end_unit_id": {
+                  "type": "integer",
+                  "minimum": 1
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

internal/modules/dnd/codec/locationoccurrences/codec.go

@@ -0,0 +1,126 @@
+// Package locationoccurrences encodes durable D&D location occurrence artifacts.
+package locationoccurrences
+
+import (
+	"embed"
+	"fmt"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/candidatejson"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+const (
+	SchemaID      = "notarius.dnd.location_occurrences"
+	SchemaName    = "notarius_dnd_location_occurrences_v1"
+	SchemaVersion = "v1"
+	MediaType     = "application/json"
+)
+
+//go:embed assets/schemas/dnd_location_occurrences.v1.json
+var schemaAssets embed.FS
+
+var _ contracts.ArtifactCodec[dnd.LocationOccurrenceList] = (*Codec)(nil)
+
+type Codec struct{}
+
+func New() *Codec { return &Codec{} }
+
+func (c *Codec) Kind() contracts.ArtifactKind { return dnd.LocationOccurrenceListKind }
+
+func (c *Codec) Schema() contracts.ArtifactSchema {
+	raw, err := schemaAssets.ReadFile("assets/schemas/dnd_location_occurrences.v1.json")
+	if err != nil {
+		return contracts.ArtifactSchema{}
+	}
+	return contracts.ArtifactSchema{
+		ID:         SchemaID,
+		Name:       SchemaName,
+		Version:    SchemaVersion,
+		JSONSchema: append([]byte(nil), raw...),
+	}
+}
+
+func (c *Codec) MediaType() string { return MediaType }
+
+func (c *Codec) Metadata(value dnd.LocationOccurrenceList) map[string]any {
+	return map[string]any{"occurrence_count": len(value.Occurrences)}
+}
+
+func (c *Codec) Encode(value dnd.LocationOccurrenceList) ([]byte, error) {
+	if err := validate(value); err != nil {
+		return nil, fmt.Errorf("encode dnd location occurrence list: %w", err)
+	}
+	return c.EncodeCandidate(value)
+}
+
+// EncodeCandidate provides the durable representation before semantic
+// validators have approved a value.
+func (c *Codec) EncodeCandidate(value dnd.LocationOccurrenceList) ([]byte, error) {
+	return candidatejson.EncodeCandidate("dnd location occurrence list", value)
+}
+
+func (c *Codec) Decode(content []byte) (dnd.LocationOccurrenceList, error) {
+	value, err := c.DecodeCandidate(content)
+	if err != nil {
+		return dnd.LocationOccurrenceList{}, err
+	}
+	if err := validate(value); err != nil {
+		return dnd.LocationOccurrenceList{}, fmt.Errorf("decode dnd location occurrence list: %w", err)
+	}
+	return value, nil
+}
+
+// DecodeCandidate reads one strict durable JSON value before semantic
+// validators have approved it.
+func (c *Codec) DecodeCandidate(content []byte) (dnd.LocationOccurrenceList, error) {
+	return candidatejson.DecodeCandidate[dnd.LocationOccurrenceList]("dnd location occurrence list", content)
+}
+
+func validate(value dnd.LocationOccurrenceList) error {
+	if value.Occurrences == nil {
+		return fmt.Errorf("occurrences must be present")
+	}
+	for index, occurrence := range value.Occurrences {
+		prefix := fmt.Sprintf("occurrences[%d]", index)
+		if !identity.IsValidID(occurrence.LocationID) {
+			return fmt.Errorf("%s.location_id must match location ID pattern", prefix)
+		}
+		if strings.TrimSpace(occurrence.Name) == "" {
+			return fmt.Errorf("%s.name must not be empty", prefix)
+		}
+		if !validOccurrenceKind(occurrence.Kind) {
+			return fmt.Errorf("%s.kind must be supported", prefix)
+		}
+		if len(occurrence.SourceRefs) == 0 {
+			return fmt.Errorf("%s.source_refs must contain at least one reference", prefix)
+		}
+		for refIndex, ref := range occurrence.SourceRefs {
+			refPrefix := fmt.Sprintf("%s.source_refs[%d]", prefix, refIndex)
+			if strings.TrimSpace(ref.SourceID) == "" {
+				return fmt.Errorf("%s.source_id must not be empty", refPrefix)
+			}
+			if ref.StartUnitID <= 0 {
+				return fmt.Errorf("%s.start_unit_id must be positive", refPrefix)
+			}
+			if ref.EndUnitID <= 0 {
+				return fmt.Errorf("%s.end_unit_id must be positive", refPrefix)
+			}
+		}
+	}
+	return nil
+}
+
+func validOccurrenceKind(value dnd.LocationOccurrenceKind) bool {
+	switch value {
+	case dnd.LocationOccurrenceKindVisited,
+		dnd.LocationOccurrenceKindPlanned,
+		dnd.LocationOccurrenceKindRecalled,
+		dnd.LocationOccurrenceKindMentioned:
+		return true
+	default:
+		return false
+	}
+}

internal/modules/dnd/codec/locationoccurrences/codec_test.go

@@ -0,0 +1,163 @@
+package locationoccurrences
+
+import (
+	"bytes"
+	"encoding/json"
+	"os"
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+func validList() dnd.LocationOccurrenceList {
+	refs := []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 1, EndUnitID: 2}}
+	return dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{{
+		LocationID: identity.DeriveID("The Old Tavern", refs),
+		Name:       "The Old Tavern",
+		Kind:       dnd.LocationOccurrenceKindVisited,
+		SourceRefs: refs,
+	}}}
+}
+
+func TestCodecMatchesMaintainedDurableFixture(t *testing.T) {
+	raw, err := os.ReadFile("testdata/dnd_location_occurrences.v1.json")
+	if err != nil {
+		t.Fatalf("read durable fixture: %v", err)
+	}
+	codec := New()
+	value, err := codec.Decode(raw)
+	if err != nil {
+		t.Fatalf("Decode() error = %v, want nil", err)
+	}
+	if want := validList(); !reflect.DeepEqual(value, want) {
+		t.Fatalf("Decode() = %#v, want %#v", value, want)
+	}
+	encoded, err := codec.Encode(value)
+	if err != nil {
+		t.Fatalf("Encode() error = %v, want nil", err)
+	}
+	var compact bytes.Buffer
+	if err := json.Compact(&compact, raw); err != nil {
+		t.Fatalf("compact durable fixture: %v", err)
+	}
+	if !bytes.Equal(encoded, compact.Bytes()) {
+		t.Fatalf("Encode() = %s, want %s", encoded, compact.Bytes())
+	}
+}
+
+func TestCodecOwnsDurableSchemaAndMetadata(t *testing.T) {
+	codec := New()
+	schema := codec.Schema()
+	if codec.Kind() != dnd.LocationOccurrenceListKind || codec.MediaType() != MediaType {
+		t.Fatalf("codec identity = %q/%q", codec.Kind(), codec.MediaType())
+	}
+	if schema.ID != SchemaID || schema.Name != SchemaName || schema.Version != SchemaVersion || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("schema = %#v, want durable location occurrence schema", schema)
+	}
+	var document map[string]any
+	if err := json.Unmarshal(schema.JSONSchema, &document); err != nil || document["$id"] != SchemaID {
+		t.Fatalf("durable schema document = %#v, %v", document, err)
+	}
+	registry := pipeline.NewArtifactCodecRegistry()
+	if err := pipeline.RegisterArtifactCodec(registry, codec); err != nil {
+		t.Fatalf("RegisterArtifactCodec() error = %v", err)
+	}
+	if spec, ok := registry.Spec(dnd.LocationOccurrenceListKind); !ok || spec.SchemaDigest != contracts.DigestArtifactSchema(schema) {
+		t.Fatalf("registered spec = %#v, %t", spec, ok)
+	}
+	first := schema.JSONSchema
+	first[0] = '['
+	if next := codec.Schema().JSONSchema; !json.Valid(next) || next[0] == '[' {
+		t.Fatal("Schema() returned shared bytes")
+	}
+	metadata := codec.Metadata(validList())
+	metadata["other"] = true
+	if next := codec.Metadata(validList()); len(next) != 1 || next["occurrence_count"] != 1 {
+		t.Fatalf("Metadata() = %#v", next)
+	}
+}
+
+func TestCodecSupportsEmptyListsAndCandidateSemanticFailures(t *testing.T) {
+	codec := New()
+	empty := dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{}}
+	if content, err := codec.Encode(empty); err != nil || string(content) != `{"occurrences":[]}` {
+		t.Fatalf("Encode() = %s, %v", content, err)
+	}
+	for _, candidate := range []dnd.LocationOccurrenceList{
+		{},
+		empty,
+		{Occurrences: []dnd.LocationOccurrence{{LocationID: "bad", Name: " ", Kind: "unsupported", SourceRefs: nil}}},
+		{Occurrences: []dnd.LocationOccurrence{{LocationID: "bad", Name: " ", Kind: "unsupported", SourceRefs: []source.SourceRef{}}}},
+		{Occurrences: []dnd.LocationOccurrence{{LocationID: "bad", Name: " ", Kind: "unsupported", SourceRefs: []source.SourceRef{{StartUnitID: 0, EndUnitID: -1}}}}},
+	} {
+		content, err := codec.EncodeCandidate(candidate)
+		if err != nil || !json.Valid(content) {
+			t.Fatalf("EncodeCandidate() = %s, %v", content, err)
+		}
+		decoded, err := codec.DecodeCandidate(content)
+		if err != nil || !reflect.DeepEqual(decoded, candidate) {
+			t.Fatalf("DecodeCandidate() = %#v, %v, want %#v", decoded, err, candidate)
+		}
+	}
+}
+
+func TestCodecRejectsStructuralJSONBeforeSemanticApproval(t *testing.T) {
+	valid := `{"occurrences":[{"location_id":"location:sha256:0000000000000000000000000000000000000000000000000000000000000000","name":"The Tavern","kind":"visited","source_refs":[{"source_id":"session","start_unit_id":1,"end_unit_id":1}]}]}`
+	for _, test := range []struct{ name, raw, want string }{
+		{"malformed", `{`, "decode dnd location occurrence list"},
+		{"unknown top-level", `{"occurrences":[],"unexpected":true}`, "unknown field"},
+		{"unknown occurrence field", strings.Replace(valid, `"kind":"visited"`, `"kind":"visited","unexpected":true`, 1), "unknown field"},
+		{"unknown source reference field", strings.Replace(valid, `"end_unit_id":1`, `"end_unit_id":1,"unexpected":true`, 1), "unknown field"},
+		{"invalid type", `{"occurrences":"not-an-array"}`, "cannot unmarshal string"},
+		{"trailing", `{"occurrences":[]} {}`, "multiple JSON values"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := New().DecodeCandidate([]byte(test.raw)); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("DecodeCandidate() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestCodecRejectsApprovedShapeEnumAndReferenceBoundaries(t *testing.T) {
+	base := validList().Occurrences[0]
+	for _, test := range []struct {
+		name  string
+		value dnd.LocationOccurrenceList
+		want  string
+	}{
+		{"nil occurrences", dnd.LocationOccurrenceList{}, "occurrences must be present"},
+		{"invalid ID", dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{{LocationID: "bad", Name: base.Name, Kind: base.Kind, SourceRefs: base.SourceRefs}}}, "location_id must match location ID pattern"},
+		{"blank name", dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{{LocationID: base.LocationID, Name: " ", Kind: base.Kind, SourceRefs: base.SourceRefs}}}, "name must not be empty"},
+		{"invalid kind", dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{{LocationID: base.LocationID, Name: base.Name, Kind: "unsupported", SourceRefs: base.SourceRefs}}}, "kind must be supported"},
+		{"empty source refs", dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{{LocationID: base.LocationID, Name: base.Name, Kind: base.Kind}}}, "source_refs must contain"},
+		{"malformed source reference", dnd.LocationOccurrenceList{Occurrences: []dnd.LocationOccurrence{{LocationID: base.LocationID, Name: base.Name, Kind: base.Kind, SourceRefs: []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 0}}}}}, "end_unit_id must be positive"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := New().Encode(test.value); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Encode() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestCodecAcceptsEveryOccurrenceKind(t *testing.T) {
+	for _, kind := range []dnd.LocationOccurrenceKind{
+		dnd.LocationOccurrenceKindVisited,
+		dnd.LocationOccurrenceKindPlanned,
+		dnd.LocationOccurrenceKindRecalled,
+		dnd.LocationOccurrenceKindMentioned,
+	} {
+		value := validList()
+		value.Occurrences[0].Kind = kind
+		if _, err := New().Encode(value); err != nil {
+			t.Fatalf("Encode(%q) error = %v", kind, err)
+		}
+	}
+}

internal/modules/dnd/codec/locationoccurrences/testdata/dnd_location_occurrences.v1.json

@@ -0,0 +1,12 @@
+{
+  "occurrences": [
+    {
+      "location_id": "location:sha256:cdd2b57615b56a4e92d506cf053cd3985e11d4df4895c5383b30ec2ce9f39ed0",
+      "name": "The Old Tavern",
+      "kind": "visited",
+      "source_refs": [
+        {"source_id": "session-alpha", "start_unit_id": 1, "end_unit_id": 2}
+      ]
+    }
+  ]
+}

internal/modules/dnd/codec/locations/assets/schemas/dnd_locations.v1.json

@@ -0,0 +1,50 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.locations",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["locations"],
+  "properties": {
+    "locations": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["id", "name", "source_refs"],
+        "properties": {
+          "id": {
+            "type": "string",
+            "pattern": "^location:sha256:[0-9a-f]{64}$"
+          },
+          "name": {
+            "type": "string",
+            "minLength": 1
+          },
+          "source_refs": {
+            "type": "array",
+            "minItems": 1,
+            "items": {
+              "type": "object",
+              "additionalProperties": false,
+              "required": ["source_id", "start_unit_id", "end_unit_id"],
+              "properties": {
+                "source_id": {
+                  "type": "string",
+                  "minLength": 1
+                },
+                "start_unit_id": {
+                  "type": "integer",
+                  "minimum": 1
+                },
+                "end_unit_id": {
+                  "type": "integer",
+                  "minimum": 1
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

internal/modules/dnd/codec/locations/codec.go

@@ -0,0 +1,111 @@
+// Package locations encodes durable D&D location artifacts.
+package locations
+
+import (
+	"embed"
+	"fmt"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/candidatejson"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+const (
+	SchemaID      = "notarius.dnd.locations"
+	SchemaName    = "notarius_dnd_locations_v1"
+	SchemaVersion = "v1"
+	MediaType     = "application/json"
+)
+
+//go:embed assets/schemas/dnd_locations.v1.json
+var schemaAssets embed.FS
+
+var _ contracts.ArtifactCodec[dnd.LocationList] = (*Codec)(nil)
+
+type Codec struct{}
+
+func New() *Codec { return &Codec{} }
+
+func (c *Codec) Kind() contracts.ArtifactKind { return dnd.LocationListKind }
+
+func (c *Codec) Schema() contracts.ArtifactSchema {
+	raw, err := schemaAssets.ReadFile("assets/schemas/dnd_locations.v1.json")
+	if err != nil {
+		return contracts.ArtifactSchema{}
+	}
+	return contracts.ArtifactSchema{
+		ID:         SchemaID,
+		Name:       SchemaName,
+		Version:    SchemaVersion,
+		JSONSchema: append([]byte(nil), raw...),
+	}
+}
+
+func (c *Codec) MediaType() string { return MediaType }
+
+func (c *Codec) Metadata(value dnd.LocationList) map[string]any {
+	return map[string]any{"location_count": len(value.Locations)}
+}
+
+func (c *Codec) Encode(value dnd.LocationList) ([]byte, error) {
+	if err := validate(value); err != nil {
+		return nil, fmt.Errorf("encode dnd location list: %w", err)
+	}
+	return c.EncodeCandidate(value)
+}
+
+// EncodeCandidate provides the durable representation before semantic
+// validators have approved a value.
+func (c *Codec) EncodeCandidate(value dnd.LocationList) ([]byte, error) {
+	return candidatejson.EncodeCandidate("dnd location list", value)
+}
+
+func (c *Codec) Decode(content []byte) (dnd.LocationList, error) {
+	value, err := c.DecodeCandidate(content)
+	if err != nil {
+		return dnd.LocationList{}, err
+	}
+	if err := validate(value); err != nil {
+		return dnd.LocationList{}, fmt.Errorf("decode dnd location list: %w", err)
+	}
+	return value, nil
+}
+
+// DecodeCandidate reads one strict durable JSON value before semantic
+// validators have approved it.
+func (c *Codec) DecodeCandidate(content []byte) (dnd.LocationList, error) {
+	return candidatejson.DecodeCandidate[dnd.LocationList]("dnd location list", content)
+}
+
+func validate(value dnd.LocationList) error {
+	if value.Locations == nil {
+		return fmt.Errorf("locations must be present")
+	}
+	for index, location := range value.Locations {
+		prefix := fmt.Sprintf("locations[%d]", index)
+		if !identity.IsValidID(location.ID) {
+			return fmt.Errorf("%s.id must match location ID pattern", prefix)
+		}
+		if strings.TrimSpace(location.Name) == "" {
+			return fmt.Errorf("%s.name must not be empty", prefix)
+		}
+		if len(location.SourceRefs) == 0 {
+			return fmt.Errorf("%s.source_refs must contain at least one reference", prefix)
+		}
+		for refIndex, ref := range location.SourceRefs {
+			refPrefix := fmt.Sprintf("%s.source_refs[%d]", prefix, refIndex)
+			if strings.TrimSpace(ref.SourceID) == "" {
+				return fmt.Errorf("%s.source_id must not be empty", refPrefix)
+			}
+			if ref.StartUnitID <= 0 {
+				return fmt.Errorf("%s.start_unit_id must be positive", refPrefix)
+			}
+			if ref.EndUnitID <= 0 {
+				return fmt.Errorf("%s.end_unit_id must be positive", refPrefix)
+			}
+		}
+	}
+	return nil
+}

internal/modules/dnd/codec/locations/codec_test.go

@@ -0,0 +1,146 @@
+package locations
+
+import (
+	"bytes"
+	"encoding/json"
+	"os"
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+func validList() dnd.LocationList {
+	refs := []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 1, EndUnitID: 2}}
+	return dnd.LocationList{Locations: []dnd.Location{{
+		ID:         identity.DeriveID("The Old Tavern", refs),
+		Name:       "The Old Tavern",
+		SourceRefs: refs,
+	}}}
+}
+
+func TestCodecMatchesMaintainedDurableFixture(t *testing.T) {
+	raw, err := os.ReadFile("testdata/dnd_locations.v1.json")
+	if err != nil {
+		t.Fatalf("read durable fixture: %v", err)
+	}
+	codec := New()
+	value, err := codec.Decode(raw)
+	if err != nil {
+		t.Fatalf("Decode() error = %v, want nil", err)
+	}
+	if want := validList(); !reflect.DeepEqual(value, want) {
+		t.Fatalf("Decode() = %#v, want %#v", value, want)
+	}
+	encoded, err := codec.Encode(value)
+	if err != nil {
+		t.Fatalf("Encode() error = %v, want nil", err)
+	}
+	var compact bytes.Buffer
+	if err := json.Compact(&compact, raw); err != nil {
+		t.Fatalf("compact durable fixture: %v", err)
+	}
+	if !bytes.Equal(encoded, compact.Bytes()) {
+		t.Fatalf("Encode() = %s, want %s", encoded, compact.Bytes())
+	}
+}
+
+func TestCodecOwnsDurableSchemaAndMetadata(t *testing.T) {
+	codec := New()
+	schema := codec.Schema()
+	if codec.Kind() != dnd.LocationListKind || codec.MediaType() != MediaType {
+		t.Fatalf("codec identity = %q/%q", codec.Kind(), codec.MediaType())
+	}
+	if schema.ID != SchemaID || schema.Name != SchemaName || schema.Version != SchemaVersion || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("schema = %#v, want durable location schema", schema)
+	}
+	var document map[string]any
+	if err := json.Unmarshal(schema.JSONSchema, &document); err != nil || document["$id"] != SchemaID {
+		t.Fatalf("durable schema document = %#v, %v", document, err)
+	}
+	registry := pipeline.NewArtifactCodecRegistry()
+	if err := pipeline.RegisterArtifactCodec(registry, codec); err != nil {
+		t.Fatalf("RegisterArtifactCodec() error = %v", err)
+	}
+	if spec, ok := registry.Spec(dnd.LocationListKind); !ok || spec.SchemaDigest != contracts.DigestArtifactSchema(schema) {
+		t.Fatalf("registered spec = %#v, %t", spec, ok)
+	}
+	first := schema.JSONSchema
+	first[0] = '['
+	if next := codec.Schema().JSONSchema; !json.Valid(next) || next[0] == '[' {
+		t.Fatal("Schema() returned shared bytes")
+	}
+	metadata := codec.Metadata(validList())
+	metadata["other"] = true
+	if next := codec.Metadata(validList()); len(next) != 1 || next["location_count"] != 1 {
+		t.Fatalf("Metadata() = %#v", next)
+	}
+}
+
+func TestCodecSupportsEmptyListsAndCandidateSemanticFailures(t *testing.T) {
+	codec := New()
+	empty := dnd.LocationList{Locations: []dnd.Location{}}
+	if content, err := codec.Encode(empty); err != nil || string(content) != `{"locations":[]}` {
+		t.Fatalf("Encode() = %s, %v", content, err)
+	}
+	for _, candidate := range []dnd.LocationList{
+		{},
+		empty,
+		{Locations: []dnd.Location{{ID: "bad", Name: " ", SourceRefs: nil}}},
+		{Locations: []dnd.Location{{ID: "bad", Name: " ", SourceRefs: []source.SourceRef{}}}},
+		{Locations: []dnd.Location{{ID: "bad", Name: " ", SourceRefs: []source.SourceRef{{StartUnitID: 0, EndUnitID: -1}}}}},
+	} {
+		content, err := codec.EncodeCandidate(candidate)
+		if err != nil || !json.Valid(content) {
+			t.Fatalf("EncodeCandidate() = %s, %v", content, err)
+		}
+		decoded, err := codec.DecodeCandidate(content)
+		if err != nil || !reflect.DeepEqual(decoded, candidate) {
+			t.Fatalf("DecodeCandidate() = %#v, %v, want %#v", decoded, err, candidate)
+		}
+	}
+}
+
+func TestCodecRejectsStructuralJSONBeforeSemanticApproval(t *testing.T) {
+	valid := `{"locations":[{"id":"location:sha256:0000000000000000000000000000000000000000000000000000000000000000","name":"The Tavern","source_refs":[{"source_id":"session","start_unit_id":1,"end_unit_id":1}]}]}`
+	for _, test := range []struct{ name, raw, want string }{
+		{"malformed", `{`, "decode dnd location list"},
+		{"unknown top-level", `{"locations":[],"unexpected":true}`, "unknown field"},
+		{"unknown location field", strings.Replace(valid, `"name":"The Tavern"`, `"name":"The Tavern","unexpected":true`, 1), "unknown field"},
+		{"unknown source reference field", strings.Replace(valid, `"end_unit_id":1`, `"end_unit_id":1,"unexpected":true`, 1), "unknown field"},
+		{"invalid type", `{"locations":"not-an-array"}`, "cannot unmarshal string"},
+		{"trailing", `{"locations":[]} {}`, "multiple JSON values"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := New().DecodeCandidate([]byte(test.raw)); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("DecodeCandidate() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestCodecRejectsApprovedShapeBoundaries(t *testing.T) {
+	base := validList().Locations[0]
+	for _, test := range []struct {
+		name  string
+		value dnd.LocationList
+		want  string
+	}{
+		{"nil locations", dnd.LocationList{}, "locations must be present"},
+		{"invalid ID", dnd.LocationList{Locations: []dnd.Location{{ID: "bad", Name: base.Name, SourceRefs: base.SourceRefs}}}, "id must match location ID pattern"},
+		{"blank name", dnd.LocationList{Locations: []dnd.Location{{ID: base.ID, Name: " ", SourceRefs: base.SourceRefs}}}, "name must not be empty"},
+		{"empty source refs", dnd.LocationList{Locations: []dnd.Location{{ID: base.ID, Name: base.Name, SourceRefs: nil}}}, "source_refs must contain"},
+		{"malformed source reference", dnd.LocationList{Locations: []dnd.Location{{ID: base.ID, Name: base.Name, SourceRefs: []source.SourceRef{{SourceID: "session", StartUnitID: 0, EndUnitID: 1}}}}}, "start_unit_id must be positive"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := New().Encode(test.value); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Encode() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}

internal/modules/dnd/codec/locations/testdata/dnd_locations.v1.json

@@ -0,0 +1,11 @@
+{
+  "locations": [
+    {
+      "id": "location:sha256:cdd2b57615b56a4e92d506cf053cd3985e11d4df4895c5383b30ec2ce9f39ed0",
+      "name": "The Old Tavern",
+      "source_refs": [
+        {"source_id": "session-alpha", "start_unit_id": 1, "end_unit_id": 2}
+      ]
+    }
+  ]
+}

internal/modules/dnd/enemyevents/enemyevents.go

@@ -0,0 +1,107 @@
+// Package enemyevents owns canonical ordering and identity policy for D&D
+// enemy-event artifacts.
+package enemyevents
+
+import (
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/npcs/identity"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+// NormalizeDisplay returns the durable display form for an enemy subject.
+func NormalizeDisplay(value string) string { return identity.NormalizeDisplay(value) }
+
+// ComparisonKey returns the shared D&D identity key for an enemy subject.
+func ComparisonKey(value string) string { return identity.ComparisonKey(NormalizeDisplay(value)) }
+
+// SupportedKind reports whether kind belongs to the durable enemy-event vocabulary.
+func SupportedKind(kind dnd.EnemyEventKind) bool {
+	_, ok := KindRank(kind)
+	return ok
+}
+
+// KindRank returns the explicit chronological tie-break order for event kinds.
+func KindRank(kind dnd.EnemyEventKind) (int, bool) {
+	switch kind {
+	case dnd.EnemyEventKindEngaged:
+		return 0, true
+	case dnd.EnemyEventKindIncapacitated:
+		return 1, true
+	case dnd.EnemyEventKindCaptured:
+		return 2, true
+	case dnd.EnemyEventKindFled:
+		return 3, true
+	case dnd.EnemyEventKindKilled:
+		return 4, true
+	default:
+		return 0, false
+	}
+}
+
+// SourceRefsEqual reports whether two reference sequences are equal after
+// canonical ordering and exact duplicate removal.
+func SourceRefsEqual(order shared.SourceRefOrder, left, right []source.SourceRef) bool {
+	left = order.Canonicalize(left)
+	right = order.Canonicalize(right)
+	if (left == nil) != (right == nil) || len(left) != len(right) {
+		return false
+	}
+	for index := range left {
+		if left[index] != right[index] {
+			return false
+		}
+	}
+	return true
+}
+
+// ExactEqual reports whether events have the same subject identity, kind, and
+// complete canonical evidence sequence.
+func ExactEqual(order shared.SourceRefOrder, left, right dnd.EnemyEvent) bool {
+	return ComparisonKey(left.Name) == ComparisonKey(right.Name) &&
+		left.Kind == right.Kind &&
+		SourceRefsEqual(order, left.SourceRefs, right.SourceRefs)
+}
+
+// Less defines the canonical stable event order. Invalid source references
+// remain comparable through SourceRefOrder's literal fallback so malformed
+// candidates can still be sorted for later validation.
+func Less(order shared.SourceRefOrder, left, right dnd.EnemyEvent) bool {
+	leftPosition, leftHasEvidence := order.EarliestValid(left.SourceRefs)
+	rightPosition, rightHasEvidence := order.EarliestValid(right.SourceRefs)
+	if leftHasEvidence != rightHasEvidence {
+		return leftHasEvidence
+	}
+	if leftHasEvidence && leftPosition != rightPosition {
+		return leftPosition < rightPosition
+	}
+	if leftKey, rightKey := ComparisonKey(left.Name), ComparisonKey(right.Name); leftKey != rightKey {
+		return leftKey < rightKey
+	}
+	if leftName, rightName := NormalizeDisplay(left.Name), NormalizeDisplay(right.Name); leftName != rightName {
+		return leftName < rightName
+	}
+	if leftRank, leftKnown := KindRank(left.Kind); leftKnown {
+		if rightRank, rightKnown := KindRank(right.Kind); rightKnown && leftRank != rightRank {
+			return leftRank < rightRank
+		} else if !rightKnown {
+			return true
+		}
+	} else if _, rightKnown := KindRank(right.Kind); rightKnown {
+		return false
+	}
+	if left.Kind != right.Kind {
+		return left.Kind < right.Kind
+	}
+	return sourceRefsLess(order, order.Canonicalize(left.SourceRefs), order.Canonicalize(right.SourceRefs))
+}
+
+func sourceRefsLess(order shared.SourceRefOrder, left, right []source.SourceRef) bool {
+	for index := 0; index < len(left) && index < len(right); index++ {
+		if left[index] == right[index] {
+			continue
+		}
+		return order.Less(left[index], right[index])
+	}
+	return len(left) < len(right)
+}

internal/modules/dnd/enemyevents/enemyevents_test.go

@@ -0,0 +1,90 @@
+package enemyevents
+
+import (
+	"sort"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+func TestSupportedKindAndRank(t *testing.T) {
+	tests := []struct {
+		kind dnd.EnemyEventKind
+		rank int
+	}{
+		{dnd.EnemyEventKindEngaged, 0},
+		{dnd.EnemyEventKindIncapacitated, 1},
+		{dnd.EnemyEventKindCaptured, 2},
+		{dnd.EnemyEventKindFled, 3},
+		{dnd.EnemyEventKindKilled, 4},
+	}
+	for _, test := range tests {
+		if rank, ok := KindRank(test.kind); !ok || rank != test.rank || !SupportedKind(test.kind) {
+			t.Fatalf("kind %q = (%d, %t), supported %t", test.kind, rank, ok, SupportedKind(test.kind))
+		}
+	}
+	if _, ok := KindRank("unknown"); ok || SupportedKind("unknown") {
+		t.Fatal("unsupported kind was accepted")
+	}
+}
+
+func TestLessOrdersChronologyThenKind(t *testing.T) {
+	order := testOrder()
+	events := []dnd.EnemyEvent{
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindKilled, SourceRefs: refs(20)},
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindFled, SourceRefs: refs(20)},
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindCaptured, SourceRefs: refs(20)},
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindIncapacitated, SourceRefs: refs(20)},
+		{Name: "Ashfang", Kind: dnd.EnemyEventKindEngaged, SourceRefs: refs(20)},
+		{Name: "Later", Kind: dnd.EnemyEventKindEngaged, SourceRefs: refs(30)},
+		{Name: "Earlier", Kind: dnd.EnemyEventKindKilled, SourceRefs: refs(10)},
+	}
+	sort.SliceStable(events, func(left, right int) bool { return Less(order, events[left], events[right]) })
+	want := []dnd.EnemyEventKind{
+		dnd.EnemyEventKindKilled,
+		dnd.EnemyEventKindEngaged,
+		dnd.EnemyEventKindIncapacitated,
+		dnd.EnemyEventKindCaptured,
+		dnd.EnemyEventKindFled,
+		dnd.EnemyEventKindKilled,
+		dnd.EnemyEventKindEngaged,
+	}
+	for index, kind := range want {
+		if events[index].Kind != kind {
+			t.Fatalf("event %d kind = %q, want %q", index, events[index].Kind, kind)
+		}
+	}
+}
+
+func TestExactEqualUsesSubjectIdentityAndCanonicalEvidence(t *testing.T) {
+	order := testOrder()
+	first := dnd.EnemyEvent{Name: "  ASHFANG ", Kind: dnd.EnemyEventKindFled, SourceRefs: []source.SourceRef{
+		{SourceID: "session", StartUnitID: 30, EndUnitID: 30},
+		{SourceID: "session", StartUnitID: 10, EndUnitID: 10},
+		{SourceID: "session", StartUnitID: 10, EndUnitID: 10},
+	}}
+	second := dnd.EnemyEvent{Name: "Ashfang", Kind: dnd.EnemyEventKindFled, SourceRefs: []source.SourceRef{
+		{SourceID: "session", StartUnitID: 10, EndUnitID: 10},
+		{SourceID: "session", StartUnitID: 30, EndUnitID: 30},
+	}}
+	if !ExactEqual(order, first, second) || !SourceRefsEqual(order, first.SourceRefs, second.SourceRefs) {
+		t.Fatal("canonical duplicate identity was not recognized")
+	}
+	differentKind := second
+	differentKind.Kind = dnd.EnemyEventKindKilled
+	differentEvidence := second
+	differentEvidence.SourceRefs = []source.SourceRef{{SourceID: "session", StartUnitID: 20, EndUnitID: 20}}
+	if ExactEqual(order, first, differentKind) || ExactEqual(order, first, differentEvidence) {
+		t.Fatal("distinct event was treated as an exact duplicate")
+	}
+}
+
+func refs(unitID int) []source.SourceRef {
+	return []source.SourceRef{{SourceID: "session", StartUnitID: unitID, EndUnitID: unitID}}
+}
+
+func testOrder() shared.SourceRefOrder {
+	return shared.NewSourceRefOrder(&source.SourceDocument{ID: "session", Units: []source.SourceUnit{{ID: 10}, {ID: 20}, {ID: 30}}})
+}

internal/modules/dnd/extract/enemyevents/assets.go

@@ -0,0 +1,6 @@
+package enemyevents
+
+import "embed"
+
+//go:embed assets/schemas/dnd_enemy_events_llm.v1.json assets/prompts/*.yaml assets/prompts/*.md
+var embeddedAssets embed.FS

internal/modules/dnd/extract/enemyevents/assets/prompts/dnd.enemy_events.yaml

@@ -0,0 +1,55 @@
+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: npcs
+    required: true
+    content_type: application/json
+  - name: combat_turns
+    required: true
+    content_type: application/json
+  - name: npc_interactions
+    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.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./sharedassets/common-dnd-npcs.md
+  - role: user
+    content_file: ./grounding.md
+  - role: user
+    content_file: ./task.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

internal/modules/dnd/extract/enemyevents/assets/prompts/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_interactions" }}

internal/modules/dnd/extract/enemyevents/assets/prompts/instructions.md

@@ -0,0 +1,14 @@
+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, a scene ending, combat-opponent grounding, or any auxiliary
+artifact. Auxiliary inputs can guide attention but cannot prove or supply an
+event. Cite only narrow current-transcript ranges that establish each event.
+
+Return the `events` array even when no enemy event is established. Every event
+must contain only `name`, `kind`, and `source_refs`. Use exactly one kind:
+`engaged`, `killed`, `fled`, `captured`, or `incapacitated`. Each source range
+uses integer `start_unit_id` and `end_unit_id`; omit `source_id` because
+Notarius assigns the current source identity.

internal/modules/dnd/extract/enemyevents/assets/prompts/task.md

@@ -0,0 +1,20 @@
+Extract Dungeons & Dragons enemy events from the supplied combat transcript.
+
+Return an `engaged` event only when the transcript directly establishes that a
+subject is actively opposing the party in combat. Return `killed`, `fled`,
+`captured`, or `incapacitated` only when the transcript explicitly establishes
+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 engagement 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.
+
+Use a normalized NPC registry spelling when the transcript identifies that
+named NPC. 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.

internal/modules/dnd/extract/enemyevents/assets/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"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

internal/modules/dnd/extract/enemyevents/extractor.go

@@ -0,0 +1,210 @@
+package enemyevents
+
+import (
+	"context"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	sceneregistry "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/scenedescriptions/registry"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const (
+	Key             = "dnd/enemy-events"
+	mappingPolicy   = "dnd.enemy_events.extract_mapping.v1"
+	sceneGatePolicy = "dnd.enemy_events.scene_gate.v1"
+)
+
+var requiredCapabilities = []string{
+	"chunks",
+	"source.transcript",
+}
+
+var providedCapabilities = []string{
+	"dnd.enemy_events",
+}
+
+var _ contracts.Extractor[dnd.EnemyEventList] = (*Extractor)(nil)
+var _ contracts.ManifestMetadataProvider = (*Extractor)(nil)
+var _ pipeline.CheckpointFingerprintProvider = (*Extractor)(nil)
+
+type Options struct{}
+
+type Extractor struct {
+	llm       contracts.StructuredLLMClient
+	grounding *groundingResolver
+
+	promptSHA         string
+	responseSchemaSHA string
+}
+
+func New(llmClient contracts.StructuredLLMClient, _ Options, references ...contracts.ReferenceSet) (*Extractor, error) {
+	if llmClient == nil {
+		return nil, extractorErrorf("LLM client must not be nil")
+	}
+	if len(references) > 1 {
+		return nil, extractorErrorf("at most one reference set may be supplied")
+	}
+	var referenceSet contracts.ReferenceSet
+	if len(references) == 1 {
+		referenceSet = references[0]
+	}
+	grounding, err := newGroundingResolver(referenceSet)
+	if err != nil {
+		return nil, extractorErrorf("prepare grounding: %w", err)
+	}
+	promptSHA, err := promptAssetMetadata()
+	if err != nil {
+		return nil, extractorErrorf("load prompt metadata: %w", err)
+	}
+	responseSchema, err := loadResponseSchema()
+	if err != nil {
+		return nil, extractorErrorf("load response schema: %w", err)
+	}
+	return &Extractor{
+		llm:               llmClient,
+		grounding:         grounding,
+		promptSHA:         promptSHA,
+		responseSchemaSHA: responseSchema.SHA256,
+	}, nil
+}
+
+func (e *Extractor) Key() string { return Key }
+
+func (e *Extractor) ReferenceSlots() []contracts.ReferenceSlot { return referenceSlots() }
+
+func (e *Extractor) ManifestMetadata() map[string]any {
+	if e == nil {
+		return nil
+	}
+	return map[string]any{
+		"prompt_id":               PromptID,
+		"prompt_version":          SchemaVersion,
+		"prompt_sha256":           e.promptSHA,
+		"mapping_policy":          mappingPolicy,
+		"scene_gate_policy":       sceneGatePolicy,
+		"response_schema_key":     string(ResponseSchemaKey),
+		"response_schema_id":      ResponseSchemaID,
+		"response_schema_name":    ResponseSchemaName,
+		"response_schema_version": SchemaVersion,
+		"response_schema_sha256":  e.responseSchemaSHA,
+	}
+}
+
+func (e *Extractor) CheckpointFingerprints() []pipeline.CheckpointFingerprint {
+	if e == nil {
+		return nil
+	}
+	return []pipeline.CheckpointFingerprint{
+		{Name: "prompt", Value: e.promptSHA},
+		{Name: "response_schema", Value: e.responseSchemaSHA},
+		{Name: "mapping_policy", Value: mappingPolicy},
+		{Name: "scene_gate_policy", Value: sceneGatePolicy},
+	}
+}
+
+func (e *Extractor) Extract(ctx context.Context, req contracts.TypedExtractionRequest) (contracts.TypedExtractionResult[dnd.EnemyEventList], error) {
+	if e == nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("extractor must not be nil")
+	}
+	if e.llm == nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("LLM client must not be nil")
+	}
+	if e.grounding == nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("grounding resolver must not be nil")
+	}
+	sourceInput, err := shared.PrepareChunkExtraction(ctx, req)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("%w", err)
+	}
+	match, err := e.grounding.SceneMatch(req.References, req.Chunk)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("resolve scene eligibility: %w", err)
+	}
+	switch match.State {
+	case sceneregistry.MatchExact:
+		if match.Kind != dnd.SceneKindCombat {
+			return emptyResult(), nil
+		}
+	case sceneregistry.MatchMissing, sceneregistry.MatchMismatched:
+		return unavailableSceneResult(), nil
+	default:
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("unsupported scene eligibility match state %q", match.State)
+	}
+	grounding, err := e.grounding.Resolve(req.References)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("resolve enemy-event grounding: %w", err)
+	}
+	inputs := shared.PromptInputs(sourceInput, req.References)
+	for name, input := range grounding.PromptInputs() {
+		inputs[name] = input
+	}
+	var response extractionResponse
+	if _, err := e.llm.CompleteStructured(ctx, contracts.StructuredCompletionRequest{
+		StageName:     Key,
+		PromptID:      PromptID,
+		PromptVersion: SchemaVersion,
+		ProfileID:     req.LLMProfile,
+		SessionID:     req.SessionID,
+		Inputs:        inputs,
+	}, &response); err != nil {
+		return contracts.TypedExtractionResult[dnd.EnemyEventList]{}, extractorErrorf("complete structured output: %w", err)
+	}
+	return contracts.TypedExtractionResult[dnd.EnemyEventList]{
+		Value: canonicalEnemyEventList(response, shared.NewSourceRefOrder(req.Source), req.Source.ID),
+	}, nil
+}
+
+func emptyResult() contracts.TypedExtractionResult[dnd.EnemyEventList] {
+	return contracts.TypedExtractionResult[dnd.EnemyEventList]{Value: dnd.EnemyEventList{Events: []dnd.EnemyEvent{}}}
+}
+
+func unavailableSceneResult() contracts.TypedExtractionResult[dnd.EnemyEventList] {
+	result := emptyResult()
+	result.Warnings = []contracts.Warning{{
+		Scope:      SceneDescriptionReferenceSlot,
+		ReasonCode: "scene_classification_unavailable",
+		Message:    "No exact scene classification was available; enemy-event extraction was skipped.",
+	}}
+	return result
+}
+
+func ModuleSpec() pipeline.ModuleSpec {
+	return pipeline.ModuleSpec{
+		Key:            Key,
+		Stage:          pipeline.StageExtract,
+		ExecutionClass: contracts.ExecutionClassLLMBacked,
+		Requires:       append([]string(nil), requiredCapabilities...),
+		Provides:       append([]string(nil), providedCapabilities...),
+		ArtifactKind:   dnd.EnemyEventListKind,
+		ReferenceSlots: referenceSlots(),
+	}
+}
+
+func Register(registry *pipeline.ExtractorRegistry) error {
+	return pipeline.RegisterExtractorBuilder(registry, ModuleSpec(), validateOptions, func(request pipeline.BuildRequest) (contracts.Extractor[dnd.EnemyEventList], error) {
+		options, err := DecodeOptions(request.Options)
+		if err != nil {
+			return nil, err
+		}
+		return New(request.Dependencies.LLM, options, request.References)
+	})
+}
+
+func validateOptions(options map[string]any) error {
+	_, err := DecodeOptions(options)
+	return err
+}
+
+func DecodeOptions(options map[string]any) (Options, error) {
+	if err := pipeline.RejectUnknownOptions(options); err != nil {
+		return Options{}, extractorErrorf("%w", err)
+	}
+	return Options{}, nil
+}
+
+func extractorErrorf(format string, args ...any) error {
+	return fmt.Errorf("dnd enemy events extractor: "+format, args...)
+}

internal/modules/dnd/extract/enemyevents/extractor_test.go

@@ -0,0 +1,229 @@
+package enemyevents
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+)
+
+func TestExtractMapsEnemyEventsInSourceOrder(t *testing.T) {
+	client := &fakeEnemyEventsLLMClient{response: extractionResponse{Events: []enemyEventResponse{
+		{Name: "Ashfang", Kind: "killed", SourceRefs: []enemySourceRefResponse{{StartUnitID: 4, EndUnitID: 4}}},
+		{Name: "Ashfang", Kind: "engaged", SourceRefs: []enemySourceRefResponse{{StartUnitID: 1, EndUnitID: 1}, {StartUnitID: 1, EndUnitID: 1}}},
+		{Name: "Orcs", Kind: "fled", SourceRefs: []enemySourceRefResponse{{StartUnitID: 3, EndUnitID: 3}}},
+		{Name: "One orc", Kind: "captured", SourceRefs: []enemySourceRefResponse{{StartUnitID: 2, EndUnitID: 2}}},
+		{Name: "Ashfang", Kind: "incapacitated", SourceRefs: []enemySourceRefResponse{{StartUnitID: 2, EndUnitID: 2}}},
+	}}}
+
+	result, err := newEnemyExtractor(t, client).Extract(context.Background(), enemyExtractionRequest(t))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if got := []dnd.EnemyEventKind{result.Value.Events[0].Kind, result.Value.Events[1].Kind, result.Value.Events[2].Kind, result.Value.Events[3].Kind, result.Value.Events[4].Kind}; !reflect.DeepEqual(got, []dnd.EnemyEventKind{
+		dnd.EnemyEventKindEngaged,
+		dnd.EnemyEventKindCaptured,
+		dnd.EnemyEventKindIncapacitated,
+		dnd.EnemyEventKindFled,
+		dnd.EnemyEventKindKilled,
+	}) {
+		t.Fatalf("event order = %#v", got)
+	}
+	if refs := result.Value.Events[0].SourceRefs; !reflect.DeepEqual(refs, []source.SourceRef{{SourceID: "combat-session", StartUnitID: 1, EndUnitID: 1}}) {
+		t.Fatalf("canonical source refs = %#v", refs)
+	}
+	if len(client.requests) != 1 {
+		t.Fatalf("LLM calls = %d, want 1", len(client.requests))
+	}
+	request := client.requests[0]
+	if request.StageName != Key || request.PromptID != PromptID || request.PromptVersion != SchemaVersion || request.ProfileID != "enemy-profile" || request.SessionID != "session-123" {
+		t.Fatalf("LLM request identity = %#v", request)
+	}
+	if got := string(request.Inputs[CombatTurnReferenceSlot].Content); !strings.Contains(got, `"actor":"Ashfang"`) || strings.Contains(got, "source_ref") {
+		t.Fatalf("combat grounding = %s", got)
+	}
+	if got := string(request.Inputs[NPCInteractionReferenceSlot].Content); !strings.Contains(got, `"kind":"combat_opponent"`) || strings.Contains(got, "Aria") {
+		t.Fatalf("interaction grounding = %s", got)
+	}
+}
+
+func TestExtractPreservesSemanticCandidatesAndResponseOwnership(t *testing.T) {
+	client := &fakeEnemyEventsLLMClient{response: extractionResponse{Events: []enemyEventResponse{{
+		Name: " ", Kind: "unsupported", SourceRefs: []enemySourceRefResponse{{StartUnitID: 99, EndUnitID: 0}},
+	}}}}
+	result, err := newEnemyExtractor(t, client).Extract(context.Background(), enemyExtractionRequest(t))
+	if err != nil {
+		t.Fatal(err)
+	}
+	event := result.Value.Events[0]
+	if event.Name != " " || event.Kind != "unsupported" || event.SourceRefs[0] != (source.SourceRef{SourceID: "combat-session", StartUnitID: 99}) {
+		t.Fatalf("semantic candidate = %#v", event)
+	}
+	result.Value.Events[0].SourceRefs[0].StartUnitID = 7
+	if client.response.Events[0].SourceRefs[0].StartUnitID != 99 {
+		t.Fatal("mapped event aliases model-owned source references")
+	}
+}
+
+func TestExtractSkipsModelForIneligibleScenes(t *testing.T) {
+	for _, test := range []struct {
+		name        string
+		kind        dnd.SceneKind
+		change      func(*source.Chunk)
+		wantWarning bool
+	}{
+		{name: "non-combat", kind: dnd.SceneKindNarrative},
+		{name: "missing", kind: dnd.SceneKindCombat, change: func(chunk *source.Chunk) { chunk.ID = "other" }, wantWarning: true},
+		{name: "mismatched", kind: dnd.SceneKindCombat, change: func(chunk *source.Chunk) { chunk.Ref.EndUnitID++ }, wantWarning: true},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			client := &fakeEnemyEventsLLMClient{}
+			request := enemyExtractionRequest(t)
+			request.References = groundingReferences(t, "Ashfang", test.kind)
+			if test.change != nil {
+				test.change(request.Chunk)
+			}
+			result, err := newEnemyExtractor(t, client).Extract(context.Background(), request)
+			if err != nil {
+				t.Fatal(err)
+			}
+			if len(client.requests) != 0 || result.Value.Events == nil || len(result.Value.Events) != 0 {
+				t.Fatalf("result = %#v, calls = %d", result, len(client.requests))
+			}
+			if test.wantWarning {
+				if len(result.Warnings) != 1 || result.Warnings[0].ReasonCode != "scene_classification_unavailable" || result.Warnings[0].Scope != SceneDescriptionReferenceSlot {
+					t.Fatalf("warnings = %#v", result.Warnings)
+				}
+			} else if len(result.Warnings) != 0 {
+				t.Fatalf("warnings = %#v", result.Warnings)
+			}
+		})
+	}
+}
+
+func TestExtractReportsRequiredGroundingAndProviderFailures(t *testing.T) {
+	client := &fakeEnemyEventsLLMClient{}
+	request := enemyExtractionRequest(t)
+	request.References = withoutSlot(request.References, NPCRegistryReferenceSlot)
+	if _, err := (&Extractor{llm: client, grounding: mustGroundingResolver(t, request.References)}).Extract(context.Background(), request); err == nil || !strings.Contains(err.Error(), "NPC registry") {
+		t.Fatalf("Extract() error = %v, want required grounding context", err)
+	}
+
+	provider := &fakeEnemyEventsLLMClient{err: errors.New("provider unavailable")}
+	if _, err := newEnemyExtractor(t, provider).Extract(context.Background(), enemyExtractionRequest(t)); err == nil || !strings.Contains(err.Error(), "complete structured output") || !strings.Contains(err.Error(), "provider unavailable") {
+		t.Fatalf("Extract() error = %v, want provider context", err)
+	}
+}
+
+func TestConstructorSpecOptionsAndSafeMetadata(t *testing.T) {
+	if _, err := New(nil, Options{}); err == nil || !strings.Contains(err.Error(), "LLM client") {
+		t.Fatalf("New(nil) error = %v", err)
+	}
+	if _, err := New(&fakeEnemyEventsLLMClient{}, Options{}, contracts.ReferenceSet{}, contracts.ReferenceSet{}); err == nil || !strings.Contains(err.Error(), "at most one") {
+		t.Fatalf("New() error = %v", err)
+	}
+	if _, err := DecodeOptions(map[string]any{"unexpected": true}); err == nil || !strings.Contains(err.Error(), "unknown option") {
+		t.Fatalf("DecodeOptions() error = %v", err)
+	}
+
+	first := ModuleSpec()
+	first.Requires[0] = "changed"
+	first.ReferenceSlots[0].AcceptedMediaTypes[0] = "changed"
+	second := ModuleSpec()
+	if second.Requires[0] != "chunks" || second.ArtifactKind != dnd.EnemyEventListKind || second.ExecutionClass != contracts.ExecutionClassLLMBacked || second.ReferenceSlots[0].AcceptedMediaTypes[0] == "changed" {
+		t.Fatalf("ModuleSpec() reused mutable state: %#v", second)
+	}
+	registry := pipeline.NewExtractorRegistry()
+	if err := Register(registry); err != nil {
+		t.Fatal(err)
+	}
+	if spec, ok := registry.Spec(Key); !ok || spec.Key != Key || spec.ArtifactKind != dnd.EnemyEventListKind {
+		t.Fatalf("registered spec = %#v, %t", spec, ok)
+	}
+
+	extractor := newEnemyExtractor(t, &fakeEnemyEventsLLMClient{}, groundingReferences(t, "Ashfang", dnd.SceneKindCombat))
+	metadata := extractor.ManifestMetadata()
+	encoded, err := json.Marshal(metadata)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if strings.Contains(string(encoded), "Ashfang") || metadata["mapping_policy"] != mappingPolicy || metadata["scene_gate_policy"] != sceneGatePolicy {
+		t.Fatalf("unsafe or incomplete metadata = %s", encoded)
+	}
+	fingerprints := extractor.CheckpointFingerprints()
+	if len(fingerprints) != 4 || fingerprints[0].Value != metadata["prompt_sha256"] || fingerprints[1].Value != metadata["response_schema_sha256"] || fingerprints[2].Value != mappingPolicy || fingerprints[3].Value != sceneGatePolicy {
+		t.Fatalf("fingerprints = %#v", fingerprints)
+	}
+}
+
+func newEnemyExtractor(t *testing.T, client contracts.StructuredLLMClient, references ...contracts.ReferenceSet) *Extractor {
+	t.Helper()
+	if len(references) == 0 {
+		references = []contracts.ReferenceSet{groundingReferences(t, "Ashfang", dnd.SceneKindCombat)}
+	}
+	extractor, err := New(client, Options{}, references...)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return extractor
+}
+
+func mustGroundingResolver(t *testing.T, references contracts.ReferenceSet) *groundingResolver {
+	t.Helper()
+	resolver, err := newGroundingResolver(references)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return resolver
+}
+
+func enemyExtractionRequest(t *testing.T) contracts.TypedExtractionRequest {
+	t.Helper()
+	chunk := combatChunk()
+	chunk.Units = []source.SourceUnit{{ID: 1}, {ID: 2}, {ID: 3}, {ID: 4}}
+	chunk.MediaType = "application/json"
+	chunk.Content = []byte(`{"id":"combat-scene","units":[1,2,3,4]}`)
+	return contracts.TypedExtractionRequest{
+		Source:      &source.SourceDocument{ID: "combat-session", Units: append([]source.SourceUnit(nil), chunk.Units...)},
+		Chunk:       chunk,
+		SourceInput: contracts.NewLLMInputMaterial("source", "application/json", chunk.Content, digest(chunk.Content), "file:///combat-session.json"),
+		References:  groundingReferences(t, "Ashfang", dnd.SceneKindCombat),
+		LLMProfile:  "enemy-profile",
+		SessionID:   "session-123",
+	}
+}
+
+type fakeEnemyEventsLLMClient struct {
+	response extractionResponse
+	err      error
+	requests []contracts.StructuredCompletionRequest
+}
+
+func (client *fakeEnemyEventsLLMClient) CompleteStructured(_ context.Context, request contracts.StructuredCompletionRequest, out any) (contracts.StructuredCompletionResponse, error) {
+	client.requests = append(client.requests, cloneEnemyRequest(request))
+	if client.err != nil {
+		return contracts.StructuredCompletionResponse{}, client.err
+	}
+	target, ok := out.(*extractionResponse)
+	if !ok {
+		return contracts.StructuredCompletionResponse{}, errors.New("unexpected output target")
+	}
+	*target = client.response
+	content, err := json.Marshal(client.response)
+	if err != nil {
+		return contracts.StructuredCompletionResponse{}, err
+	}
+	return contracts.StructuredCompletionResponse{Content: content}, nil
+}
+
+func cloneEnemyRequest(request contracts.StructuredCompletionRequest) contracts.StructuredCompletionRequest {
+	request.Inputs = request.Inputs.Clone()
+	return request
+}

internal/modules/dnd/extract/enemyevents/grounding.go

@@ -0,0 +1,293 @@
+// Package enemyevents prepares validated D&D combat grounding for enemy-event
+// extraction.
+package enemyevents
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"encoding/json"
+	"fmt"
+	"mime"
+	"sort"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	combatturncodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/combatturns"
+	interactioncodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/npcinteractions"
+	npccodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/npcs"
+	scenecodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/scenedescriptions"
+	npcregistry "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/npcs/registry"
+	sceneregistry "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/scenedescriptions/registry"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const (
+	NPCRegistryReferenceSlot      = npcregistry.ReferenceSlot
+	SceneDescriptionReferenceSlot = sceneregistry.ReferenceSlot
+	CombatTurnReferenceSlot       = "combat_turns"
+	NPCInteractionReferenceSlot   = "npc_interactions"
+	ReferenceMaxBytes             = 1048576
+	promptProjectionMediaType     = "application/json"
+)
+
+var referenceSlotDescriptions = shared.ReferenceSlotDescriptions{
+	Glossary: "Optional campaign glossary reference material used only for disambiguation.",
+	Party:    "Optional party roster reference material used only for disambiguation.",
+	Players:  "Optional player list reference material used only for disambiguation.",
+	Roster:   "Deprecated alias for party roster reference material used only for disambiguation.",
+}
+
+func referenceSlots() []contracts.ReferenceSlot {
+	slots := shared.ReferenceSlots(referenceSlotDescriptions)
+	slots = append(slots,
+		contracts.ReferenceSlot{
+			Name:                  NPCRegistryReferenceSlot,
+			Description:           "Required normalized NPC registry used only for enemy-subject grounding, never as event evidence.",
+			Required:              true,
+			AcceptedMediaTypes:    []string{npccodec.MediaType},
+			AcceptedArtifactKinds: []contracts.ArtifactKind{dnd.NPCListKind},
+			MaxBytes:              ReferenceMaxBytes,
+		},
+		contracts.ReferenceSlot{
+			Name:                  SceneDescriptionReferenceSlot,
+			Description:           "Required scene descriptions used only to determine exact combat eligibility, never as event evidence.",
+			Required:              true,
+			AcceptedMediaTypes:    []string{scenecodec.MediaType},
+			AcceptedArtifactKinds: []contracts.ArtifactKind{dnd.SceneDescriptionListKind},
+			MaxBytes:              ReferenceMaxBytes,
+		},
+		contracts.ReferenceSlot{
+			Name:                  CombatTurnReferenceSlot,
+			Description:           "Required combat-turn artifact used only as source-free enemy-event grounding.",
+			Required:              true,
+			AcceptedMediaTypes:    []string{combatturncodec.MediaType},
+			AcceptedArtifactKinds: []contracts.ArtifactKind{dnd.CombatTurnListKind},
+			MaxBytes:              ReferenceMaxBytes,
+		},
+		contracts.ReferenceSlot{
+			Name:                  NPCInteractionReferenceSlot,
+			Description:           "Required NPC-interaction artifact used only as source-free enemy-event grounding.",
+			Required:              true,
+			AcceptedMediaTypes:    []string{interactioncodec.MediaType},
+			AcceptedArtifactKinds: []contracts.ArtifactKind{dnd.NPCInteractionListKind},
+			MaxBytes:              ReferenceMaxBytes,
+		},
+	)
+	sort.Slice(slots, func(left, right int) bool { return slots[left].Name < slots[right].Name })
+	return contracts.CloneReferenceSlots(slots)
+}
+
+// groundingResolver retains only validated, compact construction-time views.
+// Per-operation generated references are decoded when supplied and never
+// become static metadata.
+type groundingResolver struct {
+	npcs   *npcregistry.Resolver
+	scenes *sceneregistry.Resolver
+
+	combatTurns     *contracts.LLMInputMaterial
+	npcInteractions *contracts.LLMInputMaterial
+}
+
+type grounding struct {
+	npcInput            contracts.LLMInputMaterial
+	combatTurnInput     contracts.LLMInputMaterial
+	npcInteractionInput contracts.LLMInputMaterial
+}
+
+func newGroundingResolver(references contracts.ReferenceSet) (*groundingResolver, error) {
+	npcs, err := npcregistry.NewResolver(references)
+	if err != nil {
+		return nil, fmt.Errorf("prepare NPC registry grounding: %w", err)
+	}
+	scenes, err := sceneregistry.NewResolver(references)
+	if err != nil {
+		return nil, fmt.Errorf("prepare scene eligibility: %w", err)
+	}
+	combatTurns, err := prepareCombatTurnInput(references)
+	if err != nil {
+		return nil, err
+	}
+	npcInteractions, err := prepareNPCInteractionInput(references)
+	if err != nil {
+		return nil, err
+	}
+	return &groundingResolver{
+		npcs:            npcs,
+		scenes:          scenes,
+		combatTurns:     combatTurns,
+		npcInteractions: npcInteractions,
+	}, nil
+}
+
+func (r *groundingResolver) Resolve(references contracts.ReferenceSet) (grounding, error) {
+	if r == nil {
+		return grounding{}, fmt.Errorf("grounding resolver must not be nil")
+	}
+	npcs, err := r.npcs.Resolve(references)
+	if err != nil {
+		return grounding{}, fmt.Errorf("resolve NPC registry grounding: %w", err)
+	}
+	if !npcs.Bound() {
+		return grounding{}, fmt.Errorf("NPC registry reference is required")
+	}
+	combatTurns, err := resolveInput(references, CombatTurnReferenceSlot, r.combatTurns, prepareCombatTurnInput)
+	if err != nil {
+		return grounding{}, err
+	}
+	npcInteractions, err := resolveInput(references, NPCInteractionReferenceSlot, r.npcInteractions, prepareNPCInteractionInput)
+	if err != nil {
+		return grounding{}, err
+	}
+	return grounding{
+		npcInput:            npcs.PromptInput(),
+		combatTurnInput:     combatTurns,
+		npcInteractionInput: npcInteractions,
+	}, nil
+}
+
+func (r *groundingResolver) SceneMatch(references contracts.ReferenceSet, chunk *source.Chunk) (sceneregistry.ChunkMatch, error) {
+	if r == nil {
+		return sceneregistry.ChunkMatch{}, fmt.Errorf("grounding resolver must not be nil")
+	}
+	scenes, err := r.resolveScenes(references)
+	if err != nil {
+		return sceneregistry.ChunkMatch{}, err
+	}
+	return scenes.Match(chunk), nil
+}
+
+func (r *groundingResolver) resolveScenes(references contracts.ReferenceSet) (*sceneregistry.Registry, error) {
+	scenes, err := r.scenes.Resolve(references)
+	if err != nil {
+		return nil, fmt.Errorf("resolve scene eligibility: %w", err)
+	}
+	if !scenes.Bound() {
+		return nil, fmt.Errorf("scene descriptions reference is required")
+	}
+	return scenes, nil
+}
+
+func resolveInput(references contracts.ReferenceSet, slot string, seeded *contracts.LLMInputMaterial, prepare func(contracts.ReferenceSet) (*contracts.LLMInputMaterial, error)) (contracts.LLMInputMaterial, error) {
+	if _, ok := references.Slots[slot]; ok {
+		value, err := prepare(references)
+		if err != nil {
+			return contracts.LLMInputMaterial{}, err
+		}
+		if value == nil {
+			return contracts.LLMInputMaterial{}, fmt.Errorf("%s reference is required", slot)
+		}
+		return value.Clone(), nil
+	}
+	if seeded == nil {
+		return contracts.LLMInputMaterial{}, fmt.Errorf("%s reference is required", slot)
+	}
+	return seeded.Clone(), nil
+}
+
+func (g grounding) PromptInputs() contracts.LLMInputSet {
+	return contracts.LLMInputSet{
+		NPCRegistryReferenceSlot:    g.npcInput.Clone(),
+		CombatTurnReferenceSlot:     g.combatTurnInput.Clone(),
+		NPCInteractionReferenceSlot: g.npcInteractionInput.Clone(),
+	}
+}
+
+func prepareCombatTurnInput(references contracts.ReferenceSet) (*contracts.LLMInputMaterial, error) {
+	item, ok, err := referenceItem(references, CombatTurnReferenceSlot, combatturncodec.MediaType)
+	if err != nil || !ok {
+		return nil, err
+	}
+	value, err := combatturncodec.New().Decode(item.Content)
+	if err != nil {
+		return nil, fmt.Errorf("decode combat-turn grounding: invalid approved combat-turn JSON")
+	}
+	content, err := json.Marshal(struct {
+		CombatTurns []combatTurnProjection `json:"combat_turns"`
+	}{CombatTurns: projectCombatTurns(value.CombatTurns)})
+	if err != nil {
+		return nil, fmt.Errorf("encode combat-turn grounding: %w", err)
+	}
+	return newPromptInput(CombatTurnReferenceSlot, content), nil
+}
+
+func prepareNPCInteractionInput(references contracts.ReferenceSet) (*contracts.LLMInputMaterial, error) {
+	item, ok, err := referenceItem(references, NPCInteractionReferenceSlot, interactioncodec.MediaType)
+	if err != nil || !ok {
+		return nil, err
+	}
+	value, err := interactioncodec.New().Decode(item.Content)
+	if err != nil {
+		return nil, fmt.Errorf("decode NPC-interaction grounding: invalid approved NPC-interaction JSON")
+	}
+	content, err := json.Marshal(struct {
+		Interactions []npcInteractionProjection `json:"npc_interactions"`
+	}{Interactions: projectNPCInteractions(value.Interactions)})
+	if err != nil {
+		return nil, fmt.Errorf("encode NPC-interaction grounding: %w", err)
+	}
+	return newPromptInput(NPCInteractionReferenceSlot, content), nil
+}
+
+func referenceItem(references contracts.ReferenceSet, slotName, expectedMediaType string) (contracts.ReferenceItem, bool, error) {
+	slot, ok := references.Slots[slotName]
+	if !ok {
+		return contracts.ReferenceItem{}, false, nil
+	}
+	if len(slot.Items) == 0 {
+		return contracts.ReferenceItem{}, false, nil
+	}
+	if len(slot.Items) != 1 {
+		return contracts.ReferenceItem{}, false, fmt.Errorf("reference slot %q must contain exactly one item", slotName)
+	}
+	item := slot.Items[0]
+	mediaType, _, err := mime.ParseMediaType(item.MediaType)
+	if err != nil {
+		return contracts.ReferenceItem{}, false, fmt.Errorf("reference slot %q item media type is invalid", slotName)
+	}
+	if !strings.EqualFold(mediaType, expectedMediaType) {
+		return contracts.ReferenceItem{}, false, fmt.Errorf("reference slot %q item media type must be %s", slotName, expectedMediaType)
+	}
+	if len(item.Content) > ReferenceMaxBytes {
+		return contracts.ReferenceItem{}, false, fmt.Errorf("reference slot %q item is %d bytes, limit %d", slotName, len(item.Content), ReferenceMaxBytes)
+	}
+	return item, true, nil
+}
+
+type combatTurnProjection struct {
+	Actor    string             `json:"actor"`
+	TurnKind dnd.CombatTurnKind `json:"turn_kind"`
+}
+
+func projectCombatTurns(turns []dnd.CombatTurn) []combatTurnProjection {
+	if turns == nil {
+		return nil
+	}
+	projection := make([]combatTurnProjection, len(turns))
+	for index, turn := range turns {
+		projection[index] = combatTurnProjection{Actor: turn.Actor, TurnKind: turn.TurnKind}
+	}
+	return projection
+}
+
+type npcInteractionProjection struct {
+	Name string                 `json:"name"`
+	Kind dnd.NPCInteractionKind `json:"kind"`
+}
+
+func projectNPCInteractions(interactions []dnd.NPCInteraction) []npcInteractionProjection {
+	projection := make([]npcInteractionProjection, 0, len(interactions))
+	for _, interaction := range interactions {
+		if interaction.Kind == dnd.NPCInteractionKindCombatOpponent {
+			projection = append(projection, npcInteractionProjection{Name: interaction.Name, Kind: interaction.Kind})
+		}
+	}
+	return projection
+}
+
+func newPromptInput(name string, content []byte) *contracts.LLMInputMaterial {
+	sum := sha256.Sum256(content)
+	material := contracts.NewLLMInputMaterial(name, promptProjectionMediaType, content, "sha256:"+hex.EncodeToString(sum[:]), "")
+	return &material
+}

internal/modules/dnd/extract/enemyevents/grounding_test.go

@@ -0,0 +1,279 @@
+package enemyevents
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	combatturncodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/combatturns"
+	interactioncodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/npcinteractions"
+	npccodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/npcs"
+	scenecodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/scenedescriptions"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/npcs/identity"
+	sceneregistry "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/scenedescriptions/registry"
+)
+
+func TestReferenceSlotsDescribeRequiredTypedArtifacts(t *testing.T) {
+	slots := referenceSlots()
+	if len(slots) != 8 {
+		t.Fatalf("ReferenceSlots() count = %d, want 8", len(slots))
+	}
+	byName := make(map[string]contracts.ReferenceSlot, len(slots))
+	for index, slot := range slots {
+		if index > 0 && slots[index-1].Name > slot.Name {
+			t.Fatalf("ReferenceSlots() is not sorted: %#v", slots)
+		}
+		byName[slot.Name] = slot
+	}
+	for _, want := range []struct {
+		name string
+		kind contracts.ArtifactKind
+	}{
+		{NPCRegistryReferenceSlot, dnd.NPCListKind},
+		{SceneDescriptionReferenceSlot, dnd.SceneDescriptionListKind},
+		{CombatTurnReferenceSlot, dnd.CombatTurnListKind},
+		{NPCInteractionReferenceSlot, dnd.NPCInteractionListKind},
+	} {
+		slot, ok := byName[want.name]
+		if !ok || !slot.Required || slot.MaxBytes != ReferenceMaxBytes || len(slot.AcceptedMediaTypes) != 1 || slot.AcceptedMediaTypes[0] != "application/json" || len(slot.AcceptedArtifactKinds) != 1 || slot.AcceptedArtifactKinds[0] != want.kind {
+			t.Fatalf("slot %q = %#v", want.name, slot)
+		}
+	}
+
+	slots[0].AcceptedMediaTypes[0] = "changed"
+	if referenceSlots()[0].AcceptedMediaTypes[0] == "changed" {
+		t.Fatal("ReferenceSlots() returned caller-owned storage")
+	}
+}
+
+func TestGroundingProducesExactSourceFreePromptInputs(t *testing.T) {
+	resolver, err := newGroundingResolver(groundingReferences(t, "Ashfang", dnd.SceneKindCombat))
+	if err != nil {
+		t.Fatal(err)
+	}
+	resolved, err := resolver.Resolve(contracts.ReferenceSet{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	inputs := resolved.PromptInputs()
+	want := map[string]string{
+		NPCRegistryReferenceSlot:    `{"npcs":[{"name":"Ashfang"}]}`,
+		CombatTurnReferenceSlot:     `{"combat_turns":[{"actor":"Ashfang","turn_kind":"turn"},{"actor":"Aria","turn_kind":"reaction"}]}`,
+		NPCInteractionReferenceSlot: `{"npc_interactions":[{"name":"Ashfang","kind":"combat_opponent"}]}`,
+	}
+	if len(inputs) != len(want) {
+		t.Fatalf("PromptInputs() = %#v", inputs)
+	}
+	for name, content := range want {
+		input, ok := inputs[name]
+		if !ok || string(input.Content) != content || input.Digest != digest([]byte(content)) || input.OriginURI != "" || input.MediaType != "application/json" {
+			t.Fatalf("PromptInputs()[%q] = %#v, want %q", name, input, content)
+		}
+		for _, forbidden := range []string{"source_ref", "npc-ashfang", "origin", "summary", "combat-session"} {
+			if strings.Contains(string(input.Content), forbidden) {
+				t.Fatalf("PromptInputs()[%q] leaked %q: %s", name, forbidden, input.Content)
+			}
+		}
+	}
+	if _, ok := inputs[SceneDescriptionReferenceSlot]; ok {
+		t.Fatal("scene descriptions were rendered as prompt grounding")
+	}
+
+	inputs[CombatTurnReferenceSlot] = contracts.NewLLMInputMaterial("changed", "text/plain", []byte("changed"), "", "")
+	if next := resolved.PromptInputs()[CombatTurnReferenceSlot]; string(next.Content) != want[CombatTurnReferenceSlot] {
+		t.Fatal("PromptInputs() did not return a defensive copy")
+	}
+}
+
+func TestGroundingResolvesGeneratedReferencesAndSceneEligibility(t *testing.T) {
+	prepared := groundingReferences(t, "Ashfang", dnd.SceneKindCombat)
+	resolver, err := newGroundingResolver(prepared)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	generated := groundingReferences(t, "Grimjaw", dnd.SceneKindNarrative)
+	resolved, err := resolver.Resolve(generated)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if got := string(resolved.PromptInputs()[NPCRegistryReferenceSlot].Content); got != `{"npcs":[{"name":"Grimjaw"}]}` {
+		t.Fatalf("generated NPC projection = %s", got)
+	}
+	match, err := resolver.SceneMatch(generated, combatChunk())
+	if err != nil {
+		t.Fatal(err)
+	}
+	if match != (sceneregistry.ChunkMatch{State: sceneregistry.MatchExact, Kind: dnd.SceneKindNarrative}) {
+		t.Fatalf("generated scene match = %#v", match)
+	}
+	match, err = resolver.SceneMatch(generated, &source.Chunk{ID: "other", Ref: combatChunk().Ref})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if match.State != sceneregistry.MatchMissing {
+		t.Fatal("missing scene was not reported")
+	}
+	mismatched := combatChunk()
+	mismatched.Ref.EndUnitID++
+	match, err = resolver.SceneMatch(generated, mismatched)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if match.State != sceneregistry.MatchMismatched {
+		t.Fatal("mismatched scene was not reported")
+	}
+
+	static, err := resolver.Resolve(contracts.ReferenceSet{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if got := string(static.PromptInputs()[NPCRegistryReferenceSlot].Content); got != `{"npcs":[{"name":"Ashfang"}]}` {
+		t.Fatalf("static NPC projection changed after generated resolution: %s", got)
+	}
+	match, err = resolver.SceneMatch(contracts.ReferenceSet{}, combatChunk())
+	if err != nil {
+		t.Fatal(err)
+	}
+	if match != (sceneregistry.ChunkMatch{State: sceneregistry.MatchExact, Kind: dnd.SceneKindCombat}) {
+		t.Fatalf("static scene match = %#v", match)
+	}
+}
+
+func TestGroundingRejectsMissingAndInvalidReferences(t *testing.T) {
+	valid := groundingReferences(t, "Ashfang", dnd.SceneKindCombat)
+	for _, test := range []struct {
+		slot string
+		want string
+	}{
+		{NPCRegistryReferenceSlot, "NPC registry"},
+		{SceneDescriptionReferenceSlot, "scene descriptions"},
+		{CombatTurnReferenceSlot, CombatTurnReferenceSlot},
+		{NPCInteractionReferenceSlot, NPCInteractionReferenceSlot},
+	} {
+		t.Run("missing "+test.slot, func(t *testing.T) {
+			resolver, err := newGroundingResolver(withoutSlot(valid, test.slot))
+			if err != nil {
+				t.Fatal(err)
+			}
+			if test.slot == SceneDescriptionReferenceSlot {
+				_, err = resolver.SceneMatch(contracts.ReferenceSet{}, combatChunk())
+			} else {
+				_, err = resolver.Resolve(contracts.ReferenceSet{})
+			}
+			if err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("grounding error = %v, want missing %q reference", err, test.slot)
+			}
+		})
+	}
+
+	tests := []struct {
+		name string
+		set  contracts.ReferenceSet
+	}{
+		{"multiple items", replaceSlot(valid, CombatTurnReferenceSlot, contracts.ResolvedReferenceSlot{Items: []contracts.ReferenceItem{{MediaType: "application/json"}, {MediaType: "application/json"}}})},
+		{"malformed durable content", replaceItem(valid, NPCInteractionReferenceSlot, contracts.ReferenceItem{MediaType: "application/json", Content: []byte(`{}`)})},
+		{"wrong media type", replaceItem(valid, CombatTurnReferenceSlot, contracts.ReferenceItem{MediaType: "text/plain", Content: valid.Slots[CombatTurnReferenceSlot].Items[0].Content})},
+		{"oversize", replaceItem(valid, CombatTurnReferenceSlot, contracts.ReferenceItem{MediaType: "application/json", Content: make([]byte, ReferenceMaxBytes+1)})},
+	}
+	resolver, err := newGroundingResolver(valid)
+	if err != nil {
+		t.Fatal(err)
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := resolver.Resolve(test.set); err == nil {
+				t.Fatal("Resolve() error = nil")
+			}
+		})
+	}
+}
+
+func groundingReferences(t *testing.T, enemy string, sceneKind dnd.SceneKind) contracts.ReferenceSet {
+	t.Helper()
+	npcContent, err := npccodec.New().Encode(dnd.NPCList{NPCs: []dnd.NPC{{
+		ID:         identity.DeriveID(enemy),
+		Name:       enemy,
+		SourceRefs: []source.SourceRef{{SourceID: "combat-session", StartUnitID: 1, EndUnitID: 1}},
+	}}})
+	if err != nil {
+		t.Fatal(err)
+	}
+	sceneContent, err := scenecodec.New().Encode(dnd.SceneDescriptionList{Scenes: []dnd.SceneDescription{{
+		ID:        "combat-scene",
+		SourceRef: source.SourceRef{SourceID: "combat-session", StartUnitID: 1, EndUnitID: 4},
+		Kind:      sceneKind,
+		Title:     "Combat session",
+		Summary:   "A detailed scene summary that must not reach the prompt.",
+	}}})
+	if err != nil {
+		t.Fatal(err)
+	}
+	turnContent, err := combatturncodec.New().Encode(dnd.CombatTurnList{CombatTurns: []dnd.CombatTurn{
+		{Actor: enemy, TurnKind: dnd.CombatTurnKindTurn, SourceRefs: []source.SourceRef{{SourceID: "combat-session", StartUnitID: 1, EndUnitID: 1}}},
+		{Actor: "Aria", TurnKind: dnd.CombatTurnKindReaction, SourceRefs: []source.SourceRef{{SourceID: "combat-session", StartUnitID: 2, EndUnitID: 2}}},
+	}})
+	if err != nil {
+		t.Fatal(err)
+	}
+	interactionContent, err := interactioncodec.New().Encode(dnd.NPCInteractionList{Interactions: []dnd.NPCInteraction{
+		{Name: enemy, Kind: dnd.NPCInteractionKindCombatOpponent, SourceRefs: []source.SourceRef{{SourceID: "combat-session", StartUnitID: 1, EndUnitID: 1}}},
+		{Name: "Aria", Kind: dnd.NPCInteractionKindCombatAlly, SourceRefs: []source.SourceRef{{SourceID: "combat-session", StartUnitID: 2, EndUnitID: 2}}},
+	}})
+	if err != nil {
+		t.Fatal(err)
+	}
+	return contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{
+		NPCRegistryReferenceSlot:      {Items: []contracts.ReferenceItem{newReferenceItem(NPCRegistryReferenceSlot, npcContent)}},
+		SceneDescriptionReferenceSlot: {Items: []contracts.ReferenceItem{newReferenceItem(SceneDescriptionReferenceSlot, sceneContent)}},
+		CombatTurnReferenceSlot:       {Items: []contracts.ReferenceItem{newReferenceItem(CombatTurnReferenceSlot, turnContent)}},
+		NPCInteractionReferenceSlot:   {Items: []contracts.ReferenceItem{newReferenceItem(NPCInteractionReferenceSlot, interactionContent)}},
+	}}
+}
+
+func newReferenceItem(slot string, content []byte) contracts.ReferenceItem {
+	return contracts.ReferenceItem{
+		SlotName:  slot,
+		MediaType: "application/json",
+		Content:   append([]byte(nil), content...),
+		Digest:    digest(content),
+		Origin:    contracts.ReferenceOrigin{Type: "file", URI: "file:///private/reference.json"},
+		Producer:  contracts.ReferenceProducer{PipelineID: "prior", StepID: "extract", LaneID: "lane", ModuleKey: "dnd/example"},
+	}
+}
+
+func withoutSlot(set contracts.ReferenceSet, name string) contracts.ReferenceSet {
+	cloned := contracts.ReferenceSet{Slots: make(map[string]contracts.ResolvedReferenceSlot, len(set.Slots)-1)}
+	for slotName, slot := range set.Slots {
+		if slotName != name {
+			cloned.Slots[slotName] = slot
+		}
+	}
+	return cloned
+}
+
+func replaceSlot(set contracts.ReferenceSet, name string, value contracts.ResolvedReferenceSlot) contracts.ReferenceSet {
+	cloned := contracts.ReferenceSet{Slots: make(map[string]contracts.ResolvedReferenceSlot, len(set.Slots))}
+	for slotName, slot := range set.Slots {
+		cloned.Slots[slotName] = slot
+	}
+	cloned.Slots[name] = value
+	return cloned
+}
+
+func replaceItem(set contracts.ReferenceSet, name string, item contracts.ReferenceItem) contracts.ReferenceSet {
+	return replaceSlot(set, name, contracts.ResolvedReferenceSlot{Items: []contracts.ReferenceItem{item}})
+}
+
+func combatChunk() *source.Chunk {
+	return &source.Chunk{ID: "combat-scene", Ref: source.SourceRef{SourceID: "combat-session", StartUnitID: 1, EndUnitID: 4}}
+}
+
+func digest(content []byte) string {
+	sum := sha256.Sum256(content)
+	return "sha256:" + hex.EncodeToString(sum[:])
+}

internal/modules/dnd/extract/enemyevents/mapping.go

@@ -0,0 +1,60 @@
+package enemyevents
+
+import (
+	"sort"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+type orderedEnemyEvent struct {
+	value       dnd.EnemyEvent
+	earliest    int
+	hasEvidence bool
+}
+
+func canonicalEnemyEventList(response extractionResponse, order shared.SourceRefOrder, sourceID string) dnd.EnemyEventList {
+	if response.Events == nil {
+		return dnd.EnemyEventList{}
+	}
+	ordered := make([]orderedEnemyEvent, len(response.Events))
+	for index, event := range response.Events {
+		refs := order.Canonicalize(sourceRefs(event.SourceRefs, sourceID))
+		earliest, hasEvidence := order.EarliestValid(refs)
+		ordered[index] = orderedEnemyEvent{
+			value: dnd.EnemyEvent{
+				Name:       event.Name,
+				Kind:       dnd.EnemyEventKind(event.Kind),
+				SourceRefs: refs,
+			},
+			earliest:    earliest,
+			hasEvidence: hasEvidence,
+		}
+	}
+	sort.SliceStable(ordered, func(left, right int) bool {
+		if ordered[left].hasEvidence != ordered[right].hasEvidence {
+			return ordered[left].hasEvidence
+		}
+		if !ordered[left].hasEvidence {
+			return false
+		}
+		return ordered[left].earliest < ordered[right].earliest
+	})
+	events := make([]dnd.EnemyEvent, len(ordered))
+	for index := range ordered {
+		events[index] = ordered[index].value
+	}
+	return dnd.EnemyEventList{Events: events}
+}
+
+func sourceRefs(refs []enemySourceRefResponse, sourceID string) []source.SourceRef {
+	if refs == nil {
+		return nil
+	}
+	values := make([]source.SourceRef, len(refs))
+	for index, ref := range refs {
+		values[index] = source.SourceRef{SourceID: sourceID, StartUnitID: ref.StartUnitID, EndUnitID: ref.EndUnitID}
+	}
+	return values
+}

internal/modules/dnd/extract/enemyevents/model.go

@@ -0,0 +1,16 @@
+package enemyevents
+
+type extractionResponse struct {
+	Events []enemyEventResponse `json:"events"`
+}
+
+type enemyEventResponse struct {
+	Name       string                   `json:"name"`
+	Kind       string                   `json:"kind"`
+	SourceRefs []enemySourceRefResponse `json:"source_refs"`
+}
+
+type enemySourceRefResponse struct {
+	StartUnitID int `json:"start_unit_id"`
+	EndUnitID   int `json:"end_unit_id"`
+}

internal/modules/dnd/extract/enemyevents/prompt_assets.go

@@ -0,0 +1,54 @@
+package enemyevents
+
+import (
+	"fmt"
+	"sync"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/promptfs"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const promptAssetRoot = "assets/prompts"
+
+var promptAssetManifest = shared.PromptAssetManifest{
+	ModuleDir: "dnd.enemy_events",
+	ModuleFiles: []promptfs.ModulePromptFile{
+		{Name: "dnd.enemy_events.yaml", Path: "assets/prompts/dnd.enemy_events.yaml"},
+		{Name: "grounding.md", Path: "assets/prompts/grounding.md"},
+		{Name: "task.md", Path: "assets/prompts/task.md"},
+		{Name: "instructions.md", Path: "assets/prompts/instructions.md"},
+	},
+	SharedFiles: []string{
+		"common-dnd-system.md",
+		"common-dnd-extraction-evidence.md",
+		"common-dnd-identity.md",
+		"common-dnd-transcript.md",
+		"common-dnd-references.md",
+		"common-dnd-npcs.md",
+	},
+}
+
+func RegisterPromptAssets(registry *llm.AssetRegistry) error {
+	promptFS, err := promptAssetManifest.PromptFS(embeddedAssets)
+	if err != nil {
+		return fmt.Errorf("prepare enemy-event prompt assets: %w", err)
+	}
+	if err := registry.RegisterPromptFS(promptFS, promptAssetRoot); err != nil {
+		return err
+	}
+	return registry.RegisterSchemaFS(embeddedAssets, "assets/schemas")
+}
+
+func promptAssetMetadata() (string, error) {
+	promptAssetHashOnce.Do(func() {
+		promptAssetHash, promptAssetHashErr = promptAssetManifest.Hash(embeddedAssets)
+	})
+	return promptAssetHash, promptAssetHashErr
+}
+
+var (
+	promptAssetHashOnce sync.Once
+	promptAssetHash     string
+	promptAssetHashErr  error
+)

internal/modules/dnd/extract/enemyevents/prompt_assets_test.go

@@ -0,0 +1,123 @@
+package enemyevents
+
+import (
+	"context"
+	"io/fs"
+	"strings"
+	"testing"
+	"time"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/promptkit"
+)
+
+func TestRegisterPromptAssetsAndPrepareEnemyEventPrompt(t *testing.T) {
+	registry := llm.NewAssetRegistry()
+	if err := RegisterPromptAssets(registry); err != nil {
+		t.Fatal(err)
+	}
+	schemaFS, err := registry.SchemaFS()
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := fs.ReadFile(schemaFS, "dnd_enemy_events_llm.v1.json"); err != nil {
+		t.Fatalf("response schema asset: %v", err)
+	}
+
+	prepared := prepareEnemyEventPrompt(t)
+	if prepared.PromptID != PromptID || prepared.OutputContract.SchemaPath != "dnd_enemy_events_llm.v1.json" || prepared.SelectedProfileID != "dnd-extraction" {
+		t.Fatalf("prepared prompt = %#v", prepared)
+	}
+	transcriptIndex := renderedMessageIndex(t, prepared.Messages, "enemy-transcript")
+	for _, sentinel := range []string{"enemy-npc", "enemy-turn", "enemy-opponent", "Extract Dungeons & Dragons enemy events"} {
+		if index := renderedMessageIndex(t, prepared.Messages, sentinel); index <= transcriptIndex {
+			t.Fatalf("message containing %q has index %d, want after transcript index %d", sentinel, index, transcriptIndex)
+		}
+	}
+	instructionIndex := renderedMessageIndex(t, prepared.Messages, "Return the `events` array")
+	if instructionIndex <= transcriptIndex {
+		t.Fatalf("instruction index = %d, want after transcript index %d", instructionIndex, transcriptIndex)
+	}
+	if instructionIndex != len(prepared.Messages)-1 {
+		t.Fatalf("instruction message index = %d, want final message", instructionIndex)
+	}
+	if cache := prepared.Messages[instructionIndex].CacheControl; cache == nil || cache.Type != promptkit.CacheControlEphemeral {
+		t.Fatalf("final instruction cache control = %#v", cache)
+	}
+}
+
+func TestEnemyEventPromptRequiresGroundingInputs(t *testing.T) {
+	engine := newEnemyEventPromptEngine(t)
+	for _, inputName := range []string{"npcs", "combat_turns", "npc_interactions"} {
+		t.Run(inputName, func(t *testing.T) {
+			inputs := enemyEventPromptInputs()
+			delete(inputs, inputName)
+			_, err := engine.Prepare(context.Background(), promptkit.RunRequest{
+				PromptID: PromptID, PromptVersion: SchemaVersion, Inputs: inputs,
+			})
+			if err == nil || !strings.Contains(err.Error(), inputName) {
+				t.Fatalf("Prepare() error = %v, want required %q input", err, inputName)
+			}
+		})
+	}
+}
+
+func prepareEnemyEventPrompt(t *testing.T) *promptkit.PreparedRun {
+	t.Helper()
+	prepared, err := newEnemyEventPromptEngine(t).Prepare(context.Background(), promptkit.RunRequest{
+		PromptID: PromptID, PromptVersion: SchemaVersion, Inputs: enemyEventPromptInputs(),
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	return prepared
+}
+
+func newEnemyEventPromptEngine(t *testing.T) *promptkit.Engine {
+	t.Helper()
+	registry := llm.NewAssetRegistry()
+	if err := RegisterPromptAssets(registry); err != nil {
+		t.Fatal(err)
+	}
+	options, err := registry.PromptKitOptions()
+	if err != nil {
+		t.Fatal(err)
+	}
+	options = append(options, promptkit.WithProfiles(promptkit.OpenAICompatibleProfile(promptkit.OpenAICompatibleProfileConfig{
+		ID: "dnd-extraction", Endpoint: "http://127.0.0.1:1/v1", Model: "enemy-test-model",
+	})))
+	engine, err := promptkit.NewEngine(promptkit.Config{Timeout: time.Second}, options...)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return engine
+}
+
+func enemyEventPromptInputs() map[string]promptkit.ArtifactRef {
+	return map[string]promptkit.ArtifactRef{
+		"transcript":       promptkit.Inline(`{"value":"enemy-transcript"}`),
+		"players":          promptkit.Inline("enemy-player"),
+		"party":            promptkit.Inline("enemy-party"),
+		"glossary":         promptkit.Inline("enemy-glossary"),
+		"npcs":             promptkit.Inline(`{"npcs":[{"name":"enemy-npc"}]}`),
+		"combat_turns":     promptkit.Inline(`{"combat_turns":[{"actor":"enemy-turn","turn_kind":"turn"}]}`),
+		"npc_interactions": promptkit.Inline(`{"npc_interactions":[{"name":"enemy-opponent","kind":"combat_opponent"}]}`),
+	}
+}
+
+func renderedMessageIndex(t *testing.T, messages []promptkit.RenderedMessage, sentinel string) int {
+	t.Helper()
+	index := -1
+	occurrences := 0
+	for messageIndex, message := range messages {
+		count := strings.Count(message.Content, sentinel)
+		if count > 0 {
+			index = messageIndex
+			occurrences += count
+		}
+	}
+	if occurrences != 1 {
+		t.Fatalf("message sentinel %q rendered %d times, want exactly once", sentinel, occurrences)
+	}
+	return index
+}

internal/modules/dnd/extract/enemyevents/schema.go

@@ -1,12 +1,12 @@
-package npcs
+package enemyevents
 
 import "gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
 
 const (
-	PromptID           = "dnd.npcs.normalize"
-	ResponseSchemaKey  = llm.ResponseSchemaKey("dnd_npcs_normalize_llm")
-	ResponseSchemaID   = "notarius.dnd.npcs.normalize.llm"
-	ResponseSchemaName = "notarius_dnd_npcs_normalize_llm_v1"
+	PromptID           = "dnd.enemy_events"
+	ResponseSchemaKey  = llm.ResponseSchemaKey("dnd_enemy_events_llm")
+	ResponseSchemaID   = "notarius.dnd.enemy_events.llm"
+	ResponseSchemaName = "notarius_dnd_enemy_events_llm_v1"
 	SchemaVersion      = "v1"
 )
 
@@ -16,6 +16,6 @@ func loadResponseSchema() (llm.ResponseSchema, error) {
 		ID:        ResponseSchemaID,
 		Version:   SchemaVersion,
 		Name:      ResponseSchemaName,
-		AssetPath: "assets/schemas/dnd_npcs_normalize_llm.v1.json",
+		AssetPath: "assets/schemas/dnd_enemy_events_llm.v1.json",
 	})
 }

internal/modules/dnd/extract/enemyevents/schema_test.go

@@ -0,0 +1,97 @@
+package enemyevents
+
+import (
+	"bytes"
+	"encoding/json"
+	"strings"
+	"testing"
+
+	"github.com/santhosh-tekuri/jsonschema/v6"
+)
+
+func TestResponseSchemaDefinesPrivateStructuralBoundary(t *testing.T) {
+	schema, err := loadResponseSchema()
+	if err != nil {
+		t.Fatal(err)
+	}
+	if schema.Key != ResponseSchemaKey || schema.ID != ResponseSchemaID || schema.Name != ResponseSchemaName || schema.Version != SchemaVersion || !strings.HasPrefix(schema.SHA256, "sha256:") || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("schema = %#v", schema)
+	}
+	valid := validEnemyResponse()
+	content, err := json.Marshal(valid)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if err := validateEnemySchema(content, schema.JSONSchema); err != nil {
+		t.Fatalf("valid response rejected: %v", err)
+	}
+	semantic := validEnemyResponse()
+	event := semantic["events"].([]any)[0].(map[string]any)
+	event["name"] = ""
+	event["kind"] = "unsupported"
+	ref := event["source_refs"].([]any)[0].(map[string]any)
+	ref["start_unit_id"] = 0
+	ref["end_unit_id"] = -1
+	content, err = json.Marshal(semantic)
+	if err != nil || validateEnemySchema(content, schema.JSONSchema) != nil {
+		t.Fatalf("validator-owned semantics were rejected: %v", err)
+	}
+}
+
+func TestResponseSchemaRejectsInvalidStructure(t *testing.T) {
+	schema, err := loadResponseSchema()
+	if err != nil {
+		t.Fatal(err)
+	}
+	for _, mutate := range []func(map[string]any){
+		func(event map[string]any) { delete(event, "name") },
+		func(event map[string]any) { event["kind"] = 1 },
+		func(event map[string]any) { event["unexpected"] = true },
+		func(event map[string]any) { event["source_refs"].([]any)[0].(map[string]any)["source_id"] = "session" },
+	} {
+		candidate := validEnemyResponse()
+		mutate(candidate["events"].([]any)[0].(map[string]any))
+		content, err := json.Marshal(candidate)
+		if err != nil {
+			t.Fatal(err)
+		}
+		if err := validateEnemySchema(content, schema.JSONSchema); err == nil {
+			t.Fatal("private schema accepted structurally invalid response")
+		}
+	}
+	first, err := loadResponseSchema()
+	if err != nil {
+		t.Fatal(err)
+	}
+	first.JSONSchema[0] = '['
+	second, err := loadResponseSchema()
+	if err != nil || !json.Valid(second.JSONSchema) || bytes.Equal(first.JSONSchema, second.JSONSchema) {
+		t.Fatalf("schema defensive copy = %s, %v", second.JSONSchema, err)
+	}
+}
+
+func validEnemyResponse() map[string]any {
+	return map[string]any{"events": []any{map[string]any{
+		"name": "Ashfang", "kind": "engaged", "source_refs": []any{map[string]any{"start_unit_id": 1, "end_unit_id": 1}},
+	}}}
+}
+
+func validateEnemySchema(instanceContent, schemaContent []byte) error {
+	instance, err := jsonschema.UnmarshalJSON(bytes.NewReader(instanceContent))
+	if err != nil {
+		return err
+	}
+	document, err := jsonschema.UnmarshalJSON(bytes.NewReader(schemaContent))
+	if err != nil {
+		return err
+	}
+	compiler := jsonschema.NewCompiler()
+	if err := compiler.AddResource("schema.json", document); err != nil {
+		return err
+	}
+	schema, err := compiler.Compile("schema.json")
+	if err != nil {
+		return err
+	}
+	return schema.Validate(instance)
+}

internal/modules/dnd/extract/locationoccurrences/assets.go

@@ -0,0 +1,6 @@
+package locationoccurrences
+
+import "embed"
+
+//go:embed assets/prompts/*.yaml assets/prompts/*.md assets/schemas/*.json
+var embeddedAssets embed.FS

internal/modules/dnd/extract/locationoccurrences/assets/prompts/dnd.location_occurrences.yaml

@@ -0,0 +1,47 @@
+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: locations
+    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.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./locations.md
+  - role: user
+    content_file: ./task.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

internal/modules/dnd/extract/locationoccurrences/assets/prompts/instructions.md

@@ -0,0 +1,8 @@
+Return the occurrences array even when no occurrence is established. Every
+record must contain location_id, name, kind, and source_refs. Copy location_id
+and name from one supplied registry record, and cite only narrow transcript
+ranges that support both that location and its classified occurrence.
+
+Do not summarize location descriptions, infer a missing registry record, or
+use registry context as evidence. Omit source_id; Notarius assigns the current
+transcript source identity.

internal/modules/dnd/extract/locationoccurrences/assets/prompts/locations.md

@@ -0,0 +1,9 @@
+A normalized location registry is provided below for identity grounding. It may
+be empty. Each record contains the exact location ID and canonical display name
+to copy when the transcript establishes an occurrence of that place.
+
+Registry content is context, not occurrence evidence. Do not derive an
+occurrence or a source range from the registry, and do not infer a location
+that is absent from it.
+
+{{ input "locations" }}

internal/modules/dnd/extract/locationoccurrences/assets/prompts/task.md

@@ -0,0 +1,29 @@
+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
+only the exact ID and name pair from the supplied location registry. Return an
+empty occurrences array when no supplied location has an evidenced occurrence
+in this transcript passage.
+
+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 itself explicitly names a supplied registry
+location, it may be mentioned using the narrow passage that supports that
+reference.
+
+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.

internal/modules/dnd/extract/locationoccurrences/assets/schemas/dnd_location_occurrences_llm.v1.json

@@ -0,0 +1,34 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.location_occurrences.llm",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["occurrences"],
+  "properties": {
+    "occurrences": {
+      "type": "array",
+      "items": {
+        "type": "object",
+        "additionalProperties": false,
+        "required": ["location_id", "name", "kind", "source_refs"],
+        "properties": {
+          "location_id": {"type": "string"},
+          "name": {"type": "string"},
+          "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"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

internal/modules/dnd/extract/locationoccurrences/canonicalize.go

@@ -0,0 +1,106 @@
+package locationoccurrences
+
+import (
+	"reflect"
+	"sort"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+type orderedOccurrence struct {
+	value       dnd.LocationOccurrence
+	earliest    int
+	hasEvidence bool
+}
+
+func canonicalOccurrenceList(response extractionResponse, order shared.SourceRefOrder, sourceID string) dnd.LocationOccurrenceList {
+	if response.Occurrences == nil {
+		return dnd.LocationOccurrenceList{}
+	}
+	ordered := make([]orderedOccurrence, len(response.Occurrences))
+	for index, occurrence := range response.Occurrences {
+		refs := order.Canonicalize(canonicalSourceRefs(occurrence.SourceRefs, sourceID))
+		earliest, hasEvidence := order.EarliestValid(refs)
+		ordered[index] = orderedOccurrence{value: dnd.LocationOccurrence{
+			LocationID: occurrence.LocationID,
+			Name:       occurrence.Name,
+			Kind:       dnd.LocationOccurrenceKind(occurrence.Kind),
+			SourceRefs: refs,
+		}, earliest: earliest, hasEvidence: hasEvidence}
+	}
+	sort.SliceStable(ordered, func(left, right int) bool {
+		return lessOccurrence(ordered[left], ordered[right], order)
+	})
+	occurrences := make([]dnd.LocationOccurrence, 0, len(ordered))
+	for _, occurrence := range ordered {
+		if len(occurrences) == 0 || !sameOccurrence(occurrences[len(occurrences)-1], occurrence.value) {
+			occurrences = append(occurrences, occurrence.value)
+		}
+	}
+	return dnd.LocationOccurrenceList{Occurrences: occurrences}
+}
+
+func lessOccurrence(left, right orderedOccurrence, order shared.SourceRefOrder) bool {
+	if left.hasEvidence != right.hasEvidence {
+		return left.hasEvidence
+	}
+	if left.hasEvidence && left.earliest != right.earliest {
+		return left.earliest < right.earliest
+	}
+	if left.value.LocationID != right.value.LocationID {
+		return left.value.LocationID < right.value.LocationID
+	}
+	if left.value.Name != right.value.Name {
+		return left.value.Name < right.value.Name
+	}
+	if kindOrder(left.value.Kind) != kindOrder(right.value.Kind) {
+		return kindOrder(left.value.Kind) < kindOrder(right.value.Kind)
+	}
+	return lessReferences(left.value.SourceRefs, right.value.SourceRefs, order)
+}
+
+func kindOrder(kind dnd.LocationOccurrenceKind) int {
+	switch kind {
+	case dnd.LocationOccurrenceKindVisited:
+		return 0
+	case dnd.LocationOccurrenceKindPlanned:
+		return 1
+	case dnd.LocationOccurrenceKindRecalled:
+		return 2
+	case dnd.LocationOccurrenceKindMentioned:
+		return 3
+	default:
+		return 4
+	}
+}
+
+func lessReferences(left, right []source.SourceRef, order shared.SourceRefOrder) bool {
+	limit := len(left)
+	if len(right) < limit {
+		limit = len(right)
+	}
+	for index := 0; index < limit; index++ {
+		if left[index] == right[index] {
+			continue
+		}
+		return order.Less(left[index], right[index])
+	}
+	return len(left) < len(right)
+}
+
+func sameOccurrence(left, right dnd.LocationOccurrence) bool {
+	return left.LocationID == right.LocationID && left.Name == right.Name && left.Kind == right.Kind && reflect.DeepEqual(left.SourceRefs, right.SourceRefs)
+}
+
+func canonicalSourceRefs(values []occurrenceSourceRefResponse, sourceID string) []source.SourceRef {
+	if values == nil {
+		return nil
+	}
+	refs := make([]source.SourceRef, len(values))
+	for index, value := range values {
+		refs[index] = source.SourceRef{SourceID: sourceID, StartUnitID: value.StartUnitID, EndUnitID: value.EndUnitID}
+	}
+	return refs
+}

internal/modules/dnd/extract/locationoccurrences/extractor.go

@@ -0,0 +1,188 @@
+// Package locationoccurrences extracts source-grounded D&D location occurrences.
+package locationoccurrences
+
+import (
+	"context"
+	"fmt"
+	"sort"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	locationregistry "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/registry"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const (
+	Key           = "dnd/location-occurrences"
+	mappingPolicy = "dnd.location_occurrences.extract_mapping.v1"
+)
+
+const (
+	LocationRegistryReferenceSlot = locationregistry.ReferenceSlot
+	LocationRegistryMaxBytes      = locationregistry.MaxBytes
+)
+
+var requiredCapabilities = []string{"chunks", "source.transcript"}
+var providedCapabilities = []string{"dnd.location_occurrences"}
+
+var referenceSlotDescriptions = shared.ReferenceSlotDescriptions{
+	Glossary: "Optional campaign glossary reference material used only for location-occurrence disambiguation.",
+	Party:    "Optional party roster reference material used only for location-occurrence disambiguation.",
+	Players:  "Optional player list reference material used only for location-occurrence disambiguation.",
+	Roster:   "Deprecated alias for party roster reference material used only for location-occurrence disambiguation.",
+}
+
+func referenceSlots() []contracts.ReferenceSlot {
+	slots := shared.ReferenceSlots(referenceSlotDescriptions)
+	slots = append(slots, contracts.ReferenceSlot{
+		Name:                  LocationRegistryReferenceSlot,
+		Description:           "Required normalized location registry used only for location identity grounding, never as occurrence evidence.",
+		Required:              true,
+		AcceptedMediaTypes:    []string{"application/json"},
+		AcceptedArtifactKinds: []contracts.ArtifactKind{dnd.LocationListKind},
+		MaxBytes:              LocationRegistryMaxBytes,
+	})
+	sort.Slice(slots, func(left, right int) bool { return slots[left].Name < slots[right].Name })
+	return slots
+}
+
+var _ contracts.Extractor[dnd.LocationOccurrenceList] = (*Extractor)(nil)
+var _ contracts.ManifestMetadataProvider = (*Extractor)(nil)
+var _ pipeline.CheckpointFingerprintProvider = (*Extractor)(nil)
+
+type Options struct{}
+
+type Extractor struct {
+	llm               contracts.StructuredLLMClient
+	locationResolver  *locationregistry.Resolver
+	promptSHA         string
+	responseSchemaSHA string
+}
+
+func New(llmClient contracts.StructuredLLMClient, _ Options, references ...contracts.ReferenceSet) (*Extractor, error) {
+	if llmClient == nil {
+		return nil, extractorErrorf("LLM client must not be nil")
+	}
+	if len(references) > 1 {
+		return nil, extractorErrorf("at most one reference set may be supplied")
+	}
+	var referenceSet contracts.ReferenceSet
+	if len(references) == 1 {
+		referenceSet = references[0]
+	}
+	locationResolver, err := locationregistry.NewResolver(referenceSet)
+	if err != nil {
+		return nil, extractorErrorf("prepare location registry prompt input: %w", err)
+	}
+	promptSHA, err := promptAssetMetadata()
+	if err != nil {
+		return nil, extractorErrorf("load prompt metadata: %w", err)
+	}
+	responseSchema, err := loadResponseSchema()
+	if err != nil {
+		return nil, extractorErrorf("load response schema: %w", err)
+	}
+	return &Extractor{llm: llmClient, locationResolver: locationResolver, promptSHA: promptSHA, responseSchemaSHA: responseSchema.SHA256}, nil
+}
+
+func (e *Extractor) Key() string { return Key }
+
+func (e *Extractor) ReferenceSlots() []contracts.ReferenceSlot { return referenceSlots() }
+
+func (e *Extractor) ManifestMetadata() map[string]any {
+	if e == nil {
+		return nil
+	}
+	metadata := map[string]any{
+		"prompt_id":               PromptID,
+		"prompt_version":          SchemaVersion,
+		"prompt_sha256":           e.promptSHA,
+		"mapping_policy":          mappingPolicy,
+		"response_schema_key":     string(ResponseSchemaKey),
+		"response_schema_id":      ResponseSchemaID,
+		"response_schema_name":    ResponseSchemaName,
+		"response_schema_version": SchemaVersion,
+		"response_schema_sha256":  e.responseSchemaSHA,
+	}
+	seeded := e.locationResolver.Seeded()
+	if seeded.Bound() {
+		metadata["location_registry_digest"] = seeded.Digest()
+		metadata["location_count"] = seeded.Count()
+	}
+	return metadata
+}
+
+func (e *Extractor) CheckpointFingerprints() []pipeline.CheckpointFingerprint {
+	if e == nil {
+		return nil
+	}
+	return []pipeline.CheckpointFingerprint{
+		{Name: "prompt", Value: e.promptSHA},
+		{Name: "response_schema", Value: e.responseSchemaSHA},
+		{Name: "mapping_policy", Value: mappingPolicy},
+		{Name: "location_registry", Value: e.locationResolver.Seeded().ProjectionDigest()},
+	}
+}
+
+func (e *Extractor) Extract(ctx context.Context, req contracts.TypedExtractionRequest) (contracts.TypedExtractionResult[dnd.LocationOccurrenceList], error) {
+	if e == nil {
+		return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{}, extractorErrorf("extractor must not be nil")
+	}
+	if e.llm == nil {
+		return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{}, extractorErrorf("LLM client must not be nil")
+	}
+	sourceInput, err := shared.PrepareChunkExtraction(ctx, req)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{}, extractorErrorf("%w", err)
+	}
+	registry, err := e.locationResolver.Resolve(req.References)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{}, extractorErrorf("resolve location registry: %w", err)
+	}
+	if !registry.Bound() {
+		return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{}, extractorErrorf("location registry reference is required")
+	}
+
+	var response extractionResponse
+	inputs := shared.PromptInputs(sourceInput, req.References)
+	inputs[LocationRegistryReferenceSlot] = registry.PromptInput()
+	if _, err := e.llm.CompleteStructured(ctx, contracts.StructuredCompletionRequest{
+		StageName: Key, PromptID: PromptID, PromptVersion: SchemaVersion,
+		ProfileID: req.LLMProfile, SessionID: req.SessionID, Inputs: inputs,
+	}, &response); err != nil {
+		return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{}, extractorErrorf("complete structured output: %w", err)
+	}
+	return contracts.TypedExtractionResult[dnd.LocationOccurrenceList]{Value: canonicalOccurrenceList(response, shared.NewSourceRefOrder(req.Source), req.Source.ID)}, nil
+}
+
+func ModuleSpec() pipeline.ModuleSpec {
+	return pipeline.ModuleSpec{
+		Key: Key, Stage: pipeline.StageExtract, ExecutionClass: contracts.ExecutionClassLLMBacked,
+		Requires: append([]string(nil), requiredCapabilities...), Provides: append([]string(nil), providedCapabilities...),
+		ArtifactKind: dnd.LocationOccurrenceListKind, ReferenceSlots: referenceSlots(),
+	}
+}
+
+func Register(registry *pipeline.ExtractorRegistry) error {
+	return pipeline.RegisterExtractorBuilder(registry, ModuleSpec(), validateOptions, func(request pipeline.BuildRequest) (contracts.Extractor[dnd.LocationOccurrenceList], error) {
+		options, err := DecodeOptions(request.Options)
+		if err != nil {
+			return nil, err
+		}
+		return New(request.Dependencies.LLM, options, request.References)
+	})
+}
+
+func validateOptions(options map[string]any) error { _, err := DecodeOptions(options); return err }
+
+func DecodeOptions(options map[string]any) (Options, error) {
+	if err := pipeline.RejectUnknownOptions(options); err != nil {
+		return Options{}, extractorErrorf("%w", err)
+	}
+	return Options{}, nil
+}
+
+func extractorErrorf(format string, args ...any) error {
+	return fmt.Errorf("dnd location occurrences extractor: "+format, args...)
+}

internal/modules/dnd/extract/locationoccurrences/extractor_test.go

@@ -0,0 +1,243 @@
+package locationoccurrences
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	locationcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/locations"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+func TestExtractMapsKindsOrdersOccurrencesAndPreservesIndependentFacts(t *testing.T) {
+	locations := locationRegistry(t, "The Tavern", "The Tavern")
+	first, second := locations.Locations[0], locations.Locations[1]
+	client := &fakeOccurrencesLLMClient{response: extractionResponse{Occurrences: []occurrenceResponse{
+		{LocationID: second.ID, Name: second.Name, Kind: "mentioned", SourceRefs: occurrenceRefs(30, 30)},
+		{LocationID: first.ID, Name: first.Name, Kind: "mentioned", SourceRefs: occurrenceRefs(10, 10)},
+		{LocationID: first.ID, Name: first.Name, Kind: "recalled", SourceRefs: occurrenceRefs(10, 10)},
+		{LocationID: first.ID, Name: first.Name, Kind: "planned", SourceRefs: occurrenceRefs(10, 10)},
+		{LocationID: first.ID, Name: first.Name, Kind: "visited", SourceRefs: append(occurrenceRefs(10, 10), occurrenceRefs(10, 10)...)},
+		{LocationID: first.ID, Name: first.Name, Kind: "visited", SourceRefs: occurrenceRefs(20, 20)},
+		{LocationID: first.ID, Name: first.Name, Kind: "visited", SourceRefs: occurrenceRefs(10, 10)},
+	}}}
+	references := registryReferences(t, locations)
+	req := extractionRequest()
+	req.References = references
+	result, err := newExtractor(t, client, references).Extract(context.Background(), req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(result.Value.Occurrences) != 6 {
+		t.Fatalf("occurrences = %#v, want exact duplicate removed", result.Value.Occurrences)
+	}
+	got := result.Value.Occurrences
+	if kinds := []dnd.LocationOccurrenceKind{got[0].Kind, got[1].Kind, got[2].Kind, got[3].Kind}; !reflect.DeepEqual(kinds, []dnd.LocationOccurrenceKind{dnd.LocationOccurrenceKindVisited, dnd.LocationOccurrenceKindPlanned, dnd.LocationOccurrenceKindRecalled, dnd.LocationOccurrenceKindMentioned}) {
+		t.Fatalf("same-evidence kind order = %#v", kinds)
+	}
+	if got[4].Kind != dnd.LocationOccurrenceKindVisited || got[4].SourceRefs[0].StartUnitID != 20 || got[5].LocationID != second.ID || got[5].SourceRefs[0].StartUnitID != 30 {
+		t.Fatalf("occurrence order = %#v", got)
+	}
+	if !reflect.DeepEqual(got[0].SourceRefs, []source.SourceRef{{SourceID: req.Source.ID, StartUnitID: 10, EndUnitID: 10}}) {
+		t.Fatalf("canonical evidence = %#v", got[0].SourceRefs)
+	}
+}
+
+func TestExtractUsesIDsNamesAndCurrentTranscriptEvidenceOnly(t *testing.T) {
+	locations := locationRegistry(t, "The Tavern", "The Tavern")
+	first, second := locations.Locations[0], locations.Locations[1]
+	client := &fakeOccurrencesLLMClient{response: extractionResponse{Occurrences: []occurrenceResponse{{
+		LocationID: second.ID, Name: second.Name, Kind: "visited", SourceRefs: occurrenceRefs(10, 10),
+	}}}}
+	references := registryReferences(t, locations)
+	req := extractionRequest()
+	req.References = references
+	result, err := newExtractor(t, client, references).Extract(context.Background(), req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if occurrence := result.Value.Occurrences[0]; occurrence.LocationID != second.ID || occurrence.Name != second.Name || occurrence.SourceRefs[0].SourceID != req.Source.ID {
+		t.Fatalf("occurrence = %#v", occurrence)
+	}
+	input := client.requests[0].Inputs[LocationRegistryReferenceSlot]
+	if input.Name != LocationRegistryReferenceSlot || !strings.Contains(string(input.Content), first.ID) || !strings.Contains(string(input.Content), second.ID) {
+		t.Fatalf("location prompt input = %#v", input)
+	}
+	for _, forbidden := range []string{"source_refs", "source_id", "other-session"} {
+		if strings.Contains(string(input.Content), forbidden) {
+			t.Fatalf("location prompt leaked %q: %s", forbidden, input.Content)
+		}
+	}
+	if strings.Contains(string(client.requests[0].Inputs["transcript"].Content), "other-session") {
+		t.Fatal("transcript input contains registry evidence")
+	}
+	metadata, err := json.Marshal(newExtractor(t, &fakeOccurrencesLLMClient{}, references).ManifestMetadata())
+	if err != nil || strings.Contains(string(metadata), "other-session") || strings.Contains(string(metadata), first.ID) {
+		t.Fatalf("manifest metadata = %s, %v", metadata, err)
+	}
+}
+
+func TestExtractPreservesUnknownOrMismatchedGroundingForValidators(t *testing.T) {
+	locations := locationRegistry(t, "The Mill")
+	known := locations.Locations[0]
+	client := &fakeOccurrencesLLMClient{response: extractionResponse{Occurrences: []occurrenceResponse{
+		{LocationID: "location:sha256:unknown", Name: "The Mill", Kind: "mentioned", SourceRefs: occurrenceRefs(10, 10)},
+		{LocationID: known.ID, Name: "A Different Mill", Kind: "mentioned", SourceRefs: occurrenceRefs(20, 20)},
+	}}}
+	references := registryReferences(t, locations)
+	req := extractionRequest()
+	req.References = references
+	result, err := newExtractor(t, client, references).Extract(context.Background(), req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if result.Value.Occurrences[0].LocationID != "location:sha256:unknown" || result.Value.Occurrences[1].Name != "A Different Mill" {
+		t.Fatalf("extractor repaired validator-owned grounding errors: %#v", result.Value.Occurrences)
+	}
+}
+
+func TestExtractRequiresRegistryAndAcceptsEmptyRegistryWithNoOccurrences(t *testing.T) {
+	client := &fakeOccurrencesLLMClient{response: extractionResponse{Occurrences: []occurrenceResponse{}}}
+	if _, err := newExtractor(t, client).Extract(context.Background(), extractionRequest()); err == nil || !strings.Contains(err.Error(), "location registry reference is required") {
+		t.Fatalf("Extract() error = %v", err)
+	}
+	if len(client.requests) != 0 {
+		t.Fatalf("LLM calls = %d", len(client.requests))
+	}
+	empty := dnd.LocationList{Locations: []dnd.Location{}}
+	references := registryReferences(t, empty)
+	req := extractionRequest()
+	req.References = references
+	result, err := newExtractor(t, client, references).Extract(context.Background(), req)
+	if err != nil || result.Value.Occurrences == nil || len(result.Value.Occurrences) != 0 {
+		t.Fatalf("empty registry result = %#v, %v", result, err)
+	}
+}
+
+func TestExtractResolvesGeneratedRegistryAtOperationTimeAndDoesNotMutateResponse(t *testing.T) {
+	locations := locationRegistry(t, "The Mill")
+	location := locations.Locations[0]
+	client := &fakeOccurrencesLLMClient{response: extractionResponse{Occurrences: []occurrenceResponse{{
+		LocationID: location.ID, Name: location.Name, Kind: "mentioned", SourceRefs: occurrenceRefs(30, 30),
+	}}}}
+	references := registryReferences(t, locations)
+	req := extractionRequest()
+	req.References = references
+	extractor := newExtractor(t, client)
+	result, err := extractor.Extract(context.Background(), req)
+	if err != nil || result.Value.Occurrences[0].Name != "The Mill" {
+		t.Fatalf("Extract() = %#v, %v", result, err)
+	}
+	if input := client.requests[0].Inputs[LocationRegistryReferenceSlot]; !strings.Contains(string(input.Content), location.ID) || input.OriginURI != "" {
+		t.Fatalf("generated registry prompt input = %#v", input)
+	}
+	if _, ok := extractor.ManifestMetadata()["location_registry_digest"]; ok {
+		t.Fatalf("operation registry leaked into static metadata: %#v", extractor.ManifestMetadata())
+	}
+	if client.response.Occurrences[0].SourceRefs[0].StartUnitID != 30 {
+		t.Fatalf("model response mutated: %#v", client.response)
+	}
+}
+
+func TestExtractorContractsMetadataAndFailures(t *testing.T) {
+	if _, err := New(nil, Options{}); err == nil || !strings.Contains(err.Error(), "LLM client") {
+		t.Fatalf("New(nil) error = %v", err)
+	}
+	if _, err := New(&fakeOccurrencesLLMClient{}, Options{}, contracts.ReferenceSet{}, contracts.ReferenceSet{}); err == nil || !strings.Contains(err.Error(), "at most one reference set") {
+		t.Fatalf("New() error = %v", err)
+	}
+	malformed := contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{LocationRegistryReferenceSlot: {
+		Items: []contracts.ReferenceItem{{SlotName: LocationRegistryReferenceSlot, MediaType: "application/json", Content: []byte(`{"secret":"registry evidence"}`)}},
+	}}}
+	if _, err := New(&fakeOccurrencesLLMClient{}, Options{}, malformed); err == nil || !strings.Contains(err.Error(), "prepare location registry") || strings.Contains(err.Error(), "registry evidence") {
+		t.Fatalf("New() error = %v", err)
+	}
+
+	locations := locationRegistry(t, "The Mill")
+	references := registryReferences(t, locations)
+	req := extractionRequest()
+	req.References = references
+	extractor := newExtractor(t, &fakeOccurrencesLLMClient{}, references)
+	var nilExtractor *Extractor
+	for _, test := range []struct {
+		name      string
+		extractor *Extractor
+		req       contracts.TypedExtractionRequest
+		want      string
+	}{
+		{"nil extractor", nilExtractor, req, "extractor"},
+		{"nil client", &Extractor{}, req, "LLM client"},
+		{"invalid request", extractor, mismatchedSourceInputRequest(req), "must match chunk"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := test.extractor.Extract(context.Background(), test.req); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Extract() error = %v", err)
+			}
+		})
+	}
+	if _, err := newExtractor(t, &fakeOccurrencesLLMClient{err: errors.New("provider unavailable")}, references).Extract(context.Background(), req); err == nil || !strings.Contains(err.Error(), "provider unavailable") {
+		t.Fatalf("provider error = %v", err)
+	}
+
+	spec := ModuleSpec()
+	if spec.Key != Key || spec.Stage != pipeline.StageExtract || spec.ExecutionClass != contracts.ExecutionClassLLMBacked || spec.ArtifactKind != dnd.LocationOccurrenceListKind {
+		t.Fatalf("ModuleSpec() = %#v", spec)
+	}
+	var slot contracts.ReferenceSlot
+	for _, candidate := range spec.ReferenceSlots {
+		if candidate.Name == LocationRegistryReferenceSlot {
+			slot = candidate
+		}
+	}
+	if !slot.Required || !reflect.DeepEqual(slot.AcceptedArtifactKinds, []contracts.ArtifactKind{dnd.LocationListKind}) || slot.MaxBytes != LocationRegistryMaxBytes {
+		t.Fatalf("location registry slot = %#v", slot)
+	}
+	registry := pipeline.NewExtractorRegistry()
+	if err := Register(registry); err != nil {
+		t.Fatal(err)
+	}
+	if _, ok := registry.Spec(Key); !ok {
+		t.Fatalf("registration missing %q", Key)
+	}
+	if _, err := DecodeOptions(map[string]any{"unexpected": true}); err == nil {
+		t.Fatal("DecodeOptions() accepted unknown options")
+	}
+	metadata := newExtractor(t, &fakeOccurrencesLLMClient{}, references).ManifestMetadata()
+	for _, key := range []string{"prompt_sha256", "response_schema_sha256", "location_registry_digest"} {
+		if value, ok := metadata[key].(string); !ok || !strings.HasPrefix(value, "sha256:") {
+			t.Fatalf("metadata[%q] = %#v", key, metadata[key])
+		}
+	}
+	if got := newExtractor(t, &fakeOccurrencesLLMClient{}, references).CheckpointFingerprints(); len(got) != 4 || got[3].Name != "location_registry" {
+		t.Fatalf("fingerprints = %#v", got)
+	}
+}
+
+func locationRegistry(t *testing.T, names ...string) dnd.LocationList {
+	t.Helper()
+	locations := make([]dnd.Location, len(names))
+	for index, name := range names {
+		refs := []source.SourceRef{{SourceID: "other-session", StartUnitID: index + 1, EndUnitID: index + 1}}
+		locations[index] = dnd.Location{ID: identity.DeriveID(name, refs), Name: name, SourceRefs: refs}
+	}
+	return dnd.LocationList{Locations: locations}
+}
+
+func registryReferences(t *testing.T, locations dnd.LocationList) contracts.ReferenceSet {
+	t.Helper()
+	content, err := locationcodec.New().Encode(locations)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{LocationRegistryReferenceSlot: {
+		Slot:  contracts.ReferenceSlot{Name: LocationRegistryReferenceSlot},
+		Items: []contracts.ReferenceItem{{SlotName: LocationRegistryReferenceSlot, MediaType: locationcodec.MediaType, Content: content, Origin: contracts.ReferenceOrigin{Type: "generated"}}},
+	}}}
+}

internal/modules/dnd/extract/locationoccurrences/model.go

@@ -0,0 +1,17 @@
+package locationoccurrences
+
+type extractionResponse struct {
+	Occurrences []occurrenceResponse `json:"occurrences"`
+}
+
+type occurrenceResponse struct {
+	LocationID string                        `json:"location_id"`
+	Name       string                        `json:"name"`
+	Kind       string                        `json:"kind"`
+	SourceRefs []occurrenceSourceRefResponse `json:"source_refs"`
+}
+
+type occurrenceSourceRefResponse struct {
+	StartUnitID int `json:"start_unit_id"`
+	EndUnitID   int `json:"end_unit_id"`
+}

internal/modules/dnd/extract/locationoccurrences/prompt_assets.go

@@ -0,0 +1,53 @@
+package locationoccurrences
+
+import (
+	"fmt"
+	"sync"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/promptfs"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const promptAssetRoot = "assets/prompts"
+
+var promptAssetManifest = shared.PromptAssetManifest{
+	ModuleDir: PromptID,
+	ModuleFiles: []promptfs.ModulePromptFile{
+		{Name: "dnd.location_occurrences.yaml", Path: "assets/prompts/dnd.location_occurrences.yaml"},
+		{Name: "locations.md", Path: "assets/prompts/locations.md"},
+		{Name: "task.md", Path: "assets/prompts/task.md"},
+		{Name: "instructions.md", Path: "assets/prompts/instructions.md"},
+	},
+	SharedFiles: []string{
+		"common-dnd-system.md",
+		"common-dnd-extraction-evidence.md",
+		"common-dnd-identity.md",
+		"common-dnd-transcript.md",
+		"common-dnd-references.md",
+	},
+}
+
+func RegisterPromptAssets(registry *llm.AssetRegistry) error {
+	promptFS, err := promptAssetManifest.PromptFS(embeddedAssets)
+	if err != nil {
+		return fmt.Errorf("prepare location-occurrence prompt assets: %w", err)
+	}
+	if err := registry.RegisterPromptFS(promptFS, promptAssetRoot); err != nil {
+		return err
+	}
+	return registry.RegisterSchemaFS(embeddedAssets, "assets/schemas")
+}
+
+func promptAssetMetadata() (string, error) {
+	promptAssetHashOnce.Do(func() {
+		promptAssetHash, promptAssetHashErr = promptAssetManifest.Hash(embeddedAssets)
+	})
+	return promptAssetHash, promptAssetHashErr
+}
+
+var (
+	promptAssetHashOnce sync.Once
+	promptAssetHash     string
+	promptAssetHashErr  error
+)

internal/modules/dnd/extract/locationoccurrences/prompt_assets_test.go

@@ -0,0 +1,64 @@
+package locationoccurrences
+
+import (
+	"context"
+	"strings"
+	"testing"
+	"time"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/promptkit"
+)
+
+func TestRegisterPromptAssetsPreparesLocationOccurrencePrompt(t *testing.T) {
+	registry := llm.NewAssetRegistry()
+	if err := RegisterPromptAssets(registry); err != nil {
+		t.Fatal(err)
+	}
+	options, err := registry.PromptKitOptions()
+	if err != nil {
+		t.Fatal(err)
+	}
+	options = append(options, promptkit.WithProfiles(promptkit.OpenAICompatibleProfile(promptkit.OpenAICompatibleProfileConfig{
+		ID: "location-occurrences-test", Endpoint: "http://127.0.0.1:1/v1", Model: "test",
+	})))
+	engine, err := promptkit.NewEngine(promptkit.Config{Timeout: time.Second}, options...)
+	if err != nil {
+		t.Fatal(err)
+	}
+	prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{
+		PromptID: PromptID, PromptVersion: SchemaVersion, ProfileID: "location-occurrences-test",
+		Inputs: map[string]promptkit.ArtifactRef{
+			"transcript": promptkit.Inline(`{"units":[1]}`), "players": promptkit.Inline(" "), "party": promptkit.Inline(" "), "glossary": promptkit.Inline(" "),
+			"locations": promptkit.Inline(`{"locations":[{"id":"location:sha256:test","name":"The Mill"}]}`),
+		},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if prepared.PromptID != PromptID || prepared.OutputContract.SchemaPath != "dnd_location_occurrences_llm.v1.json" {
+		t.Fatalf("prepared prompt = %#v", prepared)
+	}
+	var registryMessage string
+	registryIndex := -1
+	evidenceIndex := -1
+	taskIndex := -1
+	for index, message := range prepared.Messages {
+		if strings.Contains(message.Content, "normalized location registry") {
+			registryMessage = message.Content
+			registryIndex = index
+		}
+		if strings.Contains(message.Content, "Transcript units are the only evidence") {
+			evidenceIndex = index
+		}
+		if strings.Contains(message.Content, "Extract Dungeons & Dragons location occurrences") {
+			taskIndex = index
+		}
+	}
+	if !strings.Contains(registryMessage, "location:sha256:test") || !strings.Contains(registryMessage, "The Mill") || strings.Contains(registryMessage, "source_refs") {
+		t.Fatalf("rendered prompt did not preserve source-free registry grounding: %s", registryMessage)
+	}
+	if evidenceIndex < 0 || taskIndex < 0 || registryIndex <= evidenceIndex || registryIndex >= taskIndex {
+		t.Fatalf("registry prompt placement = evidence %d, registry %d, task %d", evidenceIndex, registryIndex, taskIndex)
+	}
+}

internal/modules/dnd/extract/locationoccurrences/schema.go

@@ -0,0 +1,21 @@
+package locationoccurrences
+
+import "gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+
+const (
+	PromptID           = "dnd.location_occurrences"
+	ResponseSchemaKey  = llm.ResponseSchemaKey("dnd_location_occurrences_llm")
+	ResponseSchemaID   = "notarius.dnd.location_occurrences.llm"
+	ResponseSchemaName = "notarius_dnd_location_occurrences_llm_v1"
+	SchemaVersion      = "v1"
+)
+
+func loadResponseSchema() (llm.ResponseSchema, error) {
+	return llm.LoadResponseSchema(embeddedAssets, llm.ResponseSchemaDefinition{
+		Key:       ResponseSchemaKey,
+		ID:        ResponseSchemaID,
+		Version:   SchemaVersion,
+		Name:      ResponseSchemaName,
+		AssetPath: "assets/schemas/dnd_location_occurrences_llm.v1.json",
+	})
+}

internal/modules/dnd/extract/locationoccurrences/schema_test.go

@@ -0,0 +1,96 @@
+package locationoccurrences
+
+import (
+	"bytes"
+	"encoding/json"
+	"strings"
+	"testing"
+
+	"github.com/santhosh-tekuri/jsonschema/v6"
+)
+
+func TestResponseSchemaRestrictsPrivateOccurrenceStructureAndKinds(t *testing.T) {
+	schema, err := loadResponseSchema()
+	if err != nil {
+		t.Fatal(err)
+	}
+	if schema.Key != ResponseSchemaKey || schema.ID != ResponseSchemaID || schema.Name != ResponseSchemaName || schema.Version != SchemaVersion || !strings.HasPrefix(schema.SHA256, "sha256:") || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("schema = %#v", schema)
+	}
+	valid := map[string]any{"occurrences": []any{map[string]any{
+		"location_id": "location:sha256:test", "name": "The Mill", "kind": "visited",
+		"source_refs": []any{map[string]any{"start_unit_id": 1, "end_unit_id": 2}},
+	}}}
+	if err := validateSchema(t, valid, schema.JSONSchema); err != nil {
+		t.Fatalf("valid response rejected: %v", err)
+	}
+	for _, mutate := range []func(map[string]any){
+		func(value map[string]any) { delete(value, "location_id") },
+		func(value map[string]any) { value["kind"] = "other" },
+		func(value map[string]any) { value["unexpected"] = true },
+		func(value map[string]any) {
+			value["source_refs"].([]any)[0].(map[string]any)["source_id"] = "assigned later"
+		},
+	} {
+		candidate := cloneCandidate(t, valid)
+		mutate(candidate["occurrences"].([]any)[0].(map[string]any))
+		if err := validateSchema(t, candidate, schema.JSONSchema); err == nil {
+			t.Fatal("schema accepted structurally invalid response")
+		}
+	}
+}
+
+func TestResponseSchemaIsDefensiveAndContentSafe(t *testing.T) {
+	first, err := loadResponseSchema()
+	if err != nil {
+		t.Fatal(err)
+	}
+	first.JSONSchema[0] = '['
+	second, err := loadResponseSchema()
+	if err != nil || !json.Valid(second.JSONSchema) || bytes.Equal(first.JSONSchema, second.JSONSchema) {
+		t.Fatalf("schema = %s, %v", second.JSONSchema, err)
+	}
+	if diagnostics := second.DiagnosticsMap(); diagnostics["key"] != ResponseSchemaKey || diagnostics["id"] != ResponseSchemaID {
+		t.Fatalf("diagnostics = %#v", diagnostics)
+	} else if _, ok := diagnostics["json_schema"]; ok {
+		t.Fatalf("diagnostics leaked schema content: %#v", diagnostics)
+	}
+}
+
+func cloneCandidate(t *testing.T, value map[string]any) map[string]any {
+	t.Helper()
+	content, err := json.Marshal(value)
+	if err != nil {
+		t.Fatal(err)
+	}
+	var clone map[string]any
+	if err := json.Unmarshal(content, &clone); err != nil {
+		t.Fatal(err)
+	}
+	return clone
+}
+
+func validateSchema(t *testing.T, value map[string]any, schemaContent []byte) error {
+	t.Helper()
+	content, err := json.Marshal(value)
+	if err != nil {
+		return err
+	}
+	instance, err := jsonschema.UnmarshalJSON(bytes.NewReader(content))
+	if err != nil {
+		return err
+	}
+	document, err := jsonschema.UnmarshalJSON(bytes.NewReader(schemaContent))
+	if err != nil {
+		return err
+	}
+	compiler := jsonschema.NewCompiler()
+	if err := compiler.AddResource("schema.json", document); err != nil {
+		return err
+	}
+	schema, err := compiler.Compile("schema.json")
+	if err != nil {
+		return err
+	}
+	return schema.Validate(instance)
+}

internal/modules/dnd/extract/locationoccurrences/test_helpers_test.go

@@ -0,0 +1,82 @@
+package locationoccurrences
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+)
+
+func extractionRequest() contracts.TypedExtractionRequest {
+	doc := sourceDocument()
+	chunk := &source.Chunk{
+		ID: "session-occurrences:chunk:0", SourceID: doc.ID, Index: 0,
+		Ref:     source.SourceRef{SourceID: doc.ID, StartUnitID: 10, EndUnitID: 30},
+		Content: []byte(`{"units":[10,20,30]}`), MediaType: "application/json",
+		Units: append([]source.SourceUnit(nil), doc.Units...),
+	}
+	return contracts.TypedExtractionRequest{
+		Source: doc, Chunk: chunk,
+		SourceInput: contracts.NewLLMInputMaterial("source", chunk.MediaType, chunk.Content, "sha256:chunk", "file:///session-occurrences.json"),
+		SessionID:   "occurrence-session", LLMProfile: "occurrence-profile",
+	}
+}
+
+func sourceDocument() *source.SourceDocument {
+	return &source.SourceDocument{ID: "session-occurrences", Kind: "transcript", Format: "application/json", Digest: "sha256:test", Units: []source.SourceUnit{
+		{ID: 10, Kind: "transcript_segment", Text: "The party returns to the tavern."},
+		{ID: 20, Kind: "transcript_segment", Text: "They plan to travel to the tavern tomorrow."},
+		{ID: 30, Kind: "transcript_segment", Text: "They recall their first visit to the tavern."},
+	}}
+}
+
+func occurrenceRefs(start, end int) []occurrenceSourceRefResponse {
+	return []occurrenceSourceRefResponse{{StartUnitID: start, EndUnitID: end}}
+}
+
+func newExtractor(t *testing.T, client contracts.StructuredLLMClient, references ...contracts.ReferenceSet) *Extractor {
+	t.Helper()
+	extractor, err := New(client, Options{}, references...)
+	if err != nil {
+		t.Fatalf("New() error = %v", err)
+	}
+	return extractor
+}
+
+func mismatchedSourceInputRequest(req contracts.TypedExtractionRequest) contracts.TypedExtractionRequest {
+	req.SourceInput = contracts.NewLLMInputMaterial("source", "application/json", []byte(`{"different":true}`), "sha256:other", "")
+	return req
+}
+
+type fakeOccurrencesLLMClient struct {
+	response extractionResponse
+	err      error
+	requests []contracts.StructuredCompletionRequest
+}
+
+func (client *fakeOccurrencesLLMClient) CompleteStructured(_ context.Context, req contracts.StructuredCompletionRequest, out any) (contracts.StructuredCompletionResponse, error) {
+	client.requests = append(client.requests, cloneStructuredCompletionRequest(req))
+	if client.err != nil {
+		return contracts.StructuredCompletionResponse{}, client.err
+	}
+	target, ok := out.(*extractionResponse)
+	if !ok {
+		return contracts.StructuredCompletionResponse{}, errors.New("unexpected output target")
+	}
+	content, err := json.Marshal(client.response)
+	if err != nil {
+		return contracts.StructuredCompletionResponse{}, err
+	}
+	if err := json.Unmarshal(content, target); err != nil {
+		return contracts.StructuredCompletionResponse{}, err
+	}
+	return contracts.StructuredCompletionResponse{Content: content}, nil
+}
+
+func cloneStructuredCompletionRequest(req contracts.StructuredCompletionRequest) contracts.StructuredCompletionRequest {
+	req.Inputs = req.Inputs.Clone()
+	return req
+}

internal/modules/dnd/extract/locations/assets.go

@@ -0,0 +1,6 @@
+package locations
+
+import "embed"
+
+//go:embed assets/prompts/*.yaml assets/prompts/*.md assets/schemas/*.json
+var embeddedAssets embed.FS

internal/modules/dnd/extract/locations/assets/prompts/dnd.locations.yaml

@@ -0,0 +1,42 @@
+id: dnd.locations
+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.md
+    cache_control:
+      type: ephemeral
+  - role: user
+    content_file: ./sharedassets/common-dnd-extraction-evidence.md
+  - role: user
+    content_file: ./task.md
+  - role: user
+    content_file: ./instructions.md
+    cache_control:
+      type: ephemeral
+output:
+  format: json
+  validation_mode: json_schema
+  schema_path: dnd_locations_llm.v1.json
+  repair_attempts: 0

internal/modules/dnd/extract/locations/assets/prompts/instructions.md

@@ -0,0 +1,6 @@
+Return only observed location display names and narrow transcript source ranges.
+
+Exclude people, creatures, objects, organizations, abstract concepts, and
+places merely inferred from an event. Omit uncertain or unsupported places.
+Campaign references may clarify terms already present in the transcript, but
+they are not evidence and must never supply a source range.

internal/modules/dnd/extract/locations/assets/prompts/task.md

@@ -0,0 +1,8 @@
+Extract physical places established by the provided Dungeons & Dragons
+transcript and cite where each place is identified.
+
+Include planes, regions, settlements, districts, buildings, rooms, landmarks,
+routes, and geographic features. A generic label such as "the tavern" is
+allowed only when the transcript uses it for a specific place. Keep aliases and
+nested places when the transcript identifies them; do not merge or invent
+qualifiers for similarly named places.

internal/modules/dnd/extract/locations/assets/schemas/dnd_locations_llm.v1.json

@@ -0,0 +1,32 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "notarius.dnd.locations.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"}
+              }
+            }
+          }
+        }
+      }
+    }
+  }
+}

internal/modules/dnd/extract/locations/canonicalize.go

@@ -0,0 +1,86 @@
+package locations
+
+import (
+	"sort"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+type orderedLocationResponse struct {
+	value       locationResponse
+	earliest    int
+	hasEvidence bool
+}
+
+func canonicalizeResponse(response *extractionResponse, order shared.SourceRefOrder, sourceID string) {
+	if response == nil {
+		return
+	}
+	ordered := make([]orderedLocationResponse, len(response.Locations))
+	for index := range response.Locations {
+		earliest, hasEvidence := canonicalizeLocation(&response.Locations[index], order, sourceID)
+		ordered[index] = orderedLocationResponse{value: response.Locations[index], earliest: earliest, hasEvidence: hasEvidence}
+	}
+	sort.SliceStable(ordered, func(left, right int) bool {
+		if ordered[left].hasEvidence != ordered[right].hasEvidence {
+			return ordered[left].hasEvidence
+		}
+		if !ordered[left].hasEvidence {
+			return false
+		}
+		return ordered[left].earliest < ordered[right].earliest
+	})
+	for index := range ordered {
+		response.Locations[index] = ordered[index].value
+	}
+}
+
+func canonicalizeLocation(location *locationResponse, order shared.SourceRefOrder, sourceID string) (int, bool) {
+	if location == nil {
+		return 0, false
+	}
+	refs := order.Canonicalize(canonicalSourceRefs(location.SourceRefs, sourceID))
+	location.SourceRefs = locationResponseRefs(refs)
+	return order.EarliestValid(refs)
+}
+
+func canonicalLocationList(response extractionResponse, sourceID string) dnd.LocationList {
+	if response.Locations == nil {
+		return dnd.LocationList{Locations: nil}
+	}
+	locations := make([]dnd.Location, len(response.Locations))
+	for index, location := range response.Locations {
+		refs := canonicalSourceRefs(location.SourceRefs, sourceID)
+		locations[index] = dnd.Location{
+			ID:         identity.DeriveID(location.Name, refs),
+			Name:       location.Name,
+			SourceRefs: refs,
+		}
+	}
+	return dnd.LocationList{Locations: locations}
+}
+
+func canonicalSourceRefs(values []locationSourceRefResponse, sourceID string) []source.SourceRef {
+	if values == nil {
+		return nil
+	}
+	refs := make([]source.SourceRef, len(values))
+	for index, value := range values {
+		refs[index] = source.SourceRef{SourceID: sourceID, StartUnitID: value.StartUnitID, EndUnitID: value.EndUnitID}
+	}
+	return refs
+}
+
+func locationResponseRefs(values []source.SourceRef) []locationSourceRefResponse {
+	if values == nil {
+		return nil
+	}
+	refs := make([]locationSourceRefResponse, len(values))
+	for index, value := range values {
+		refs[index] = locationSourceRefResponse{StartUnitID: value.StartUnitID, EndUnitID: value.EndUnitID}
+	}
+	return refs
+}

internal/modules/dnd/extract/locations/extractor.go

@@ -0,0 +1,152 @@
+// Package locations extracts source-grounded D&D physical location candidates.
+package locations
+
+import (
+	"context"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const (
+	Key           = "dnd/locations"
+	mappingPolicy = "dnd.locations.extract_mapping.v1"
+)
+
+var requiredCapabilities = []string{"chunks", "source.transcript"}
+var providedCapabilities = []string{"dnd.locations"}
+
+var referenceSlotDescriptions = shared.ReferenceSlotDescriptions{
+	Glossary: "Optional campaign glossary reference material used only for location disambiguation.",
+	Party:    "Optional party roster reference material used only for location disambiguation.",
+	Players:  "Optional player list reference material used only for location disambiguation.",
+	Roster:   "Deprecated alias for party roster reference material used only for location disambiguation.",
+}
+
+func referenceSlots() []contracts.ReferenceSlot {
+	return shared.ReferenceSlots(referenceSlotDescriptions)
+}
+
+var _ contracts.Extractor[dnd.LocationList] = (*Extractor)(nil)
+var _ contracts.ManifestMetadataProvider = (*Extractor)(nil)
+var _ pipeline.CheckpointFingerprintProvider = (*Extractor)(nil)
+
+type Options struct{}
+
+type Extractor struct {
+	llm               contracts.StructuredLLMClient
+	promptSHA         string
+	responseSchemaSHA string
+}
+
+func New(llmClient contracts.StructuredLLMClient, _ Options, references ...contracts.ReferenceSet) (*Extractor, error) {
+	if llmClient == nil {
+		return nil, extractorErrorf("LLM client must not be nil")
+	}
+	if len(references) > 1 {
+		return nil, extractorErrorf("at most one reference set may be supplied")
+	}
+	promptSHA, err := promptAssetMetadata()
+	if err != nil {
+		return nil, extractorErrorf("load prompt metadata: %w", err)
+	}
+	responseSchema, err := loadResponseSchema()
+	if err != nil {
+		return nil, extractorErrorf("load response schema: %w", err)
+	}
+	return &Extractor{llm: llmClient, promptSHA: promptSHA, responseSchemaSHA: responseSchema.SHA256}, nil
+}
+
+func (e *Extractor) Key() string { return Key }
+
+func (e *Extractor) ReferenceSlots() []contracts.ReferenceSlot { return referenceSlots() }
+
+func (e *Extractor) ManifestMetadata() map[string]any {
+	if e == nil {
+		return nil
+	}
+	return map[string]any{
+		"prompt_id":               PromptID,
+		"prompt_version":          SchemaVersion,
+		"prompt_sha256":           e.promptSHA,
+		"response_schema_key":     string(ResponseSchemaKey),
+		"response_schema_id":      ResponseSchemaID,
+		"response_schema_name":    ResponseSchemaName,
+		"response_schema_version": SchemaVersion,
+		"response_schema_sha256":  e.responseSchemaSHA,
+		"identity_policy":         identity.Policy,
+		"mapping_policy":          mappingPolicy,
+	}
+}
+
+func (e *Extractor) CheckpointFingerprints() []pipeline.CheckpointFingerprint {
+	if e == nil {
+		return nil
+	}
+	return []pipeline.CheckpointFingerprint{
+		{Name: "prompt", Value: e.promptSHA},
+		{Name: "response_schema", Value: e.responseSchemaSHA},
+		{Name: "identity_policy", Value: identity.Policy},
+		{Name: "mapping_policy", Value: mappingPolicy},
+	}
+}
+
+func (e *Extractor) Extract(ctx context.Context, req contracts.TypedExtractionRequest) (contracts.TypedExtractionResult[dnd.LocationList], error) {
+	if e == nil {
+		return contracts.TypedExtractionResult[dnd.LocationList]{}, extractorErrorf("extractor must not be nil")
+	}
+	if e.llm == nil {
+		return contracts.TypedExtractionResult[dnd.LocationList]{}, extractorErrorf("LLM client must not be nil")
+	}
+	sourceInput, err := shared.PrepareChunkExtraction(ctx, req)
+	if err != nil {
+		return contracts.TypedExtractionResult[dnd.LocationList]{}, extractorErrorf("%w", err)
+	}
+	order := shared.NewSourceRefOrder(req.Source)
+
+	var response extractionResponse
+	if _, err := e.llm.CompleteStructured(ctx, contracts.StructuredCompletionRequest{
+		StageName: Key, PromptID: PromptID, PromptVersion: SchemaVersion,
+		ProfileID: req.LLMProfile, SessionID: req.SessionID,
+		Inputs: shared.PromptInputs(sourceInput, req.References),
+	}, &response); err != nil {
+		return contracts.TypedExtractionResult[dnd.LocationList]{}, extractorErrorf("complete structured output: %w", err)
+	}
+	canonicalizeResponse(&response, order, req.Source.ID)
+	return contracts.TypedExtractionResult[dnd.LocationList]{Value: canonicalLocationList(response, req.Source.ID)}, nil
+}
+
+func ModuleSpec() pipeline.ModuleSpec {
+	return pipeline.ModuleSpec{
+		Key: Key, Stage: pipeline.StageExtract, ExecutionClass: contracts.ExecutionClassLLMBacked,
+		Requires: append([]string(nil), requiredCapabilities...), Provides: append([]string(nil), providedCapabilities...),
+		ArtifactKind: dnd.LocationListKind, ReferenceSlots: referenceSlots(),
+	}
+}
+
+func Register(registry *pipeline.ExtractorRegistry) error {
+	return pipeline.RegisterExtractorBuilder(registry, ModuleSpec(), validateOptions, func(request pipeline.BuildRequest) (contracts.Extractor[dnd.LocationList], error) {
+		options, err := DecodeOptions(request.Options)
+		if err != nil {
+			return nil, err
+		}
+		return New(request.Dependencies.LLM, options, request.References)
+	})
+}
+
+func validateOptions(options map[string]any) error { _, err := DecodeOptions(options); return err }
+
+func DecodeOptions(options map[string]any) (Options, error) {
+	if err := pipeline.RejectUnknownOptions(options); err != nil {
+		return Options{}, extractorErrorf("%w", err)
+	}
+	return Options{}, nil
+}
+
+func extractorErrorf(format string, args ...any) error {
+	return fmt.Errorf("dnd locations extractor: "+format, args...)
+}

internal/modules/dnd/extract/locations/extractor_test.go

@@ -0,0 +1,125 @@
+package locations
+
+import (
+	"context"
+	"errors"
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+func TestExtractMapsLocationsWithOwnedEvidenceAndDeterministicOrder(t *testing.T) {
+	client := &fakeLocationsLLMClient{response: extractionResponse{Locations: []locationResponse{
+		{Name: "The Tavern", SourceRefs: responseSourceRefs(3, 3)},
+		{Name: "Old Mill", SourceRefs: []locationSourceRefResponse{{StartUnitID: 2, EndUnitID: 2}, {StartUnitID: 1, EndUnitID: 1}, {StartUnitID: 1, EndUnitID: 1}}},
+	}}}
+	result, err := newExtractor(t, client).Extract(context.Background(), extractionRequest())
+	if err != nil {
+		t.Fatalf("Extract() error = %v, want nil", err)
+	}
+	refs := []source.SourceRef{{SourceID: "session-locations", StartUnitID: 1, EndUnitID: 1}, {SourceID: "session-locations", StartUnitID: 2, EndUnitID: 2}}
+	want := dnd.LocationList{Locations: []dnd.Location{
+		{ID: identity.DeriveID("Old Mill", refs), Name: "Old Mill", SourceRefs: refs},
+		{ID: identity.DeriveID("The Tavern", []source.SourceRef{{SourceID: "session-locations", StartUnitID: 3, EndUnitID: 3}}), Name: "The Tavern", SourceRefs: []source.SourceRef{{SourceID: "session-locations", StartUnitID: 3, EndUnitID: 3}}},
+	}}
+	if !reflect.DeepEqual(result.Value, want) {
+		t.Fatalf("Value = %#v, want %#v", result.Value, want)
+	}
+	result.Value.Locations[0].SourceRefs[0].StartUnitID = 99
+	for _, location := range client.response.Locations {
+		for _, ref := range location.SourceRefs {
+			if ref.StartUnitID == 99 {
+				t.Fatal("result source references alias the model response")
+			}
+		}
+	}
+}
+
+func TestExtractRetainsSameNameLocationsAtDifferentAnchors(t *testing.T) {
+	client := &fakeLocationsLLMClient{response: extractionResponse{Locations: []locationResponse{
+		{Name: "the tavern", SourceRefs: responseSourceRefs(1, 1)},
+		{Name: "the tavern", SourceRefs: responseSourceRefs(3, 3)},
+	}}}
+	result, err := newExtractor(t, client).Extract(context.Background(), extractionRequest())
+	if err != nil || len(result.Value.Locations) != 2 {
+		t.Fatalf("Extract() = %#v, %v; want both same-name candidates", result, err)
+	}
+	if result.Value.Locations[0].ID == result.Value.Locations[1].ID || result.Value.Locations[0].Name != result.Value.Locations[1].Name {
+		t.Fatalf("locations = %#v, want distinct evidence-anchored IDs", result.Value.Locations)
+	}
+}
+
+func TestExtractPreservesInvalidCandidatesForValidators(t *testing.T) {
+	client := &fakeLocationsLLMClient{content: []byte(`{"locations":[{"name":"","source_refs":[{"start_unit_id":0,"end_unit_id":-1}]}]}`)}
+	result, err := newExtractor(t, client).Extract(context.Background(), extractionRequest())
+	if err != nil {
+		t.Fatalf("Extract() error = %v, want nil", err)
+	}
+	location := result.Value.Locations[0]
+	if location.ID != "" || location.Name != "" || !reflect.DeepEqual(location.SourceRefs, []source.SourceRef{{SourceID: "session-locations", StartUnitID: 0, EndUnitID: -1}}) {
+		t.Fatalf("location = %#v, want invalid candidate preserved", location)
+	}
+}
+
+func TestExtractPassesReferencesWithoutTreatingThemAsEvidence(t *testing.T) {
+	client := &fakeLocationsLLMClient{response: extractionResponse{Locations: []locationResponse{}}}
+	req := extractionRequest()
+	req.References = contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{
+		"glossary": {Slot: contracts.ReferenceSlot{Name: "glossary"}, Items: []contracts.ReferenceItem{{SlotName: "glossary", Content: []byte("Old Mill: abandoned granary")}}},
+	}}
+	if _, err := newExtractor(t, client).Extract(context.Background(), req); err != nil {
+		t.Fatalf("Extract() error = %v", err)
+	}
+	inputs := client.requests[0].Inputs
+	if string(inputs["glossary"].Content) != "Old Mill: abandoned granary" || strings.Contains(string(inputs["transcript"].Content), "abandoned granary") {
+		t.Fatalf("prompt inputs = %#v, want separated reference material", inputs)
+	}
+}
+
+func TestExtractDoesNotMutateRequestMaterials(t *testing.T) {
+	client := &fakeLocationsLLMClient{response: extractionResponse{Locations: []locationResponse{{Name: "Old Mill", SourceRefs: responseSourceRefs(1, 1)}}}}
+	req := extractionRequest()
+	beforeUnits := append([]source.SourceUnit(nil), req.Source.Units...)
+	beforeChunkUnits := append([]source.SourceUnit(nil), req.Chunk.Units...)
+	beforeContent := append([]byte(nil), req.Chunk.Content...)
+	if _, err := newExtractor(t, client).Extract(context.Background(), req); err != nil {
+		t.Fatalf("Extract() error = %v", err)
+	}
+	if !reflect.DeepEqual(req.Source.Units, beforeUnits) || !reflect.DeepEqual(req.Chunk.Units, beforeChunkUnits) || !reflect.DeepEqual(req.Chunk.Content, beforeContent) {
+		t.Fatalf("Extract() mutated request: %#v", req)
+	}
+}
+
+func TestExtractHandlesEmptyOutputAndLocalFailures(t *testing.T) {
+	empty, err := newExtractor(t, &fakeLocationsLLMClient{response: extractionResponse{Locations: []locationResponse{}}}).Extract(context.Background(), extractionRequest())
+	if err != nil || empty.Value.Locations == nil || len(empty.Value.Locations) != 0 {
+		t.Fatalf("empty Extract() = %#v, %v; want empty list", empty, err)
+	}
+	request := extractionRequest()
+	var nilExtractor *Extractor
+	for _, test := range []struct {
+		name      string
+		extractor *Extractor
+		req       contracts.TypedExtractionRequest
+		want      string
+	}{
+		{name: "nil extractor", extractor: nilExtractor, req: request, want: "extractor"},
+		{name: "nil client", extractor: &Extractor{}, req: request, want: "LLM client"},
+		{name: "preflight", extractor: newExtractor(t, &fakeLocationsLLMClient{}), req: mismatchedSourceInputRequest(request), want: "must match chunk"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := test.extractor.Extract(context.Background(), test.req); err == nil || !strings.Contains(err.Error(), "dnd locations") || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Extract() error = %v, want local context", err)
+			}
+		})
+	}
+	_, err = newExtractor(t, &fakeLocationsLLMClient{err: errors.New("provider unavailable")}).Extract(context.Background(), request)
+	if err == nil || !strings.Contains(err.Error(), "dnd locations") || !strings.Contains(err.Error(), "provider unavailable") {
+		t.Fatalf("provider error = %v, want contextual provider error", err)
+	}
+}

internal/modules/dnd/extract/locations/model.go

@@ -0,0 +1,15 @@
+package locations
+
+type extractionResponse struct {
+	Locations []locationResponse `json:"locations"`
+}
+
+type locationResponse struct {
+	Name       string                      `json:"name"`
+	SourceRefs []locationSourceRefResponse `json:"source_refs"`
+}
+
+type locationSourceRefResponse struct {
+	StartUnitID int `json:"start_unit_id"`
+	EndUnitID   int `json:"end_unit_id"`
+}

internal/modules/dnd/extract/locations/prompt_assets.go

@@ -0,0 +1,52 @@
+package locations
+
+import (
+	"fmt"
+	"sync"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/promptfs"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
+)
+
+const promptAssetRoot = "assets/prompts"
+
+var promptAssetManifest = shared.PromptAssetManifest{
+	ModuleDir: PromptID,
+	ModuleFiles: []promptfs.ModulePromptFile{
+		{Name: "dnd.locations.yaml", Path: "assets/prompts/dnd.locations.yaml"},
+		{Name: "task.md", Path: "assets/prompts/task.md"},
+		{Name: "instructions.md", Path: "assets/prompts/instructions.md"},
+	},
+	SharedFiles: []string{
+		"common-dnd-system.md",
+		"common-dnd-extraction-evidence.md",
+		"common-dnd-identity.md",
+		"common-dnd-transcript.md",
+		"common-dnd-references.md",
+	},
+}
+
+func RegisterPromptAssets(registry *llm.AssetRegistry) error {
+	promptFS, err := promptAssetManifest.PromptFS(embeddedAssets)
+	if err != nil {
+		return fmt.Errorf("prepare location prompt assets: %w", err)
+	}
+	if err := registry.RegisterPromptFS(promptFS, promptAssetRoot); err != nil {
+		return err
+	}
+	return registry.RegisterSchemaFS(embeddedAssets, "assets/schemas")
+}
+
+func promptAssetMetadata() (string, error) {
+	promptAssetHashOnce.Do(func() {
+		promptAssetHash, promptAssetHashErr = promptAssetManifest.Hash(embeddedAssets)
+	})
+	return promptAssetHash, promptAssetHashErr
+}
+
+var (
+	promptAssetHashOnce sync.Once
+	promptAssetHash     string
+	promptAssetHashErr  error
+)

internal/modules/dnd/extract/locations/prompt_assets_test.go

@@ -0,0 +1,50 @@
+package locations
+
+import (
+	"context"
+	"slices"
+	"strings"
+	"testing"
+	"time"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+	"gitea.maximumdirect.net/eric/promptkit"
+)
+
+func TestRegisterPromptAssetsPreparesLocationPrompt(t *testing.T) {
+	for _, name := range []string{"common-dnd-system.md", "common-dnd-identity.md", "common-dnd-references.md", "common-dnd-transcript.md", "common-dnd-extraction-evidence.md"} {
+		if !slices.Contains(promptAssetManifest.SharedFiles, name) {
+			t.Fatalf("shared prompt assets = %#v, missing %q", promptAssetManifest.SharedFiles, name)
+		}
+	}
+	registry := llm.NewAssetRegistry()
+	if err := RegisterPromptAssets(registry); err != nil {
+		t.Fatalf("RegisterPromptAssets() error = %v", err)
+	}
+	options, err := registry.PromptKitOptions()
+	if err != nil {
+		t.Fatalf("PromptKitOptions() error = %v", err)
+	}
+	options = append(options, promptkit.WithProfiles(promptkit.OpenAICompatibleProfile(promptkit.OpenAICompatibleProfileConfig{ID: "location-test-profile", Endpoint: "http://127.0.0.1:1/v1", Model: "location-test-model"})))
+	engine, err := promptkit.NewEngine(promptkit.Config{Timeout: time.Second}, options...)
+	if err != nil {
+		t.Fatalf("NewEngine() error = %v", err)
+	}
+	prepared, err := engine.Prepare(context.Background(), promptkit.RunRequest{PromptID: PromptID, PromptVersion: SchemaVersion, ProfileID: "location-test-profile", Inputs: map[string]promptkit.ArtifactRef{
+		"transcript": promptkit.Inline(`{"units":[1]}`), "players": promptkit.Inline(" "), "party": promptkit.Inline(" "), "glossary": promptkit.Inline(" "),
+	}})
+	if err != nil {
+		t.Fatalf("Prepare() error = %v", err)
+	}
+	if prepared.OutputContract.SchemaPath != "dnd_locations_llm.v1.json" {
+		t.Fatalf("output contract = %#v", prepared.OutputContract)
+	}
+	if len(prepared.Messages) != 7 || !strings.Contains(prepared.Messages[3].Content, `"units"`) || strings.Contains(prepared.Messages[3].Content, "location-test") {
+		t.Fatalf("prepared messages = %#v, want rendered transcript only in transcript message", prepared.Messages)
+	}
+	for _, index := range []int{2, 3, 6} {
+		if prepared.Messages[index].CacheControl == nil || prepared.Messages[index].CacheControl.Type != promptkit.CacheControlEphemeral {
+			t.Fatalf("message %d cache control = %#v, want ephemeral", index, prepared.Messages[index].CacheControl)
+		}
+	}
+}

internal/modules/dnd/extract/locations/registry_test.go

@@ -0,0 +1,53 @@
+package locations
+
+import (
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+func TestModuleRegistrationAndMetadata(t *testing.T) {
+	if _, err := New(nil, Options{}); err == nil || !strings.Contains(err.Error(), "LLM client") {
+		t.Fatalf("New(nil) error = %v, want client rejection", err)
+	}
+	if _, err := New(&fakeLocationsLLMClient{}, Options{}, contracts.ReferenceSet{}, contracts.ReferenceSet{}); err == nil || !strings.Contains(err.Error(), "at most one") {
+		t.Fatalf("New() error = %v, want reference-set rejection", err)
+	}
+	want := pipeline.ModuleSpec{Key: Key, Stage: pipeline.StageExtract, ExecutionClass: contracts.ExecutionClassLLMBacked, Requires: []string{"chunks", "source.transcript"}, Provides: []string{"dnd.locations"}, ArtifactKind: dnd.LocationListKind, ReferenceSlots: referenceSlots()}
+	if got := ModuleSpec(); !reflect.DeepEqual(got, want) {
+		t.Fatalf("ModuleSpec() = %#v, want %#v", got, want)
+	}
+	registry := pipeline.NewExtractorRegistry()
+	if err := Register(registry); err != nil {
+		t.Fatalf("Register() error = %v", err)
+	}
+	if got, ok := registry.Spec(Key); !ok || !reflect.DeepEqual(got, want) {
+		t.Fatalf("registry spec = %#v, present = %t", got, ok)
+	}
+	if err := Register(nil); err == nil || !strings.Contains(err.Error(), "extractor registry") {
+		t.Fatalf("Register(nil) error = %v, want registry rejection", err)
+	}
+	if _, err := DecodeOptions(map[string]any{"unknown": true}); err == nil {
+		t.Fatal("DecodeOptions() accepted unknown option")
+	}
+	extractor := newExtractor(t, &fakeLocationsLLMClient{})
+	metadata := extractor.ManifestMetadata()
+	for key, value := range map[string]string{
+		"prompt_id": PromptID, "prompt_version": SchemaVersion,
+		"response_schema_key": string(ResponseSchemaKey), "response_schema_id": ResponseSchemaID,
+		"response_schema_name": ResponseSchemaName, "response_schema_version": SchemaVersion,
+		"identity_policy": identity.Policy, "mapping_policy": mappingPolicy,
+	} {
+		if metadata[key] != value {
+			t.Fatalf("metadata[%q] = %#v, want %q", key, metadata[key], value)
+		}
+	}
+	if got := extractor.CheckpointFingerprints(); len(got) != 4 || got[0].Name != "prompt" || got[1].Name != "response_schema" || got[2].Value != identity.Policy || got[3].Value != mappingPolicy {
+		t.Fatalf("CheckpointFingerprints() = %#v", got)
+	}
+}

internal/modules/dnd/extract/locations/schema.go

@@ -0,0 +1,21 @@
+package locations
+
+import "gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
+
+const (
+	PromptID           = "dnd.locations"
+	ResponseSchemaKey  = llm.ResponseSchemaKey("dnd_locations_llm")
+	ResponseSchemaID   = "notarius.dnd.locations.llm"
+	ResponseSchemaName = "notarius_dnd_locations_llm_v1"
+	SchemaVersion      = "v1"
+)
+
+func loadResponseSchema() (llm.ResponseSchema, error) {
+	return llm.LoadResponseSchema(embeddedAssets, llm.ResponseSchemaDefinition{
+		Key:       ResponseSchemaKey,
+		ID:        ResponseSchemaID,
+		Version:   SchemaVersion,
+		Name:      ResponseSchemaName,
+		AssetPath: "assets/schemas/dnd_locations_llm.v1.json",
+	})
+}

internal/modules/dnd/extract/locations/schema_test.go

@@ -0,0 +1,25 @@
+package locations
+
+import (
+	"encoding/json"
+	"strings"
+	"testing"
+)
+
+func TestLocationResponseSchemaIsPrivateAndStructural(t *testing.T) {
+	schema, err := loadResponseSchema()
+	if err != nil {
+		t.Fatalf("loadResponseSchema() error = %v", err)
+	}
+	if schema.Key != ResponseSchemaKey || schema.ID != ResponseSchemaID || schema.Name != ResponseSchemaName || schema.Version != SchemaVersion || !strings.HasPrefix(schema.SHA256, "sha256:") || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("schema = %#v, want location response schema", schema)
+	}
+	var document map[string]any
+	if err := json.Unmarshal(schema.JSONSchema, &document); err != nil {
+		t.Fatal(err)
+	}
+	encoded, err := json.Marshal(document)
+	if err != nil || strings.Contains(string(encoded), `"id"`) {
+		t.Fatalf("schema = %s, want no durable ID field", encoded)
+	}
+}

internal/modules/dnd/extract/locations/test_helpers_test.go

@@ -0,0 +1,89 @@
+package locations
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+)
+
+func extractionRequest() contracts.TypedExtractionRequest {
+	doc := sourceDocument()
+	chunk := &source.Chunk{
+		ID: "session-locations:chunk:0", SourceID: doc.ID, Index: 0,
+		Ref:     source.SourceRef{SourceID: doc.ID, StartUnitID: 1, EndUnitID: 3},
+		Content: []byte(`{"units":[1,2,3]}`), MediaType: "application/json",
+		Units: append([]source.SourceUnit(nil), doc.Units...),
+	}
+	return contracts.TypedExtractionRequest{
+		Source: doc, Chunk: chunk,
+		SourceInput: contracts.NewLLMInputMaterial("source", chunk.MediaType, chunk.Content, "sha256:chunk", "file:///session-locations.json"),
+		SessionID:   "location-session", LLMProfile: "location-profile",
+	}
+}
+
+func sourceDocument() *source.SourceDocument {
+	return &source.SourceDocument{ID: "session-locations", Kind: "transcript", Format: "application/json", Digest: "sha256:test", Units: []source.SourceUnit{
+		{ID: 1, Kind: "transcript_segment", Text: "The party enters the Old Mill."},
+		{ID: 2, Kind: "transcript_segment", Text: "They leave the old road behind."},
+		{ID: 3, Kind: "transcript_segment", Text: "The tavern is quiet."},
+	}}
+}
+
+func responseSourceRefs(startUnitID, endUnitID int) []locationSourceRefResponse {
+	return []locationSourceRefResponse{{StartUnitID: startUnitID, EndUnitID: endUnitID}}
+}
+
+func newExtractor(t *testing.T, client contracts.StructuredLLMClient, references ...contracts.ReferenceSet) *Extractor {
+	t.Helper()
+	extractor, err := New(client, Options{}, references...)
+	if err != nil {
+		t.Fatalf("New() error = %v, want nil", err)
+	}
+	return extractor
+}
+
+func mismatchedSourceInputRequest(req contracts.TypedExtractionRequest) contracts.TypedExtractionRequest {
+	req.SourceInput = contracts.NewLLMInputMaterial("source", "application/json", []byte(`{"different":true}`), "sha256:other", "file:///other.json")
+	return req
+}
+
+type fakeLocationsLLMClient struct {
+	response extractionResponse
+	content  []byte
+	err      error
+	requests []contracts.StructuredCompletionRequest
+}
+
+func (client *fakeLocationsLLMClient) CompleteStructured(_ context.Context, req contracts.StructuredCompletionRequest, out any) (contracts.StructuredCompletionResponse, error) {
+	client.requests = append(client.requests, cloneStructuredCompletionRequest(req))
+	if client.err != nil {
+		return contracts.StructuredCompletionResponse{}, client.err
+	}
+	target, ok := out.(*extractionResponse)
+	if !ok {
+		return contracts.StructuredCompletionResponse{}, errors.New("unexpected output target")
+	}
+	content := append([]byte(nil), client.content...)
+	if len(content) != 0 {
+		if err := json.Unmarshal(content, target); err != nil {
+			return contracts.StructuredCompletionResponse{}, err
+		}
+	} else {
+		*target = client.response
+		var err error
+		content, err = json.Marshal(client.response)
+		if err != nil {
+			return contracts.StructuredCompletionResponse{}, err
+		}
+	}
+	return contracts.StructuredCompletionResponse{Content: content}, nil
+}
+
+func cloneStructuredCompletionRequest(req contracts.StructuredCompletionRequest) contracts.StructuredCompletionRequest {
+	req.Inputs = req.Inputs.Clone()
+	return req
+}

internal/modules/dnd/locations/identity/identity.go

@@ -0,0 +1,183 @@
+// Package identity implements deterministic session-scoped location identity.
+package identity
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"encoding/json"
+	"sort"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	"golang.org/x/text/cases"
+	"golang.org/x/text/unicode/norm"
+)
+
+const (
+	// IdentityPolicy identifies the durable location ID derivation policy.
+	IdentityPolicy = "dnd.locations.identity.v1"
+	// Policy is an alias for IdentityPolicy.
+	Policy = IdentityPolicy
+
+	idPrefix = "location:sha256:"
+)
+
+// IssueCode identifies one deterministic registry identity problem.
+type IssueCode string
+
+const (
+	IssueEmptyCanonicalName IssueCode = "empty_canonical_name"
+	IssueMissingEvidence    IssueCode = "missing_evidence"
+	IssueInvalidID          IssueCode = "invalid_id"
+	IssueIDMismatch         IssueCode = "id_mismatch"
+	IssueDuplicateID        IssueCode = "duplicate_id"
+)
+
+// Issue is an inspectable identity validation problem.
+type Issue struct {
+	Code        IssueCode
+	RecordIndex int
+	Value       string
+}
+
+func (i Issue) Error() string { return string(i.Code) }
+
+// NormalizeDisplay returns the durable display form without changing spelling
+// or punctuation.
+func NormalizeDisplay(value string) string {
+	return strings.Join(strings.Fields(value), " ")
+}
+
+// ComparisonKey returns the stable key used for location identity comparisons.
+func ComparisonKey(value string) string {
+	value = norm.NFKC.String(value)
+	value = strings.Map(func(r rune) rune {
+		switch r {
+		case '\u2018', '\u2019', '\u02bc':
+			return '\''
+		default:
+			return r
+		}
+	}, value)
+	value = strings.Join(strings.Fields(value), " ")
+	return cases.Fold().String(value)
+}
+
+// DeriveID derives a location ID from name and the earliest canonical evidence
+// reference. It returns an empty string when either identity component is not
+// available, leaving validation to report the problem instead of manufacturing
+// an ID.
+func DeriveID(name string, refs []source.SourceRef) string {
+	comparisonName := ComparisonKey(name)
+	anchor, ok := earliestReference(refs)
+	if comparisonName == "" || !ok {
+		return ""
+	}
+
+	input, err := json.Marshal([]any{
+		IdentityPolicy,
+		comparisonName,
+		anchor.SourceID,
+		anchor.StartUnitID,
+		anchor.EndUnitID,
+	})
+	if err != nil {
+		return ""
+	}
+	digest := sha256.Sum256(input)
+	return idPrefix + hex.EncodeToString(digest[:])
+}
+
+// IDFor is an alias for DeriveID.
+func IDFor(name string, refs []source.SourceRef) string { return DeriveID(name, refs) }
+
+// IsValidID reports whether value has the exact durable location ID syntax.
+func IsValidID(value string) bool {
+	if len(value) != len(idPrefix)+sha256.Size*2 || !strings.HasPrefix(value, idPrefix) {
+		return false
+	}
+	for _, r := range value[len(idPrefix):] {
+		if !((r >= '0' && r <= '9') || (r >= 'a' && r <= 'f')) {
+			return false
+		}
+	}
+	return true
+}
+
+// ValidID is an alias for IsValidID.
+func ValidID(value string) bool { return IsValidID(value) }
+
+// ValidateList validates the identity members of list.
+func ValidateList(list dnd.LocationList) []Issue { return ValidateRegistry(list.Locations) }
+
+// ValidateRegistry validates location IDs without modifying records or their
+// source references. Equal comparison names are allowed because their evidence
+// anchors are part of the identity policy.
+func ValidateRegistry(locations []dnd.Location) []Issue {
+	issues := make([]Issue, 0)
+	idOwners := make(map[string][]int)
+
+	for recordIndex, location := range locations {
+		if ComparisonKey(location.Name) == "" {
+			issues = append(issues, Issue{Code: IssueEmptyCanonicalName, RecordIndex: recordIndex, Value: location.Name})
+		}
+		if _, ok := earliestReference(location.SourceRefs); !ok {
+			issues = append(issues, Issue{Code: IssueMissingEvidence, RecordIndex: recordIndex})
+		}
+
+		if !IsValidID(location.ID) {
+			issues = append(issues, Issue{Code: IssueInvalidID, RecordIndex: recordIndex, Value: location.ID})
+		} else if expected := DeriveID(location.Name, location.SourceRefs); location.ID != expected {
+			issues = append(issues, Issue{Code: IssueIDMismatch, RecordIndex: recordIndex, Value: location.ID})
+		}
+		if location.ID != "" {
+			idOwners[location.ID] = append(idOwners[location.ID], recordIndex)
+		}
+	}
+
+	for recordIndex, location := range locations {
+		if location.ID != "" && len(idOwners[location.ID]) > 1 && idOwners[location.ID][0] != recordIndex {
+			issues = append(issues, Issue{Code: IssueDuplicateID, RecordIndex: recordIndex, Value: location.ID})
+		}
+	}
+
+	return issues
+}
+
+func earliestReference(refs []source.SourceRef) (source.SourceRef, bool) {
+	canonical := canonicalReferences(refs)
+	if len(canonical) == 0 {
+		return source.SourceRef{}, false
+	}
+	return canonical[0], true
+}
+
+func canonicalReferences(refs []source.SourceRef) []source.SourceRef {
+	canonical := make([]source.SourceRef, 0, len(refs))
+	for _, ref := range refs {
+		if validIdentityReference(ref) {
+			canonical = append(canonical, ref)
+		}
+	}
+	sort.Slice(canonical, func(left, right int) bool {
+		if canonical[left].SourceID != canonical[right].SourceID {
+			return canonical[left].SourceID < canonical[right].SourceID
+		}
+		if canonical[left].StartUnitID != canonical[right].StartUnitID {
+			return canonical[left].StartUnitID < canonical[right].StartUnitID
+		}
+		return canonical[left].EndUnitID < canonical[right].EndUnitID
+	})
+	unique := canonical[:0]
+	for _, ref := range canonical {
+		if len(unique) == 0 || unique[len(unique)-1] != ref {
+			unique = append(unique, ref)
+		}
+	}
+	return unique
+}
+
+func validIdentityReference(ref source.SourceRef) bool {
+	return strings.TrimSpace(ref.SourceID) == ref.SourceID && ref.SourceID != "" && ref.StartUnitID > 0 && ref.EndUnitID > 0
+}

internal/modules/dnd/locations/identity/identity_test.go

@@ -0,0 +1,146 @@
+package identity
+
+import (
+	"reflect"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+)
+
+func TestNormalizeDisplayOnlyChangesWhitespace(t *testing.T) {
+	if got, want := NormalizeDisplay("  The\u2003Old\nTavern  "), "The Old Tavern"; got != want {
+		t.Fatalf("NormalizeDisplay() = %q, want %q", got, want)
+	}
+}
+
+func TestComparisonKeyNormalizesUnicodeWhitespaceAndApostrophes(t *testing.T) {
+	for _, pair := range [][2]string{
+		{"  Caf\u00e9\u2003d\u2019Or  ", "cafe\u0301 d'Or"},
+		{"\uff34\uff48\uff45\u00a0\uff34\uff41\uff56\uff45\uff52\uff4e", "the tavern"},
+	} {
+		if left, right := ComparisonKey(pair[0]), ComparisonKey(pair[1]); left != right {
+			t.Fatalf("ComparisonKey(%q) = %q, want value equal to %q", pair[0], left, pair[1])
+		}
+	}
+}
+
+func TestDeriveIDUsesDocumentedCompactJSONInput(t *testing.T) {
+	refs := []source.SourceRef{
+		{SourceID: "session-alpha", StartUnitID: 9, EndUnitID: 9},
+		{SourceID: "session-alpha", StartUnitID: 3, EndUnitID: 9},
+		{SourceID: "session-alpha", StartUnitID: 3, EndUnitID: 9},
+	}
+	const want = "location:sha256:379bd996e754e143b47e6b613a95166e58c111c7451750df7066ca816bb1866c"
+
+	if got := DeriveID("  The\u2003Tavern ", refs); got != want {
+		t.Fatalf("DeriveID() = %q, want %q", got, want)
+	}
+	if got := IDFor("The Tavern", refs); got != want {
+		t.Fatalf("IDFor() = %q, want %q", got, want)
+	}
+}
+
+func TestDeriveIDUsesEarliestCanonicalEvidenceWithoutMutatingInput(t *testing.T) {
+	refs := []source.SourceRef{
+		{SourceID: "zeta", StartUnitID: 1, EndUnitID: 1},
+		{SourceID: "alpha", StartUnitID: 8, EndUnitID: 8},
+		{SourceID: "alpha", StartUnitID: 2, EndUnitID: 3},
+	}
+	wantRefs := append([]source.SourceRef(nil), refs...)
+
+	first := DeriveID("The Tavern", refs)
+	second := DeriveID("The Tavern", []source.SourceRef{refs[2], refs[0], refs[1]})
+	if first != second {
+		t.Fatalf("DeriveID() changed when evidence order changed: %q != %q", first, second)
+	}
+	if !reflect.DeepEqual(refs, wantRefs) {
+		t.Fatalf("DeriveID() mutated refs: got %#v, want %#v", refs, wantRefs)
+	}
+}
+
+func TestDeriveIDDistinguishesSameNameAtDifferentEvidenceAnchors(t *testing.T) {
+	first := DeriveID("The Tavern", []source.SourceRef{{SourceID: "session", StartUnitID: 3, EndUnitID: 3}})
+	second := DeriveID("the\u00a0tavern", []source.SourceRef{{SourceID: "session", StartUnitID: 18, EndUnitID: 18}})
+	if first == "" || second == "" || first == second {
+		t.Fatalf("same-name locations have IDs %q and %q, want distinct nonempty IDs", first, second)
+	}
+}
+
+func TestDeriveIDRejectsMissingIdentityComponents(t *testing.T) {
+	for _, tt := range []struct {
+		name      string
+		nameValue string
+		refs      []source.SourceRef
+	}{
+		{name: "blank name", nameValue: " \u2003 ", refs: []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 1}}},
+		{name: "missing refs", nameValue: "The Tavern"},
+		{name: "malformed ref", nameValue: "The Tavern", refs: []source.SourceRef{{SourceID: " ", StartUnitID: 1, EndUnitID: 1}}},
+	} {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := DeriveID(tt.nameValue, tt.refs); got != "" {
+				t.Fatalf("DeriveID() = %q, want empty ID", got)
+			}
+		})
+	}
+}
+
+func TestIsValidID(t *testing.T) {
+	valid := DeriveID("The Tavern", []source.SourceRef{{SourceID: "session", StartUnitID: 1, EndUnitID: 1}})
+	if !IsValidID(valid) || !ValidID(valid) {
+		t.Fatalf("derived ID %q is not valid", valid)
+	}
+	for _, value := range []string{"", "location:sha256:", "location:sha256:" + strings.Repeat("A", 64), "location:sha256:" + strings.Repeat("0", 63)} {
+		if IsValidID(value) {
+			t.Fatalf("IsValidID(%q) = true, want false", value)
+		}
+	}
+}
+
+func TestValidateRegistryAllowsSameComparisonNameWithDifferentAnchors(t *testing.T) {
+	firstRefs := []source.SourceRef{{SourceID: "session", StartUnitID: 2, EndUnitID: 2}}
+	secondRefs := []source.SourceRef{{SourceID: "session", StartUnitID: 8, EndUnitID: 8}}
+	locations := []dnd.Location{
+		{ID: DeriveID("The Tavern", firstRefs), Name: "The Tavern", SourceRefs: firstRefs},
+		{ID: DeriveID("the\u00a0tavern", secondRefs), Name: "the\u00a0tavern", SourceRefs: secondRefs},
+	}
+	if issues := ValidateRegistry(locations); len(issues) != 0 {
+		t.Fatalf("ValidateRegistry() = %#v, want no issues", issues)
+	}
+}
+
+func TestValidateRegistryReportsIdentityProblemsDeterministicallyWithoutMutation(t *testing.T) {
+	refs := []source.SourceRef{{SourceID: "session", StartUnitID: 2, EndUnitID: 2}}
+	validID := DeriveID("The Tavern", refs)
+	locations := []dnd.Location{
+		{ID: validID, Name: "The Tavern", SourceRefs: refs},
+		{ID: validID, Name: "Elsewhere", SourceRefs: refs},
+		{ID: "bad", Name: "", SourceRefs: nil},
+	}
+	wantLocations := cloneLocations(locations)
+
+	issues := ValidateList(dnd.LocationList{Locations: locations})
+	want := []Issue{
+		{Code: IssueIDMismatch, RecordIndex: 1, Value: validID},
+		{Code: IssueEmptyCanonicalName, RecordIndex: 2, Value: ""},
+		{Code: IssueMissingEvidence, RecordIndex: 2},
+		{Code: IssueInvalidID, RecordIndex: 2, Value: "bad"},
+		{Code: IssueDuplicateID, RecordIndex: 1, Value: validID},
+	}
+	if !reflect.DeepEqual(issues, want) {
+		t.Fatalf("ValidateList() = %#v, want %#v", issues, want)
+	}
+	if !reflect.DeepEqual(locations, wantLocations) {
+		t.Fatalf("ValidateList() mutated locations: got %#v, want %#v", locations, wantLocations)
+	}
+}
+
+func cloneLocations(input []dnd.Location) []dnd.Location {
+	output := make([]dnd.Location, len(input))
+	copy(output, input)
+	for index := range output {
+		output[index].SourceRefs = append([]source.SourceRef(nil), input[index].SourceRefs...)
+	}
+	return output
+}

internal/modules/dnd/locations/registry/registry.go

@@ -0,0 +1,277 @@
+// Package registry resolves normalized location artifacts into immutable
+// ID-grounding data for D&D extraction modules.
+package registry
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"encoding/json"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	locationcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/locations"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared/diagnostics"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared/registryresolver"
+)
+
+const (
+	ReferenceSlot = "locations"
+	MaxBytes      = 1048576
+	emptyPrompt   = `{"locations":[]}`
+)
+
+// Registry is an immutable, validated location registry prepared for prompt
+// grounding. All accessors return defensive copies.
+type Registry struct {
+	bound            bool
+	list             dnd.LocationList
+	canonical        []byte
+	digest           string
+	projectionDigest string
+	promptInput      contracts.LLMInputMaterial
+	lookupByID       map[string]int
+}
+
+// Resolver selects and memoizes immutable location registry views.
+type Resolver struct {
+	resolver *registryresolver.Resolver[*Registry]
+}
+
+// NewResolver validates a materialized construction-time location reference.
+// An empty slot is permitted because a generated reference is supplied only at
+// operation time.
+func NewResolver(references contracts.ReferenceSet) (*Resolver, error) {
+	resolver, err := registryresolver.New(registryResolverConfig(), references)
+	if err != nil {
+		return nil, err
+	}
+	return &Resolver{resolver: resolver}, nil
+}
+
+// Seeded returns the validated construction-time registry.
+func (r *Resolver) Seeded() *Registry {
+	if r == nil {
+		return nil
+	}
+	return r.resolver.Seeded()
+}
+
+// Resolve returns the generated operation-time registry when the locations
+// slot is present, otherwise it returns the construction-time registry.
+func (r *Resolver) Resolve(references contracts.ReferenceSet) (*Registry, error) {
+	if r == nil || r.resolver == nil {
+		return Resolve(references)
+	}
+	return r.resolver.Resolve(references)
+}
+
+// Resolve validates an optional location-list reference. An absent reference
+// uses the canonical empty projection and has no durable registry identity.
+func Resolve(references contracts.ReferenceSet) (*Registry, error) {
+	item, present, err := registryresolver.ResolveOptionalSingleItem(references, locationReferenceSpec())
+	if err != nil {
+		return nil, err
+	}
+	if !present {
+		return emptyRegistry(), nil
+	}
+	return loadRegistry(item.Content)
+}
+
+func registryResolverConfig() registryresolver.Config[*Registry] {
+	return registryresolver.Config[*Registry]{
+		Reference: locationReferenceSpec(),
+		Absent: func() (*Registry, error) {
+			return emptyRegistry(), nil
+		},
+		Load: loadRegistry,
+		SemanticIdentity: func(registry *Registry) string {
+			return registry.Digest()
+		},
+	}
+}
+
+func locationReferenceSpec() registryresolver.ReferenceSpec {
+	return registryresolver.ReferenceSpec{SlotName: ReferenceSlot, AcceptedMediaType: locationcodec.MediaType, MaxBytes: MaxBytes}
+}
+
+func emptyRegistry() *Registry {
+	content := []byte(emptyPrompt)
+	projectionDigest := semanticDigest(content)
+	return &Registry{
+		list:             dnd.LocationList{Locations: []dnd.Location{}},
+		canonical:        append([]byte(nil), content...),
+		projectionDigest: projectionDigest,
+		promptInput:      contracts.NewLLMInputMaterial(ReferenceSlot, locationcodec.MediaType, content, projectionDigest, ""),
+		lookupByID:       map[string]int{},
+	}
+}
+
+func loadRegistry(referenceContent []byte) (*Registry, error) {
+	codec := locationcodec.New()
+	value, err := codec.Decode(referenceContent)
+	if err != nil {
+		return nil, fmt.Errorf("decode location registry: invalid approved location JSON")
+	}
+	if issues := identity.ValidateList(value); len(issues) > 0 {
+		return nil, fmt.Errorf("%s", formatIdentityIssues(issues))
+	}
+	content, err := codec.Encode(value)
+	if err != nil {
+		return nil, fmt.Errorf("encode canonical location registry: approved location value could not be encoded")
+	}
+
+	list := cloneLocationList(value)
+	lookupByID := make(map[string]int, len(list.Locations))
+	for index, location := range list.Locations {
+		lookupByID[location.ID] = index
+	}
+	projection, err := promptProjection(list)
+	if err != nil {
+		return nil, fmt.Errorf("encode location prompt projection: %w", err)
+	}
+	digest := semanticDigest(content)
+	projectionDigest := semanticDigest(projection)
+	return &Registry{
+		bound:            true,
+		list:             list,
+		canonical:        append([]byte(nil), content...),
+		digest:           digest,
+		projectionDigest: projectionDigest,
+		promptInput:      contracts.NewLLMInputMaterial(ReferenceSlot, locationcodec.MediaType, projection, projectionDigest, ""),
+		lookupByID:       lookupByID,
+	}, nil
+}
+
+// Bound reports whether a location reference was supplied and validated.
+func (r *Registry) Bound() bool { return r != nil && r.bound }
+
+// Locations returns a defensive copy of the validated location records.
+func (r *Registry) Locations() []dnd.Location {
+	if r == nil {
+		return nil
+	}
+	return cloneLocations(r.list.Locations)
+}
+
+// List returns a defensive copy of the validated location list.
+func (r *Registry) List() dnd.LocationList {
+	if r == nil {
+		return dnd.LocationList{}
+	}
+	return cloneLocationList(r.list)
+}
+
+// CanonicalBytes returns a defensive copy of the canonical durable JSON.
+func (r *Registry) CanonicalBytes() []byte {
+	if r == nil {
+		return nil
+	}
+	return append([]byte(nil), r.canonical...)
+}
+
+// Digest returns the semantic digest of the canonical JSON, or an empty string
+// when the registry is unbound.
+func (r *Registry) Digest() string {
+	if r == nil {
+		return ""
+	}
+	return r.digest
+}
+
+// ProjectionDigest returns the digest of the exact source-free prompt
+// projection, including for an unbound or empty registry.
+func (r *Registry) ProjectionDigest() string {
+	if r == nil {
+		return ""
+	}
+	return r.projectionDigest
+}
+
+// Count returns the number of validated location records.
+func (r *Registry) Count() int {
+	if r == nil {
+		return 0
+	}
+	return len(r.list.Locations)
+}
+
+// PromptInput returns the ordered ID-and-name projection without evidence or
+// reference provenance.
+func (r *Registry) PromptInput() contracts.LLMInputMaterial {
+	if r == nil {
+		return contracts.LLMInputMaterial{}
+	}
+	return r.promptInput.Clone()
+}
+
+// Lookup returns the canonical location for an exact durable location ID.
+func (r *Registry) Lookup(id string) (dnd.Location, bool) {
+	if r == nil {
+		return dnd.Location{}, false
+	}
+	index, ok := r.lookupByID[id]
+	if !ok {
+		return dnd.Location{}, false
+	}
+	return cloneLocation(r.list.Locations[index]), true
+}
+
+// Matches reports whether id resolves to exactly the supplied canonical name.
+func (r *Registry) Matches(id, name string) bool {
+	location, ok := r.Lookup(id)
+	return ok && location.Name == name
+}
+
+func semanticDigest(content []byte) string {
+	sum := sha256.Sum256(content)
+	return "sha256:" + hex.EncodeToString(sum[:])
+}
+
+type projectedLocation struct {
+	ID   string `json:"id"`
+	Name string `json:"name"`
+}
+
+type projectedLocationList struct {
+	Locations []projectedLocation `json:"locations"`
+}
+
+func promptProjection(list dnd.LocationList) ([]byte, error) {
+	projection := projectedLocationList{Locations: make([]projectedLocation, len(list.Locations))}
+	for index, location := range list.Locations {
+		projection.Locations[index] = projectedLocation{ID: location.ID, Name: location.Name}
+	}
+	return json.Marshal(projection)
+}
+
+func formatIdentityIssues(issues []identity.Issue) string {
+	parts := make([]string, len(issues))
+	for index, issue := range issues {
+		parts[index] = fmt.Sprintf("%s at record %d", issue.Code, issue.RecordIndex)
+	}
+	return diagnostics.Aggregate("validate location registry identity", parts)
+}
+
+func cloneLocationList(value dnd.LocationList) dnd.LocationList {
+	return dnd.LocationList{Locations: cloneLocations(value.Locations)}
+}
+
+func cloneLocations(values []dnd.Location) []dnd.Location {
+	if values == nil {
+		return nil
+	}
+	cloned := make([]dnd.Location, len(values))
+	for index, value := range values {
+		cloned[index] = cloneLocation(value)
+	}
+	return cloned
+}
+
+func cloneLocation(value dnd.Location) dnd.Location {
+	value.SourceRefs = append([]source.SourceRef(nil), value.SourceRefs...)
+	return value
+}

internal/modules/dnd/locations/registry/registry_test.go

@@ -0,0 +1,191 @@
+package registry
+
+import (
+	"bytes"
+	"strings"
+	"testing"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+	locationcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/locations"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/locations/identity"
+)
+
+func TestResolveUnboundRegistryHasEmptyProjection(t *testing.T) {
+	registry, err := Resolve(contracts.ReferenceSet{})
+	if err != nil {
+		t.Fatal(err)
+	}
+	input := registry.PromptInput()
+	if registry.Bound() || registry.Digest() != "" || registry.Count() != 0 || string(input.Content) != emptyPrompt {
+		t.Fatalf("unbound registry = %#v, input = %#v", registry, input)
+	}
+	if registry.ProjectionDigest() == "" || input.Digest != registry.ProjectionDigest() || input.OriginURI != "" {
+		t.Fatalf("projection digest/input = %q/%#v", registry.ProjectionDigest(), input)
+	}
+}
+
+func TestResolveProjectsOrderedLocationsWithoutEvidence(t *testing.T) {
+	registry := resolveList(t, registryFixture())
+	if !registry.Bound() || registry.Count() != 2 || registry.Digest() == "" {
+		t.Fatalf("registry identity = bound %t count %d digest %q", registry.Bound(), registry.Count(), registry.Digest())
+	}
+	projection := string(registry.PromptInput().Content)
+	if !strings.Contains(projection, `"locations":[{"id":`) || !strings.Contains(projection, `"name":"The Tavern"`) || !strings.Contains(projection, `"name":"The Tavern"},{"id":`) {
+		t.Fatalf("projection ordering = %s", projection)
+	}
+	for _, forbidden := range []string{"source_refs", "source_id", "session-alpha"} {
+		if strings.Contains(projection, forbidden) {
+			t.Fatalf("projection leaked %q: %s", forbidden, projection)
+		}
+	}
+	if registry.PromptInput().Digest != registry.ProjectionDigest() || registry.Digest() == registry.ProjectionDigest() {
+		t.Fatalf("full/projection digests = %q/%q", registry.Digest(), registry.ProjectionDigest())
+	}
+}
+
+func TestRegistryLookupUsesIDAndReturnsDefensiveCopies(t *testing.T) {
+	registry := resolveList(t, registryFixture())
+	first := registry.Locations()[0]
+	if got, ok := registry.Lookup(first.ID); !ok || got.Name != first.Name || !registry.Matches(first.ID, first.Name) || registry.Matches(first.ID, "Other Tavern") {
+		t.Fatalf("ID lookup/match = %#v, %t", got, ok)
+	}
+	if _, ok := registry.Lookup("The Tavern"); ok {
+		t.Fatal("Lookup accepted a name as an ID")
+	}
+
+	locations := registry.Locations()
+	locations[0].Name = "changed"
+	locations[0].SourceRefs[0].SourceID = "changed"
+	canonical := registry.CanonicalBytes()
+	canonical[0] = '['
+	input := registry.PromptInput()
+	input.Content[0] = '['
+	if next, ok := registry.Lookup(first.ID); !ok || next.Name != first.Name || next.SourceRefs[0].SourceID != "session-alpha" {
+		t.Fatalf("registry mutated through accessor: %#v, %t", next, ok)
+	}
+	if registry.CanonicalBytes()[0] != '{' || registry.PromptInput().Content[0] != '{' {
+		t.Fatal("registry bytes mutated through accessor")
+	}
+}
+
+func TestResolveRejectsInvalidReferenceInputs(t *testing.T) {
+	valid := registryFixture()
+	content, err := locationcodec.New().Encode(valid)
+	if err != nil {
+		t.Fatal(err)
+	}
+	invalidIdentity := append([]byte(nil), content...)
+	invalidIdentity = bytes.Replace(invalidIdentity, []byte(valid.Locations[0].ID), []byte("location:sha256:0000000000000000000000000000000000000000000000000000000000000000"), 1)
+	for _, test := range []struct {
+		name string
+		set  contracts.ReferenceSet
+		want string
+	}{
+		{"empty bound slot", referenceSet(), "exactly one item"},
+		{"multiple items", referenceSet(item(content), item(content)), "exactly one item"},
+		{"malformed JSON", referenceSet(item([]byte(`{"locations":[`))), "invalid approved"},
+		{"wrong media type", referenceSet(contracts.ReferenceItem{MediaType: "text/plain", Content: content}), "media type must be"},
+		{"oversized", referenceSet(item(make([]byte, MaxBytes+1))), "limit"},
+		{"invalid identity", referenceSet(item(invalidIdentity)), "id_mismatch"},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			if _, err := Resolve(test.set); err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Resolve() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestResolverHandlesConstructionAndOperationReferences(t *testing.T) {
+	placeholder, err := NewResolver(referenceSet())
+	if err != nil || placeholder.Seeded().Bound() {
+		t.Fatalf("generated placeholder = %#v, %v; want unbound seed", placeholder, err)
+	}
+
+	validContent := encodeList(t, registryFixture())
+	malformed := referenceSet(item([]byte(`{"locations":[`)))
+	if _, err := NewResolver(malformed); err == nil {
+		t.Fatal("NewResolver(malformed) error = nil")
+	}
+	if _, err := placeholder.Resolve(malformed); err == nil {
+		t.Fatal("Resolve(malformed) error = nil")
+	}
+
+	staticContent := append([]byte(nil), validContent...)
+	staticReferences := referenceSet(item(staticContent))
+	seeded, err := NewResolver(staticReferences)
+	if err != nil {
+		t.Fatal(err)
+	}
+	resolved, err := seeded.Resolve(staticReferences)
+	if err != nil || resolved != seeded.Seeded() {
+		t.Fatalf("seed reuse = %p / %p, %v", resolved, seeded.Seeded(), err)
+	}
+	staticContent[0] = '['
+	delete(staticReferences.Slots, ReferenceSlot)
+	if seeded.Seeded().Count() != 2 || seeded.Seeded().CanonicalBytes()[0] != '{' {
+		t.Fatalf("seeded registry retained construction references: %#v", seeded.Seeded())
+	}
+	if fallback, err := seeded.Resolve(contracts.ReferenceSet{}); err != nil || fallback != seeded.Seeded() {
+		t.Fatalf("fallback = %#v, %v; want seeded registry", fallback, err)
+	}
+}
+
+func TestResolverReusesEquivalentCanonicalRegistries(t *testing.T) {
+	resolver, err := NewResolver(referenceSet())
+	if err != nil {
+		t.Fatal(err)
+	}
+	content, err := locationcodec.New().Encode(registryFixture())
+	if err != nil {
+		t.Fatal(err)
+	}
+	firstSet := referenceSet(item(content))
+	first, err := resolver.Resolve(firstSet)
+	if err != nil {
+		t.Fatal(err)
+	}
+	spaced := append([]byte("\n  "), content...)
+	spaced = append(spaced, '\n')
+	second, err := resolver.Resolve(referenceSet(contracts.ReferenceItem{MediaType: "APPLICATION/JSON; charset=utf-8", Content: spaced}))
+	if err != nil || second != first {
+		t.Fatalf("equivalent canonical registry = %p / %p, %v", first, second, err)
+	}
+}
+
+func registryFixture() dnd.LocationList {
+	firstRefs := []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 1, EndUnitID: 1}}
+	secondRefs := []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 2, EndUnitID: 2}}
+	return dnd.LocationList{Locations: []dnd.Location{
+		{ID: identity.DeriveID("The Tavern", firstRefs), Name: "The Tavern", SourceRefs: firstRefs},
+		{ID: identity.DeriveID("The Tavern", secondRefs), Name: "The Tavern", SourceRefs: secondRefs},
+	}}
+}
+
+func resolveList(t *testing.T, list dnd.LocationList) *Registry {
+	t.Helper()
+	registry, err := Resolve(referenceSet(item(encodeList(t, list))))
+	if err != nil {
+		t.Fatalf("Resolve() error = %v", err)
+	}
+	return registry
+}
+
+func encodeList(t *testing.T, list dnd.LocationList) []byte {
+	t.Helper()
+	content, err := locationcodec.New().Encode(list)
+	if err != nil {
+		t.Fatalf("Encode() error = %v", err)
+	}
+	return content
+}
+
+func item(content []byte) contracts.ReferenceItem {
+	return contracts.ReferenceItem{SlotName: ReferenceSlot, MediaType: locationcodec.MediaType, Content: content}
+}
+
+func referenceSet(items ...contracts.ReferenceItem) contracts.ReferenceSet {
+	return contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{ReferenceSlot: {Items: items}}}
+}