Introduce typed D&D spell artifacts

docs/integrations/dnd-spell-artifacts.md

@@ -6,6 +6,7 @@ extractor. Selectable extractor keys are cataloged in
 
 ## Identity
 
+- Artifact kind: `dnd/spell-list`
 - Prompt ID: `dnd.spells`
 - Response schema key: `dnd_spells`
 - Response schema ID: `notarius.dnd.spells`
@@ -13,6 +14,12 @@ extractor. Selectable extractor keys are cataloged in
 - Response schema version: `v1`
 - Media type: `application/json`
 
+The durable JSON Schema is owned by the D&D spell artifact codec. The
+extractor's private LLM response schema is a separate transport contract: its
+source-reference objects omit `source_id`, which the extractor assigns while
+mapping the response to the canonical artifact. The LLM DTO and transport
+schema are not part of this durable contract.
+
 The output contains canonical spell casts derived from transcript evidence.
 Source IDs are assigned from the input identity; source-unit ranges identify
 the evidence location.

docs/internal/modules.md

@@ -17,15 +17,14 @@ validator registry. Package-family registrars compose those leaf registrations
 into the production catalog and own family-level policy such as default
 validator chains and prompt asset collection.
 
-Production input, chunk, and output packages register strict option decoders and
-run-local builders. Preparation decodes their options into implementation-owned
-values and injects dependencies; their operation requests contain run context,
-not raw option maps or LLM clients. Production extract, merge, normalize, and
-validator packages still use the explicit legacy raw registration APIs and
-temporary adapters around zero-argument constructors. Their raw option maps and
-LLM clients remain operation inputs while that part of the catalog migrates.
-Typed registration is framework-ready, but no production artifact kind or
-codec is registered yet.
+Production input, chunk, output, and D&D spell-extract packages register strict
+option decoders and run-local builders. Preparation decodes their options into
+implementation-owned values and injects dependencies. The spell extractor is
+typed over the canonical D&D model; a temporary raw adapter preserves the
+current downstream production path. Production merge, normalize, and validator
+packages still use the explicit legacy raw registration APIs and temporary
+adapters around zero-argument constructors. Their raw option maps and LLM
+clients remain operation inputs while that part of the catalog migrates.
 
 Specs expose capability and execution metadata without constructing an
 implementation. Registry entries separately expose option validation and
@@ -106,12 +105,16 @@ The spell extractor prepares a structured request from one chunk, the
 chunk-scoped source input, the session, and optional D&D reference inputs. It
 decodes the model response, assigns the generic source identity to every source
 reference, canonicalizes duplicate references, orders spell casts by their
-earliest cited unit, and returns raw JSON plus response-schema provenance.
+earliest cited unit, and returns `dnd.SpellList`.
 
-The package owns its embedded prompt, response schemas, and prompt/schema
-manifest metadata. Shared D&D helpers keep prompt input names and source-unit
-reference conversion consistent with the scene chunker. The extractor produces
-raw output; production validators own approval policy.
+The extractor owns its private model-response DTO, embedded prompt, LLM response
+schema, strict option decoder, injected shared LLM client, and prompt/schema
+manifest metadata. The separate `internal/modules/dnd/codec/spells` package
+owns the durable schema and stable JSON representation for artifact kind
+`dnd/spell-list`. Production composition currently wraps the typed extractor
+with a raw adapter that encodes through this codec, so existing validators and
+later stages remain unchanged. Shared D&D helpers keep prompt input names and
+source-unit reference conversion consistent with the scene chunker.
 
 The durable payload and manifest metadata shapes are defined in the
 [D&D spell artifact contract](../integrations/dnd-spell-artifacts.md).

docs/internal/overview.md

@@ -62,10 +62,12 @@ run-local construction closures. Preparation injects shared dependencies and
 constructs input, chunk, validators, ordered lanes, and output before source
 parsing. Production input, chunk, and output modules use strict construction-time
 option decoding, and the LLM-backed scene chunker retains the injected shared
-client. Production artifact-lane modules and validators do not register typed
-variants yet and continue through explicitly named legacy raw registrations and
-temporary zero-argument constructor adapters. The current runner rejects a
-typed prepared lane instead of routing it through raw execution.
+client. The D&D family registers the canonical `dnd/spell-list` codec and a
+typed spell extractor. A temporary raw adapter serializes that typed result for
+the still-raw production validators, merger, normalizer, and runner. Other
+artifact-lane modules and validators continue through explicitly named legacy
+raw registrations and temporary zero-argument constructor adapters. The current
+runner rejects a typed prepared lane instead of routing it through raw execution.
 
 ## Production Extensions
 
@@ -79,7 +81,9 @@ Configuration. The implemented module packages are:
 | `internal/modules/seriatim/input/transcript` | Parses the supported Seriatim transcript format into the generic source model. |
 | `internal/modules/generic/chunk/units` | Splits ordered source units by unit count and overlap. |
 | `internal/modules/dnd/chunk/scenes` | Produces contiguous D&D scene chunks from structured model output. |
-| `internal/modules/dnd/extract/spells` | Produces source-grounded D&D spell-cast raw output. |
+| `internal/modules/dnd` | Owns the canonical D&D spell-list and spell-cast artifact types. |
+| `internal/modules/dnd/codec/spells` | Strictly decodes and stably encodes the durable D&D spell-list representation. |
+| `internal/modules/dnd/extract/spells` | Maps private structured model output to canonical source-grounded D&D spell lists. |
 | `internal/modules/generic/merge/appendorder` | Combines accepted extraction results in chunk order. |
 | `internal/modules/generic/normalize/noop` | Preserves accepted merged output. |
 | `internal/modules/generic/output/json` | Encodes manifests, lane payloads, warnings, and rejections as logical JSON files. |

docs/internal/pipeline.md

@@ -73,9 +73,11 @@ separate target namespaces; serialized registrations declare whether they
 support chunks, artifacts, or both. Duplicate variants and exact Go-type
 mismatches are rejected deterministically.
 
-Production composition initializes the artifact codec registry without adding
-codec entries, and production artifact-lane modules use the explicitly named
-legacy raw registration APIs. A raw registration cannot satisfy a typed lane.
+Production composition registers the D&D spell-list codec. The typed spell
+extractor also registers a temporary raw adapter, which resolution selects
+until its downstream production lane is typed. Other production artifact-lane
+modules use the explicitly named legacy raw registration APIs. A standalone raw
+registration cannot satisfy a typed lane.
 
 A `ModuleSpec` declares its stage plus required and provided capabilities.
 Chunk, extract, merge, and normalize specs may also declare reference slots.
@@ -84,8 +86,10 @@ and verify that a constructed implementation reports the registered key.
 Builder registrations accept `ModuleDependencies` and cloned raw options through
 one `BuildRequest`. Production input, chunk, and output builders decode those
 options and retain typed values or injected dependencies in the constructed
-implementation. Remaining production raw-stage registrations are adapted from
-their zero-argument constructors during migration.
+implementation. A typed extractor registration may explicitly supply a raw
+adapter builder for a still-raw downstream lane; the adapter is selected as one
+unit and does not expose the typed value to raw consumers. Remaining production
+raw-stage registrations are adapted from their zero-argument constructors.
 
 A `ValidatorSpec` declares a validator key and execution class. Resolution uses
 the execution class to reject incompatible profile bindings before execution.

internal/cli/compatibility_test.go

@@ -106,13 +106,12 @@ func TestProductionCompatibilitySnapshot(t *testing.T) {
 	})
 	assertAssetNames(t, assets.SchemaFS, []string{
 		"dnd_scenes.v1.json",
-		"dnd_spells.v1.json",
 		"dnd_spells_llm.v1.json",
 	})
 
 	identitySnapshot := map[string]map[string]any{
 		"scenes": sceneManifestMetadata(t),
-		"spells": spells.New().ManifestMetadata(),
+		"spells": spellManifestMetadata(t),
 	}
 	for name, metadata := range identitySnapshot {
 		for _, key := range []string{"prompt_id", "prompt_version", "prompt_sha256", "response_schema_key", "response_schema_id", "response_schema_name", "response_schema_version", "response_schema_sha256"} {
@@ -348,7 +347,10 @@ func TestProductionLLMCallersShareScheduledClient(t *testing.T) {
 	}()
 	go func() {
 		started.Done()
-		_, err := spells.New().Extract(context.Background(), contracts.ExtractionRequest{Source: doc, Chunk: &chunk, LLMClient: client})
+		extractor, err := spells.New(client, spells.Options{})
+		if err == nil {
+			_, err = extractor.Extract(context.Background(), contracts.TypedExtractionRequest{Source: doc, Chunk: &chunk})
+		}
 		errs <- err
 	}()
 	started.Wait()
@@ -520,6 +522,15 @@ func sceneManifestMetadata(t *testing.T) map[string]any {
 	return chunker.ManifestMetadata()
 }
 
+func spellManifestMetadata(t *testing.T) map[string]any {
+	t.Helper()
+	extractor, err := spells.New(orderingProductionLLMClient{}, spells.Options{})
+	if err != nil {
+		t.Fatalf("construct spell extractor: %v", err)
+	}
+	return extractor.ManifestMetadata()
+}
+
 func assertAssetNames(t *testing.T, getFS func() (fs.FS, error), want []string) {
 	t.Helper()
 	fSys, err := getFS()

internal/cli/run_test.go

@@ -138,8 +138,8 @@ func TestProductionCatalogIncludesProductionModulesValidatorsAndDefaults(t *test
 	if catalog.ArtifactCodecs == nil {
 		t.Fatal("production artifact codec registry = nil, want initialized empty registry")
 	}
-	if got := catalog.ArtifactCodecs.RegisteredKinds(); len(got) != 0 {
-		t.Fatalf("production artifact codec kinds = %#v, want legacy raw production path", got)
+	if got := catalog.ArtifactCodecs.RegisteredKinds(); !reflect.DeepEqual(got, []contracts.ArtifactKind{"dnd/spell-list"}) {
+		t.Fatalf("production artifact codec kinds = %#v, want dnd/spell-list", got)
 	}
 
 	moduleTests := []struct {

internal/framework/pipeline/extractor_registry.go

@@ -23,6 +23,7 @@ type typedExtractorEntry struct {
 	valueType       reflect.Type
 	validateOptions OptionValidator
 	builder         func(BuildRequest) (any, error)
+	rawBuilder      LegacyRawExtractorBuilder
 }
 
 func NewExtractorRegistry() *ExtractorRegistry {
@@ -92,6 +93,20 @@ func RegisterExtractor[T any](registry *ExtractorRegistry, spec ModuleSpec, cons
 }
 
 func RegisterExtractorBuilder[T any](registry *ExtractorRegistry, spec ModuleSpec, validateOptions OptionValidator, builder func(BuildRequest) (contracts.Extractor[T], error)) error {
+	return registerExtractorBuilder(registry, spec, validateOptions, builder, nil)
+}
+
+// RegisterExtractorBuilderWithRawAdapter registers a typed extractor while a
+// raw downstream remains in use. Resolution selects the adapter until the
+// registration is replaced with the typed-only builder.
+func RegisterExtractorBuilderWithRawAdapter[T any](registry *ExtractorRegistry, spec ModuleSpec, validateOptions OptionValidator, builder func(BuildRequest) (contracts.Extractor[T], error), rawBuilder LegacyRawExtractorBuilder) error {
+	if rawBuilder == nil {
+		return fmt.Errorf("extractor raw adapter builder for %q must not be nil", strings.TrimSpace(spec.Key))
+	}
+	return registerExtractorBuilder(registry, spec, validateOptions, builder, rawBuilder)
+}
+
+func registerExtractorBuilder[T any](registry *ExtractorRegistry, spec ModuleSpec, validateOptions OptionValidator, builder func(BuildRequest) (contracts.Extractor[T], error), rawBuilder LegacyRawExtractorBuilder) error {
 	if registry == nil {
 		return fmt.Errorf("extractor registry must not be nil")
 	}
@@ -118,6 +133,7 @@ func RegisterExtractorBuilder[T any](registry *ExtractorRegistry, spec ModuleSpe
 		builder: func(request BuildRequest) (any, error) {
 			return builder(cloneBuildRequest(request))
 		},
+		rawBuilder: rawBuilder,
 	}
 	if registry.typedEntries == nil {
 		registry.typedEntries = make(map[string]typedExtractorEntry)
@@ -143,6 +159,12 @@ func (r *ExtractorRegistry) BuildLegacyRawWithRequest(key string, request BuildR
 		return nil, fmt.Errorf("extractor key must not be empty")
 	}
 	builder, ok := r.legacyBuilders[normalizedKey]
+	if !ok {
+		if entry, typedOK := r.typedEntries[normalizedKey]; typedOK && entry.rawBuilder != nil {
+			builder = entry.rawBuilder
+			ok = true
+		}
+	}
 	if !ok {
 		return nil, fmt.Errorf("legacy raw extractor %q is not registered", normalizedKey)
 	}
@@ -193,6 +215,11 @@ func (r *ExtractorRegistry) typedEntry(key string) (typedExtractorEntry, bool) {
 	return entry, ok
 }
 
+func (r *ExtractorRegistry) usesRawAdapter(key string) bool {
+	entry, ok := r.typedEntry(key)
+	return ok && entry.rawBuilder != nil
+}
+
 func (r *ExtractorRegistry) RegisteredKeys() []string {
 	if r == nil {
 		return nil

internal/framework/pipeline/extractor_registry_test.go

@@ -353,10 +353,48 @@ func TestExtractorRegistryBuildRejectsEmptyKey(t *testing.T) {
 	}
 }
 
+func TestExtractorRegistryTypedRegistrationCanProvideRawAdapter(t *testing.T) {
+	registry := NewExtractorRegistry()
+	spec := ModuleSpec{Key: "typed-extractor", Stage: StageExtract, ArtifactKind: "test/value"}
+	if err := RegisterExtractorBuilderWithRawAdapter(registry, spec, func(map[string]any) error { return nil },
+		func(BuildRequest) (contracts.Extractor[registryTypedValue], error) {
+			return registryTypedExtractor{key: spec.Key}, nil
+		},
+		func(BuildRequest) (contracts.LegacyRawExtractor, error) {
+			return registryFakeExtractor{key: spec.Key}, nil
+		},
+	); err != nil {
+		t.Fatalf("RegisterExtractorBuilderWithRawAdapter() error = %v", err)
+	}
+	if !registry.usesRawAdapter(spec.Key) {
+		t.Fatal("usesRawAdapter() = false, want true")
+	}
+	if _, ok := registry.typedEntry(spec.Key); !ok {
+		t.Fatal("typedEntry() ok = false, want true")
+	}
+	adapter, err := registry.BuildLegacyRaw(spec.Key)
+	if err != nil || adapter.Key() != spec.Key {
+		t.Fatalf("BuildLegacyRaw() = %#v, %v", adapter, err)
+	}
+	if err := RegisterExtractorBuilderWithRawAdapter[registryTypedValue](NewExtractorRegistry(), spec, func(map[string]any) error { return nil }, nil, nil); err == nil || !strings.Contains(err.Error(), "raw adapter") {
+		t.Fatalf("nil raw builder error = %v, want raw adapter context", err)
+	}
+}
+
 type registryFakeExtractor struct {
 	key string
 }
 
+type registryTypedValue struct{ Value string }
+
+type registryTypedExtractor struct{ key string }
+
+func (extractor registryTypedExtractor) Key() string                     { return extractor.key }
+func (registryTypedExtractor) ReferenceSlots() []contracts.ReferenceSlot { return nil }
+func (registryTypedExtractor) Extract(context.Context, contracts.TypedExtractionRequest) (contracts.TypedExtractionResult[registryTypedValue], error) {
+	return contracts.TypedExtractionResult[registryTypedValue]{}, nil
+}
+
 func fakeExtractorConstructor(key string) LegacyRawExtractorConstructor {
 	return func() (contracts.LegacyRawExtractor, error) {
 		return registryFakeExtractor{key: key}, nil

internal/framework/pipeline/profile.go

@@ -393,6 +393,9 @@ func resolveArtifactIdentity(pipelineID, laneID string, lane *ResolvedArtifactLa
 	if !ok {
 		return nil, fmt.Errorf("pipeline %q lane %q extract module %q declares artifact kind %q without a typed registration", pipelineID, laneID, lane.Extract.Module, extractSpec.ArtifactKind)
 	}
+	if catalog.Extractors.usesRawAdapter(lane.Extract.Module) {
+		return nil, nil
+	}
 	if catalog.ArtifactCodecs == nil {
 		return nil, fmt.Errorf("pipeline %q lane %q artifact codec registry must not be nil for kind %q", pipelineID, laneID, extractSpec.ArtifactKind)
 	}
@@ -639,12 +642,16 @@ func validatePipelineReferenceDefaults(
 
 		merge := resolveBinding(laneProfile.Merge, DefaultMergeModule)
 		var artifactType reflect.Type
-		if extractSpec.ArtifactKind != "" && catalog.Extractors != nil {
+		artifactKind := extractSpec.ArtifactKind
+		if catalog.Extractors != nil && catalog.Extractors.usesRawAdapter(extract.Module) {
+			artifactKind = ""
+		}
+		if artifactKind != "" && catalog.Extractors != nil {
 			if entry, ok := catalog.Extractors.typedEntry(extract.Module); ok {
 				artifactType = entry.valueType
 			}
 		}
-		mergeSpec, err := mergerSpecForArtifact(catalog, merge.Module, extractSpec.ArtifactKind, artifactType)
+		mergeSpec, err := mergerSpecForArtifact(catalog, merge.Module, artifactKind, artifactType)
 		if err != nil {
 			return moduleLookupError(pipelineID, laneID, StageMerge, merge.Module, err)
 		}
@@ -653,7 +660,7 @@ func validatePipelineReferenceDefaults(
 		}
 
 		normalize := resolveBinding(laneProfile.Normalize, DefaultNormalizeModule)
-		normalizeSpec, err := normalizerSpecForArtifact(catalog, normalize.Module, extractSpec.ArtifactKind, artifactType)
+		normalizeSpec, err := normalizerSpecForArtifact(catalog, normalize.Module, artifactKind, artifactType)
 		if err != nil {
 			return moduleLookupError(pipelineID, laneID, StageNormalize, normalize.Module, err)
 		}

internal/modules/dnd/codec/spells/codec.go

@@ -0,0 +1,110 @@
+package spells
+
+import (
+	"bytes"
+	"embed"
+	"encoding/json"
+	"fmt"
+	"io"
+	"strings"
+
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd"
+)
+
+const (
+	SchemaID      = "notarius.dnd.spells"
+	SchemaName    = "notarius_dnd_spells_v1"
+	SchemaVersion = "v1"
+	MediaType     = "application/json"
+)
+
+//go:embed assets/schemas/dnd_spells.v1.json
+var schemaAssets embed.FS
+
+var _ contracts.ArtifactCodec[dnd.SpellList] = (*Codec)(nil)
+
+type Codec struct{}
+
+func New() *Codec { return &Codec{} }
+
+func (c *Codec) Kind() contracts.ArtifactKind { return dnd.SpellListKind }
+
+func (c *Codec) Schema() contracts.ArtifactSchema {
+	raw, err := schemaAssets.ReadFile("assets/schemas/dnd_spells.v1.json")
+	if err != nil {
+		return contracts.ArtifactSchema{}
+	}
+	return contracts.ArtifactSchema{ID: SchemaID, Name: SchemaName, Version: SchemaVersion, JSONSchema: raw}
+}
+
+func (c *Codec) MediaType() string { return MediaType }
+
+func (c *Codec) Encode(value dnd.SpellList) ([]byte, error) {
+	if err := validate(value); err != nil {
+		return nil, fmt.Errorf("encode dnd spell list: %w", err)
+	}
+	return c.EncodeCandidate(value)
+}
+
+// EncodeCandidate provides the same stable representation before typed
+// validators have approved a value on the temporary raw downstream path.
+func (c *Codec) EncodeCandidate(value dnd.SpellList) ([]byte, error) {
+	content, err := json.Marshal(value)
+	if err != nil {
+		return nil, fmt.Errorf("encode dnd spell list: %w", err)
+	}
+	return content, nil
+}
+
+func (c *Codec) Decode(content []byte) (dnd.SpellList, error) {
+	decoder := json.NewDecoder(bytes.NewReader(content))
+	decoder.DisallowUnknownFields()
+	var value dnd.SpellList
+	if err := decoder.Decode(&value); err != nil {
+		return dnd.SpellList{}, fmt.Errorf("decode dnd spell list: %w", err)
+	}
+	var trailing any
+	if err := decoder.Decode(&trailing); err != io.EOF {
+		return dnd.SpellList{}, fmt.Errorf("decode dnd spell list: multiple JSON values")
+	}
+	if err := validate(value); err != nil {
+		return dnd.SpellList{}, fmt.Errorf("decode dnd spell list: %w", err)
+	}
+	return value, nil
+}
+
+func validate(value dnd.SpellList) error {
+	if value.SpellCasts == nil {
+		return fmt.Errorf("spell_casts must be present")
+	}
+	for index, spell := range value.SpellCasts {
+		if strings.TrimSpace(spell.Caster) == "" {
+			return fmt.Errorf("spell_casts[%d].caster must not be empty", index)
+		}
+		if strings.TrimSpace(spell.Spell) == "" {
+			return fmt.Errorf("spell_casts[%d].spell must not be empty", index)
+		}
+		if strings.TrimSpace(spell.Effect) == "" {
+			return fmt.Errorf("spell_casts[%d].effect must not be empty", index)
+		}
+		if strings.TrimSpace(spell.NarrativeDescription) == "" {
+			return fmt.Errorf("spell_casts[%d].narrative_description must not be empty", index)
+		}
+		if len(spell.SourceRefs) == 0 {
+			return fmt.Errorf("spell_casts[%d].source_refs must not be empty", index)
+		}
+		for refIndex, ref := range spell.SourceRefs {
+			if strings.TrimSpace(ref.SourceID) == "" {
+				return fmt.Errorf("spell_casts[%d].source_refs[%d].source_id must not be empty", index, refIndex)
+			}
+			if ref.StartUnitID <= 0 {
+				return fmt.Errorf("spell_casts[%d].source_refs[%d].start_unit_id must be positive", index, refIndex)
+			}
+			if ref.EndUnitID <= 0 {
+				return fmt.Errorf("spell_casts[%d].source_refs[%d].end_unit_id must be positive", index, refIndex)
+			}
+		}
+	}
+	return nil
+}

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

@@ -0,0 +1,111 @@
+package spells
+
+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"
+)
+
+func TestCodecMatchesMaintainedDurableFixture(t *testing.T) {
+	raw, err := os.ReadFile("testdata/dnd_spells.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)
+	}
+	want := dnd.SpellList{SpellCasts: []dnd.SpellCast{
+		{Caster: "Aria", Spell: "Cure Wounds", Effect: "Heals an injured ally.", NarrativeDescription: "Aria restores the fighter after the fight.", SourceRefs: []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 1, EndUnitID: 2}}},
+		{Caster: "Borin", Spell: "Fire Bolt", Effect: "Scorches the wight.", NarrativeDescription: "Borin hurls fire at the wight.", SourceRefs: []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 3, EndUnitID: 3}}},
+	}}
+	if !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 stable durable JSON %s", encoded, compact.Bytes())
+	}
+	second, err := codec.Encode(value)
+	if err != nil || !bytes.Equal(second, encoded) {
+		t.Fatalf("second Encode() = %s, %v; want deterministic bytes", second, err)
+	}
+}
+
+func TestCodecOwnsDurableSchemaAndRegistersExactType(t *testing.T) {
+	codec := New()
+	schema := codec.Schema()
+	if codec.Kind() != dnd.SpellListKind || 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 spell 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)
+	}
+	spec, ok := registry.Spec(dnd.SpellListKind)
+	if !ok || spec.SchemaDigest != contracts.DigestArtifactSchema(schema) {
+		t.Fatalf("registered spec = %#v, %t", spec, ok)
+	}
+}
+
+func TestCodecStrictlyRejectsInvalidRepresentations(t *testing.T) {
+	codec := New()
+	tests := []struct {
+		name string
+		raw  string
+		want string
+	}{
+		{name: "unknown", raw: `{"spell_casts":[],"unexpected":true}`, want: "unknown field"},
+		{name: "trailing", raw: `{"spell_casts":[]} {}`, want: "multiple JSON values"},
+		{name: "missing", raw: `{}`, want: "spell_casts must be present"},
+		{name: "invalid evidence", raw: `{"spell_casts":[{"caster":"Aria","spell":"Cure Wounds","effect":"Heals","narrative_description":"Aria heals","source_refs":[{"source_id":"session","start_unit_id":0,"end_unit_id":1}]}]}`, want: "start_unit_id"},
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			_, err := codec.Decode([]byte(test.raw))
+			if err == nil || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Decode() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestCodecRejectsInvalidCanonicalValues(t *testing.T) {
+	_, err := New().Encode(dnd.SpellList{})
+	if err == nil || !strings.Contains(err.Error(), "spell_casts must be present") {
+		t.Fatalf("Encode() error = %v, want strict shape error", err)
+	}
+}
+
+func TestCodecSchemaIsMutationSafe(t *testing.T) {
+	first := New().Schema()
+	first.JSONSchema[0] = '['
+	second := New().Schema()
+	if !json.Valid(second.JSONSchema) || second.JSONSchema[0] == '[' {
+		t.Fatalf("Schema() returned shared bytes: %s", second.JSONSchema)
+	}
+}

internal/modules/dnd/codec/spells/testdata/dnd_spells.v1.json

@@ -0,0 +1,30 @@
+{
+  "spell_casts": [
+    {
+      "caster": "Aria",
+      "spell": "Cure Wounds",
+      "effect": "Heals an injured ally.",
+      "narrative_description": "Aria restores the fighter after the fight.",
+      "source_refs": [
+        {
+          "source_id": "session-alpha",
+          "start_unit_id": 1,
+          "end_unit_id": 2
+        }
+      ]
+    },
+    {
+      "caster": "Borin",
+      "spell": "Fire Bolt",
+      "effect": "Scorches the wight.",
+      "narrative_description": "Borin hurls fire at the wight.",
+      "source_refs": [
+        {
+          "source_id": "session-alpha",
+          "start_unit_id": 3,
+          "end_unit_id": 3
+        }
+      ]
+    }
+  ]
+}

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

@@ -2,5 +2,5 @@ package spells
 
 import "embed"
 
-//go:embed assets/schemas/*.json assets/prompts/*.yaml assets/prompts/*.md
+//go:embed assets/schemas/dnd_spells_llm.v1.json assets/prompts/*.yaml assets/prompts/*.md
 var embeddedAssets embed.FS

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

@@ -3,15 +3,17 @@ package spells
 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"
 )
 
-func canonicalizeResponse(response *extractionResponse, sourceID string) {
+func canonicalizeResponse(response *extractionResponse) {
 	if response == nil {
 		return
 	}
 	for index := range response.SpellCasts {
-		canonicalizeSpellCast(&response.SpellCasts[index], sourceID)
+		canonicalizeSpellCast(&response.SpellCasts[index])
 	}
 	sort.SliceStable(response.SpellCasts, func(i, j int) bool {
 		left, leftOK := earliestSourceUnit(response.SpellCasts[i])
@@ -26,9 +28,8 @@ func canonicalizeResponse(response *extractionResponse, sourceID string) {
 	})
 }
 
-func canonicalizeSpellCast(spell *spellCastResponse, sourceID string) {
+func canonicalizeSpellCast(spell *spellCastResponse) {
 	for index := range spell.SourceRefs {
-		spell.SourceRefs[index].SourceID = sourceID
 		spell.SourceRefs[index].StartUnitID = canonicalUnitRef(spell.SourceRefs[index].StartUnitID)
 		spell.SourceRefs[index].EndUnitID = canonicalUnitRef(spell.SourceRefs[index].EndUnitID)
 	}
@@ -51,12 +52,12 @@ func canonicalUnitRef(ref shared.UnitRef) shared.UnitRef {
 	return shared.UnitRefFromInt(value)
 }
 
-func dedupeSourceRefs(refs []shared.SourceRefResponse) []shared.SourceRefResponse {
+func dedupeSourceRefs(refs []spellSourceRefResponse) []spellSourceRefResponse {
 	if len(refs) < 2 {
 		return refs
 	}
 	out := refs[:0]
-	var previous shared.SourceRefResponse
+	var previous spellSourceRefResponse
 	for index, ref := range refs {
 		if index > 0 && sameSourceRef(previous, ref) {
 			continue
@@ -67,9 +68,8 @@ func dedupeSourceRefs(refs []shared.SourceRefResponse) []shared.SourceRefRespons
 	return out
 }
 
-func sameSourceRef(left shared.SourceRefResponse, right shared.SourceRefResponse) bool {
-	return left.SourceID == right.SourceID &&
-		left.StartUnitID.Int() == right.StartUnitID.Int() &&
+func sameSourceRef(left spellSourceRefResponse, right spellSourceRefResponse) bool {
+	return left.StartUnitID.Int() == right.StartUnitID.Int() &&
 		left.EndUnitID.Int() == right.EndUnitID.Int()
 }
 
@@ -90,3 +90,28 @@ func unitSortValue(ref shared.UnitRef) int {
 	}
 	return value
 }
+
+func canonicalSpellList(response extractionResponse, sourceID string) dnd.SpellList {
+	spellCasts := make([]dnd.SpellCast, len(response.SpellCasts))
+	for index, spell := range response.SpellCasts {
+		refs := make([]source.SourceRef, len(spell.SourceRefs))
+		for refIndex, ref := range spell.SourceRefs {
+			refs[refIndex] = source.SourceRef{
+				SourceID:    sourceID,
+				StartUnitID: ref.StartUnitID.Int(),
+				EndUnitID:   ref.EndUnitID.Int(),
+			}
+		}
+		spellCasts[index] = dnd.SpellCast{
+			Caster:               spell.Caster,
+			Spell:                spell.Spell,
+			Effect:               spell.Effect,
+			NarrativeDescription: spell.NarrativeDescription,
+			SourceRefs:           refs,
+		}
+	}
+	if response.SpellCasts == nil {
+		spellCasts = nil
+	}
+	return dnd.SpellList{SpellCasts: spellCasts}
+}

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

@@ -3,11 +3,11 @@ package spells
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"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/shared"
 )
 
@@ -31,12 +31,31 @@ var referenceSlotDescriptions = shared.ReferenceSlotDescriptions{
 	Roster:   "Deprecated alias for party roster reference material used only for disambiguation.",
 }
 
-var _ contracts.LegacyRawExtractor = (*Extractor)(nil)
+var _ contracts.Extractor[dnd.SpellList] = (*Extractor)(nil)
 
-type Extractor struct{}
+type Options struct{}
 
-func New() *Extractor {
-	return &Extractor{}
+type Extractor struct {
+	llm contracts.StructuredLLMClient
+}
+
+type rawAdapter struct {
+	extractor *Extractor
+	codec     RawAdapterCodec
+}
+
+type RawAdapterCodec interface {
+	contracts.ArtifactCodec[dnd.SpellList]
+	EncodeCandidate(dnd.SpellList) ([]byte, error)
+}
+
+var _ contracts.LegacyRawExtractor = (*rawAdapter)(nil)
+
+func New(llmClient contracts.StructuredLLMClient, _ Options) (*Extractor, error) {
+	if llmClient == nil {
+		return nil, extractorErrorf("LLM client must not be nil")
+	}
+	return &Extractor{llm: llmClient}, nil
 }
 
 func (e *Extractor) Key() string {
@@ -67,35 +86,35 @@ func (e *Extractor) ManifestMetadata() map[string]any {
 	return metadata
 }
 
-func (e *Extractor) Extract(ctx context.Context, req contracts.ExtractionRequest) (contracts.ExtractionResult, error) {
+func (e *Extractor) Extract(ctx context.Context, req contracts.TypedExtractionRequest) (contracts.TypedExtractionResult[dnd.SpellList], error) {
 	if e == nil {
-		return contracts.ExtractionResult{}, extractorErrorf("extractor must not be nil")
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("extractor must not be nil")
+	}
+	if e.llm == nil {
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("LLM client must not be nil")
 	}
 	if ctx == nil {
-		return contracts.ExtractionResult{}, extractorErrorf("context must not be nil")
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("context must not be nil")
 	}
 	if err := ctx.Err(); err != nil {
-		return contracts.ExtractionResult{}, extractorErrorf("context error before extraction: %w", err)
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("context error before extraction: %w", err)
 	}
 	if req.Source == nil {
-		return contracts.ExtractionResult{}, extractorErrorf("source must not be nil")
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("source must not be nil")
 	}
 	if req.Chunk == nil {
-		return contracts.ExtractionResult{}, extractorErrorf("chunk must not be nil")
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("chunk must not be nil")
 	}
 	if len(req.Chunk.Units) == 0 {
-		return contracts.ExtractionResult{}, extractorErrorf("chunk %q units must not be empty", req.Chunk.ID)
-	}
-	if req.LLMClient == nil {
-		return contracts.ExtractionResult{}, extractorErrorf("LLM client must not be nil")
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("chunk %q units must not be empty", req.Chunk.ID)
 	}
 	sourceInput, err := chunkSourceInput(req)
 	if err != nil {
-		return contracts.ExtractionResult{}, err
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, err
 	}
 
 	var response extractionResponse
-	if _, err := req.LLMClient.CompleteStructured(ctx, contracts.StructuredCompletionRequest{
+	if _, err := e.llm.CompleteStructured(ctx, contracts.StructuredCompletionRequest{
 		StageName:     Key,
 		PromptID:      PromptID,
 		PromptVersion: SchemaVersion,
@@ -103,37 +122,13 @@ func (e *Extractor) Extract(ctx context.Context, req contracts.ExtractionRequest
 		SessionID:     req.SessionID,
 		Inputs:        shared.PromptInputs(sourceInput, req.References),
 	}, &response); err != nil {
-		return contracts.ExtractionResult{}, extractorErrorf("complete structured output: %w", err)
+		return contracts.TypedExtractionResult[dnd.SpellList]{}, extractorErrorf("complete structured output: %w", err)
 	}
-	canonicalizeResponse(&response, req.Source.ID)
-	content, err := json.Marshal(response)
-	if err != nil {
-		return contracts.ExtractionResult{}, extractorErrorf("marshal canonical output: %w", err)
-	}
-	schema, err := loadResponseSchema()
-	if err != nil {
-		return contracts.ExtractionResult{}, extractorErrorf("load response schema: %w", err)
-	}
-	return contracts.ExtractionResult{
-		Output: contracts.ExtractOutput{
-			Schema: contracts.ResponseSchema{
-				ID:         ResponseSchemaID,
-				Name:       ResponseSchemaName,
-				Version:    SchemaVersion,
-				JSONSchema: append([]byte(nil), schema.JSONSchema...),
-			},
-			Payload: contracts.RawPayload{
-				Content:   content,
-				MediaType: "application/json",
-				Metadata: map[string]any{
-					"spell_cast_count": len(response.SpellCasts),
-				},
-			},
-		},
-	}, nil
+	canonicalizeResponse(&response)
+	return contracts.TypedExtractionResult[dnd.SpellList]{Value: canonicalSpellList(response, req.Source.ID)}, nil
 }
 
-func chunkSourceInput(req contracts.ExtractionRequest) (contracts.LLMInputMaterial, error) {
+func chunkSourceInput(req contracts.TypedExtractionRequest) (contracts.LLMInputMaterial, error) {
 	material := req.SourceInput.Clone()
 	if len(material.Content) == 0 {
 		material = contracts.NewLLMInputMaterial("source", req.Chunk.MediaType, req.Chunk.Content, "", "")
@@ -159,16 +154,93 @@ func ModuleSpec() pipeline.ModuleSpec {
 		Stage:          pipeline.StageExtract,
 		Requires:       append([]string(nil), requiredCapabilities...),
 		Provides:       append([]string(nil), providedCapabilities...),
+		ArtifactKind:   dnd.SpellListKind,
 		ReferenceSlots: shared.ReferenceSlots(referenceSlotDescriptions),
 	}
 }
 
 func Register(registry *pipeline.ExtractorRegistry) error {
-	return registry.RegisterLegacyRawWithSpec(ModuleSpec(), func() (contracts.LegacyRawExtractor, error) {
-		return New(), nil
+	return pipeline.RegisterExtractorBuilder(registry, ModuleSpec(), validateOptions, func(request pipeline.BuildRequest) (contracts.Extractor[dnd.SpellList], error) {
+		options, err := DecodeOptions(request.Options)
+		if err != nil {
+			return nil, err
+		}
+		return New(request.Dependencies.LLM, options)
 	})
 }
 
+// RegisterWithRawAdapter keeps existing raw downstream implementations usable
+// while the extractor itself produces the canonical typed artifact.
+func RegisterWithRawAdapter(registry *pipeline.ExtractorRegistry, codec RawAdapterCodec) error {
+	if codec == nil {
+		return extractorErrorf("artifact codec must not be nil")
+	}
+	build := func(request pipeline.BuildRequest) (*Extractor, error) {
+		options, err := DecodeOptions(request.Options)
+		if err != nil {
+			return nil, err
+		}
+		return New(request.Dependencies.LLM, options)
+	}
+	return pipeline.RegisterExtractorBuilderWithRawAdapter(registry, ModuleSpec(), validateOptions,
+		func(request pipeline.BuildRequest) (contracts.Extractor[dnd.SpellList], error) {
+			return build(request)
+		},
+		func(request pipeline.BuildRequest) (contracts.LegacyRawExtractor, error) {
+			extractor, err := build(request)
+			if err != nil {
+				return nil, err
+			}
+			return &rawAdapter{extractor: extractor, codec: codec}, nil
+		},
+	)
+}
+
+func (adapter *rawAdapter) Key() string { return Key }
+
+func (adapter *rawAdapter) ReferenceSlots() []contracts.ReferenceSlot {
+	return adapter.extractor.ReferenceSlots()
+}
+
+func (adapter *rawAdapter) ManifestMetadata() map[string]any {
+	return adapter.extractor.ManifestMetadata()
+}
+
+func (adapter *rawAdapter) Extract(ctx context.Context, req contracts.ExtractionRequest) (contracts.ExtractionResult, error) {
+	result, err := adapter.extractor.Extract(ctx, contracts.TypedExtractionRequest{
+		Source: req.Source, Chunk: req.Chunk, AmbientContext: req.AmbientContext,
+		SourceInput: req.SourceInput, SessionID: req.SessionID, References: req.References,
+		LLMProfile: req.LLMProfile, Metadata: req.Metadata,
+	})
+	if err != nil {
+		return contracts.ExtractionResult{}, err
+	}
+	content, err := adapter.codec.EncodeCandidate(result.Value)
+	if err != nil {
+		return contracts.ExtractionResult{}, extractorErrorf("encode canonical output: %w", err)
+	}
+	schema := adapter.codec.Schema()
+	return contracts.ExtractionResult{
+		Output: contracts.ExtractOutput{
+			Schema:  contracts.ResponseSchema{ID: schema.ID, Name: schema.Name, Version: schema.Version, JSONSchema: append([]byte(nil), schema.JSONSchema...)},
+			Payload: contracts.RawPayload{Content: content, MediaType: adapter.codec.MediaType(), Metadata: map[string]any{"spell_cast_count": len(result.Value.SpellCasts)}},
+		},
+		Warnings: result.Warnings,
+	}, nil
+}
+
+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 spells extractor: "+format, args...)
 }

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

@@ -2,47 +2,55 @@ package spells
 
 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/modules/dnd"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
 )
 
-func TestExtractReturnsCanonicalOutputFromStructuredResponse(t *testing.T) {
-	client := &fakeSpellsLLMClient{
-		response: extractionResponse{
-			SpellCasts: []spellCastResponse{
-				{
-					Caster:               " Aria ",
-					Spell:                " Cure Wounds ",
-					Effect:               " Heals an injured ally. ",
-					NarrativeDescription: " Aria restores the fighter after the fight. ",
-					SourceRefs:           responseSourceRefsInt("transcript", 1, 2),
-				},
-			},
+func TestExtractReturnsCanonicalSpellListFromPrivateResponse(t *testing.T) {
+	client := &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{
+		{
+			Caster:               " Aria ",
+			Spell:                " Cure Wounds ",
+			Effect:               " Heals an injured ally. ",
+			NarrativeDescription: " Aria restores the fighter after the fight. ",
+			SourceRefs:           responseSourceRefs(1, 2),
 		},
-		content: []byte(`{"spell_casts":[{"caster":" Aria ","spell":" Cure Wounds ","effect":" Heals an injured ally. ","narrative_description":" Aria restores the fighter after the fight. ","source_refs":[{"source_id":"session-alpha","start_unit_id":1,"end_unit_id":2}]}],"raw_marker":true}`),
-	}
+	}}}
+	req := extractionRequest()
 
-	extractReq := extractionRequestWithClient(client)
-	result, err := New().Extract(context.Background(), extractReq)
+	result, err := newExtractor(t, client).Extract(context.Background(), req)
 	if err != nil {
 		t.Fatalf("Extract() error = %v, want nil", err)
 	}
+	want := dnd.SpellList{SpellCasts: []dnd.SpellCast{
+		{
+			Caster:               " Aria ",
+			Spell:                " Cure Wounds ",
+			Effect:               " Heals an injured ally. ",
+			NarrativeDescription: " Aria restores the fighter after the fight. ",
+			SourceRefs:           []source.SourceRef{{SourceID: "session-alpha", StartUnitID: 1, EndUnitID: 2}},
+		},
+	}}
+	if !reflect.DeepEqual(result.Value, want) {
+		t.Fatalf("Value = %#v, want %#v", result.Value, want)
+	}
+	if len(result.Warnings) != 0 {
+		t.Fatalf("Warnings = %#v, want none", result.Warnings)
+	}
 
 	if len(client.requests) != 1 {
 		t.Fatalf("LLM calls = %d, want 1", len(client.requests))
 	}
 	llmReq := client.requests[0]
-	if llmReq.StageName != Key {
-		t.Fatalf("StageName = %q, want %q", llmReq.StageName, Key)
-	}
-	if llmReq.PromptID != PromptID || llmReq.PromptVersion != SchemaVersion {
-		t.Fatalf("prompt = %q/%q, want %q/%q", llmReq.PromptID, llmReq.PromptVersion, PromptID, SchemaVersion)
+	if llmReq.StageName != Key || llmReq.PromptID != PromptID || llmReq.PromptVersion != SchemaVersion {
+		t.Fatalf("LLM request identity = %#v, want spell prompt", llmReq)
 	}
 	if llmReq.SessionID != "session-123" || llmReq.ProfileID != "profile-spells" {
 		t.Fatalf("session/profile = %q/%q, want session-123/profile-spells", llmReq.SessionID, llmReq.ProfileID)
@@ -51,45 +59,17 @@ func TestExtractReturnsCanonicalOutputFromStructuredResponse(t *testing.T) {
 	if transcript.Name != "transcript" || transcript.MediaType != "application/json" || transcript.Digest != "sha256:chunk" || transcript.OriginURI != "file:///session-alpha.json" {
 		t.Fatalf("transcript metadata = %#v", transcript)
 	}
-	if got := string(transcript.Content); got != string(extractReq.Chunk.Content) {
-		t.Fatalf("transcript content = %q, want chunk content %q", got, extractReq.Chunk.Content)
-	}
-
-	if result.Output.Payload.MediaType != "application/json" {
-		t.Fatalf("MediaType = %q, want application/json", result.Output.Payload.MediaType)
-	}
-	if result.Output.Schema.ID != ResponseSchemaID || result.Output.Schema.Name != ResponseSchemaName || result.Output.Schema.Version != SchemaVersion {
-		t.Fatalf("schema = %#v, want response schema provenance", result.Output.Schema)
-	}
-	if !json.Valid(result.Output.Schema.JSONSchema) {
-		t.Fatalf("schema JSON is invalid or missing: %s", result.Output.Schema.JSONSchema)
-	}
-	if strings.Contains(string(result.Output.Payload.Content), "raw_marker") {
-		t.Fatalf("content = %q, want canonical payload without raw completion marker", result.Output.Payload.Content)
-	}
-
-	var payload extractionResponse
-	if err := json.Unmarshal(result.Output.Payload.Content, &payload); err != nil {
-		t.Fatalf("Unmarshal(Content) error = %v, want nil", err)
-	}
-	if len(payload.SpellCasts) != 1 || payload.SpellCasts[0].Spell != " Cure Wounds " {
-		t.Fatalf("payload = %#v, want structured response fields", payload)
-	}
-	if got := payload.SpellCasts[0].SourceRefs[0].SourceID; got != "session-alpha" {
-		t.Fatalf("source_id = %q, want canonical source document ID", got)
+	if got := string(transcript.Content); got != string(req.Chunk.Content) {
+		t.Fatalf("transcript content = %q, want chunk content %q", got, req.Chunk.Content)
 	}
 }
 
-func TestExtractorManifestMetadataIncludesPromptAndSchemaProvenance(t *testing.T) {
-	metadata := New().ManifestMetadata()
-
+func TestExtractorManifestMetadataIncludesLLMSchemaProvenance(t *testing.T) {
+	metadata := newExtractor(t, &fakeSpellsLLMClient{}).ManifestMetadata()
 	tests := 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,
+		"prompt_id": PromptID, "prompt_version": SchemaVersion,
+		"response_schema_key": string(ResponseSchemaKey), "response_schema_id": ResponseSchemaID,
+		"response_schema_name": ResponseSchemaName, "response_schema_version": SchemaVersion,
 	}
 	for key, want := range tests {
 		if metadata[key] != want {
@@ -106,373 +86,125 @@ func TestExtractorManifestMetadataIncludesPromptAndSchemaProvenance(t *testing.T
 
 func TestExtractPassesReferencesAsPromptInputs(t *testing.T) {
 	client := &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{}}}
-	req := extractionRequestWithClient(client)
-	req.References = contracts.ReferenceSet{
-		Slots: map[string]contracts.ResolvedReferenceSlot{
-			"players": {
-				Slot: contracts.ReferenceSlot{Name: "players"},
-				Items: []contracts.ReferenceItem{
-					{SlotName: "players", Content: []byte("Alice: Aria Brightmantle")},
-				},
-			},
-			"party": {
-				Slot: contracts.ReferenceSlot{Name: "party"},
-				Items: []contracts.ReferenceItem{
-					{SlotName: "party", Content: []byte("Aria Brightmantle: party cleric")},
-				},
-			},
-			"glossary": {
-				Slot: contracts.ReferenceSlot{Name: "glossary"},
-				Items: []contracts.ReferenceItem{
-					{SlotName: "glossary", Content: []byte("Brightmantle: local temple name")},
-				},
-			},
-		},
-	}
+	req := extractionRequest()
+	req.References = contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{
+		"players":  {Slot: contracts.ReferenceSlot{Name: "players"}, Items: []contracts.ReferenceItem{{SlotName: "players", Content: []byte("Alice: Aria Brightmantle")}}},
+		"party":    {Slot: contracts.ReferenceSlot{Name: "party"}, Items: []contracts.ReferenceItem{{SlotName: "party", Content: []byte("Aria Brightmantle: party cleric")}}},
+		"glossary": {Slot: contracts.ReferenceSlot{Name: "glossary"}, Items: []contracts.ReferenceItem{{SlotName: "glossary", Content: []byte("Brightmantle: local temple name")}}},
+	}}
 
-	if _, err := New().Extract(context.Background(), req); err != nil {
+	if _, err := newExtractor(t, client).Extract(context.Background(), req); err != nil {
 		t.Fatalf("Extract() error = %v, want nil", err)
 	}
-
-	if len(client.requests) != 1 {
-		t.Fatalf("LLM calls = %d, want 1", len(client.requests))
+	inputs := client.requests[0].Inputs
+	if string(inputs["players"].Content) != "Alice: Aria Brightmantle" || string(inputs["party"].Content) != "Aria Brightmantle: party cleric" || string(inputs["glossary"].Content) != "Brightmantle: local temple name" {
+		t.Fatalf("reference inputs = %#v, want configured content", inputs)
 	}
-	request := client.requests[0]
-	if request.PromptID != PromptID || request.PromptVersion != SchemaVersion {
-		t.Fatalf("prompt = %q/%q, want %q/%q", request.PromptID, request.PromptVersion, PromptID, SchemaVersion)
-	}
-	if got := string(request.Inputs["players"].Content); got != "Alice: Aria Brightmantle" {
-		t.Fatalf("players input = %q, want reference content", got)
-	}
-	if got := string(request.Inputs["party"].Content); got != "Aria Brightmantle: party cleric" {
-		t.Fatalf("party input = %q, want reference content", got)
-	}
-	if got := string(request.Inputs["glossary"].Content); got != "Brightmantle: local temple name" {
-		t.Fatalf("glossary input = %q, want reference content", got)
-	}
-	if strings.Contains(string(request.Inputs["transcript"].Content), "Aria Brightmantle: party cleric") {
-		t.Fatalf("transcript input contains reference content")
+	if strings.Contains(string(inputs["transcript"].Content), "party cleric") {
+		t.Fatal("transcript input contains reference content")
 	}
 }
 
 func TestPromptInputsMapLegacyRosterReferenceToParty(t *testing.T) {
-	inputs := shared.PromptInputs(spellSourceInput(), contracts.ReferenceSet{
-		Slots: map[string]contracts.ResolvedReferenceSlot{
-			"roster": {
-				Slot: contracts.ReferenceSlot{Name: "roster"},
-				Items: []contracts.ReferenceItem{
-					{SlotName: "roster", Content: []byte("Legacy roster text")},
-				},
-			},
-		},
-	})
-
+	inputs := shared.PromptInputs(spellSourceInput(), contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{
+		"roster": {Slot: contracts.ReferenceSlot{Name: "roster"}, Items: []contracts.ReferenceItem{{SlotName: "roster", Content: []byte("Legacy roster text")}}},
+	}})
 	if got := string(inputs["party"].Content); got != "Legacy roster text" {
 		t.Fatalf("party input = %q, want legacy roster content", got)
 	}
 	if _, ok := inputs["roster"]; ok {
-		t.Fatalf("roster prompt input was present; want only party input")
+		t.Fatal("roster prompt input was present; want only party input")
 	}
 }
 
-func TestExtractReturnsRawOutputForEmptyResponse(t *testing.T) {
-	client := &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{}}}
-
-	result, err := New().Extract(context.Background(), extractionRequestWithClient(client))
-	if err != nil {
-		t.Fatalf("Extract() error = %v, want nil", err)
-	}
-	var payload extractionResponse
-	if err := json.Unmarshal(result.Output.Payload.Content, &payload); err != nil {
-		t.Fatalf("Unmarshal(Content) error = %v, want nil", err)
-	}
-	if len(payload.SpellCasts) != 0 {
-		t.Fatalf("SpellCasts = %#v, want none", payload.SpellCasts)
-	}
-}
-
-func TestExtractReturnsCanonicalOutputForMalformedStructuredResponse(t *testing.T) {
-	client := &fakeSpellsLLMClient{response: extractionResponse{}}
-
-	result, err := New().Extract(context.Background(), extractionRequestWithClient(client))
-	if err != nil {
-		t.Fatalf("Extract() error = %v, want nil", err)
-	}
-	if string(result.Output.Payload.Content) != `{"spell_casts":null}` {
-		t.Fatalf("content = %s, want canonical structured output", result.Output.Payload.Content)
-	}
-}
-
-func TestExtractWrapsLLMClientError(t *testing.T) {
-	client := &fakeSpellsLLMClient{err: errors.New("provider unavailable")}
-
-	_, err := New().Extract(context.Background(), extractionRequestWithClient(client))
-	if err == nil {
-		t.Fatal("Extract() error = nil, want LLM error")
-	}
-	if !strings.Contains(err.Error(), "dnd spells") || !strings.Contains(err.Error(), "provider unavailable") {
-		t.Fatalf("Extract() error = %q, want wrapped LLM context", err.Error())
-	}
-}
-
-func TestExtractRejectsInvalidRequests(t *testing.T) {
-	validClient := &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{}}}
-	validReq := extractionRequestWithClient(validClient)
-	canceledCtx, cancel := context.WithCancel(context.Background())
-	cancel()
-
+func TestExtractPreservesEmptyAndMalformedValuesForTypedValidators(t *testing.T) {
 	tests := []struct {
-		name      string
-		extractor *Extractor
-		ctx       context.Context
-		req       contracts.ExtractionRequest
-		want      string
+		name     string
+		response extractionResponse
+		wantNil  bool
 	}{
-		{name: "nil extractor", extractor: nil, ctx: context.Background(), req: validReq, want: "extractor"},
-		{name: "nil context", extractor: New(), ctx: nil, req: validReq, want: "context"},
-		{name: "canceled context", extractor: New(), ctx: canceledCtx, req: validReq, want: "context"},
-		{name: "nil source", extractor: New(), ctx: context.Background(), req: contracts.ExtractionRequest{Chunk: validReq.Chunk, LLMClient: validReq.LLMClient}, want: "source"},
-		{name: "nil chunk", extractor: New(), ctx: context.Background(), req: contracts.ExtractionRequest{Source: validReq.Source, LLMClient: validReq.LLMClient}, want: "chunk"},
-		{name: "empty chunk units", extractor: New(), ctx: context.Background(), req: emptyChunkRequest(validReq), want: "units"},
-		{name: "nil LLM client", extractor: New(), ctx: context.Background(), req: contracts.ExtractionRequest{Source: validReq.Source, Chunk: validReq.Chunk}, want: "LLM client"},
-		{name: "source input mismatches chunk", extractor: New(), ctx: context.Background(), req: mismatchedSourceInputRequest(validReq), want: "must match chunk"},
+		{name: "empty", response: extractionResponse{SpellCasts: []spellCastResponse{}}},
+		{name: "missing", response: extractionResponse{}, wantNil: true},
 	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			_, err := tt.extractor.Extract(tt.ctx, tt.req)
-			if err == nil {
-				t.Fatal("Extract() error = nil, want error")
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			result, err := newExtractor(t, &fakeSpellsLLMClient{response: test.response}).Extract(context.Background(), extractionRequest())
+			if err != nil {
+				t.Fatalf("Extract() error = %v, want nil", err)
 			}
-			if !strings.Contains(err.Error(), "dnd spells") || !strings.Contains(err.Error(), tt.want) {
-				t.Fatalf("Extract() error = %q, want %q context", err.Error(), tt.want)
+			if (result.Value.SpellCasts == nil) != test.wantNil || len(result.Value.SpellCasts) != 0 {
+				t.Fatalf("SpellCasts = %#v, want empty with nil=%t", result.Value.SpellCasts, test.wantNil)
 			}
 		})
 	}
 }
 
-func TestExtractOrdersSpellCastsByEarliestSourceUnit(t *testing.T) {
-	client := &fakeSpellsLLMClient{
-		response: extractionResponse{
-			SpellCasts: []spellCastResponse{
-				{
-					Caster:               "Bandit Shaman",
-					Spell:                "Fire Bolt",
-					Effect:               "Burns.",
-					NarrativeDescription: "Second spell.",
-					SourceRefs:           responseSourceRefs("session-alpha", 2, 2),
-				},
-				{
-					Caster:               "Aria",
-					Spell:                "Cure Wounds",
-					Effect:               "Heals.",
-					NarrativeDescription: "First spell.",
-					SourceRefs:           responseSourceRefs("session-alpha", 1, 1),
-				},
-				{
-					Caster:               "Narrator",
-					Spell:                "Unknown Spell",
-					Effect:               "No cited range.",
-					NarrativeDescription: "This should sort after cited spell casts.",
-				},
-			},
-		},
+func TestExtractWrapsLLMClientError(t *testing.T) {
+	_, err := newExtractor(t, &fakeSpellsLLMClient{err: errors.New("provider unavailable")}).Extract(context.Background(), extractionRequest())
+	if err == nil || !strings.Contains(err.Error(), "dnd spells") || !strings.Contains(err.Error(), "provider unavailable") {
+		t.Fatalf("Extract() error = %v, want wrapped provider error", err)
 	}
+}
 
-	result, err := New().Extract(context.Background(), extractionRequestWithClient(client))
+func TestExtractRejectsInvalidRequests(t *testing.T) {
+	validReq := extractionRequest()
+	canceledCtx, cancel := context.WithCancel(context.Background())
+	cancel()
+	validExtractor := newExtractor(t, &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{}}})
+	tests := []struct {
+		name      string
+		extractor *Extractor
+		ctx       context.Context
+		req       contracts.TypedExtractionRequest
+		want      string
+	}{
+		{name: "nil extractor", ctx: context.Background(), req: validReq, want: "extractor"},
+		{name: "nil context", extractor: validExtractor, req: validReq, want: "context"},
+		{name: "canceled context", extractor: validExtractor, ctx: canceledCtx, req: validReq, want: "context"},
+		{name: "nil source", extractor: validExtractor, ctx: context.Background(), req: contracts.TypedExtractionRequest{Chunk: validReq.Chunk}, want: "source"},
+		{name: "nil chunk", extractor: validExtractor, ctx: context.Background(), req: contracts.TypedExtractionRequest{Source: validReq.Source}, want: "chunk"},
+		{name: "empty chunk units", extractor: validExtractor, ctx: context.Background(), req: emptyChunkRequest(validReq), want: "units"},
+		{name: "source input mismatches chunk", extractor: validExtractor, ctx: context.Background(), req: mismatchedSourceInputRequest(validReq), want: "must match chunk"},
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			_, err := test.extractor.Extract(test.ctx, test.req)
+			if err == nil || !strings.Contains(err.Error(), "dnd spells") || !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("Extract() error = %v, want %q context", err, test.want)
+			}
+		})
+	}
+}
+
+func TestExtractOrdersAndDeduplicatesEvidence(t *testing.T) {
+	client := &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{
+		{Caster: "Borin", Spell: "Fire Bolt", Effect: "Burns.", NarrativeDescription: "Second.", SourceRefs: responseSourceRefs(2, 2)},
+		{Caster: "Aria", Spell: "Cure Wounds", Effect: "Heals.", NarrativeDescription: "First.", SourceRefs: []spellSourceRefResponse{{StartUnitID: shared.UnitRefFromInt(1), EndUnitID: shared.UnitRefFromInt(2)}, {StartUnitID: shared.UnitRefFromInt(1), EndUnitID: shared.UnitRefFromInt(2)}}},
+		{Caster: "Narrator", Spell: "Unknown", Effect: "Unknown.", NarrativeDescription: "Uncited."},
+	}}}
+	result, err := newExtractor(t, client).Extract(context.Background(), extractionRequest())
 	if err != nil {
 		t.Fatalf("Extract() error = %v, want nil", err)
 	}
-	var payload extractionResponse
-	if err := json.Unmarshal(result.Output.Payload.Content, &payload); err != nil {
-		t.Fatalf("Unmarshal(Content) error = %v, want nil", err)
+	if got := []string{result.Value.SpellCasts[0].Spell, result.Value.SpellCasts[1].Spell, result.Value.SpellCasts[2].Spell}; !reflect.DeepEqual(got, []string{"Cure Wounds", "Fire Bolt", "Unknown"}) {
+		t.Fatalf("spell order = %#v, want evidence order", got)
 	}
-	if len(payload.SpellCasts) != 3 ||
-		payload.SpellCasts[0].Spell != "Cure Wounds" ||
-		payload.SpellCasts[1].Spell != "Fire Bolt" ||
-		payload.SpellCasts[2].Spell != "Unknown Spell" {
-		t.Fatalf("spell order = %#v, want earliest source-unit order with uncited spell last", payload.SpellCasts)
+	if refs := result.Value.SpellCasts[0].SourceRefs; len(refs) != 1 || refs[0] != (source.SourceRef{SourceID: "session-alpha", StartUnitID: 1, EndUnitID: 2}) {
+		t.Fatalf("source refs = %#v, want one canonical ref", refs)
 	}
 }
 
-func TestExtractCanonicalizesSourceRefs(t *testing.T) {
-	client := &fakeSpellsLLMClient{
-		response: extractionResponse{
-			SpellCasts: []spellCastResponse{
-				{
-					Caster:               "Aria",
-					Spell:                "Cure Wounds",
-					Effect:               "Heals.",
-					NarrativeDescription: "Aria heals.",
-					SourceRefs: []shared.SourceRefResponse{
-						{SourceID: "gameplay_transcript", StartUnitID: shared.UnitRefFromInt(2), EndUnitID: shared.UnitRefFromInt(2)},
-						{SourceID: "", StartUnitID: shared.UnitRefFromInt(1), EndUnitID: shared.UnitRefFromInt(2)},
-						{SourceID: "transcript", StartUnitID: shared.UnitRefFromInt(1), EndUnitID: shared.UnitRefFromInt(2)},
-					},
-				},
-			},
-		},
-	}
-
-	result, err := New().Extract(context.Background(), extractionRequestWithClient(client))
+func TestExtractPreservesInvalidEvidenceForValidators(t *testing.T) {
+	client := &fakeSpellsLLMClient{response: extractionResponse{SpellCasts: []spellCastResponse{{
+		Caster: "Aria", Spell: "Cure Wounds", Effect: "Heals.", NarrativeDescription: "Aria heals.",
+		SourceRefs: []spellSourceRefResponse{{StartUnitID: shared.UnitRefFromInt(99), EndUnitID: shared.UnitRefFromString("missing")}},
+	}}}}
+	result, err := newExtractor(t, client).Extract(context.Background(), extractionRequest())
 	if err != nil {
 		t.Fatalf("Extract() error = %v, want nil", err)
 	}
-	var payload extractionResponse
-	if err := json.Unmarshal(result.Output.Payload.Content, &payload); err != nil {
-		t.Fatalf("Unmarshal(Content) error = %v, want nil", err)
-	}
-	refs := payload.SpellCasts[0].SourceRefs
-	if len(refs) != 2 {
-		t.Fatalf("source refs = %#v, want duplicate collapsed", refs)
-	}
-	for _, ref := range refs {
-		if ref.SourceID != "session-alpha" {
-			t.Fatalf("source ref = %#v, want canonical source_id", ref)
-		}
-	}
-	if refs[0].StartUnitID.Int() != 1 || refs[0].EndUnitID.Int() != 2 ||
-		refs[1].StartUnitID.Int() != 2 || refs[1].EndUnitID.Int() != 2 {
-		t.Fatalf("source refs = %#v, want sorted unit ranges", refs)
+	ref := result.Value.SpellCasts[0].SourceRefs[0]
+	if ref != (source.SourceRef{SourceID: "session-alpha", StartUnitID: 99}) {
+		t.Fatalf("source ref = %#v, want canonical source with invalid range preserved", ref)
 	}
 }
-
-func TestExtractPreservesInvalidSourceRefsForValidators(t *testing.T) {
-	client := &fakeSpellsLLMClient{
-		response: extractionResponse{
-			SpellCasts: []spellCastResponse{
-				{
-					Caster:               "Aria",
-					Spell:                "Cure Wounds",
-					Effect:               "Heals.",
-					NarrativeDescription: "Aria heals.",
-					SourceRefs: []shared.SourceRefResponse{
-						{SourceID: "transcript", StartUnitID: shared.UnitRefFromInt(99), EndUnitID: shared.UnitRefFromString("missing")},
-					},
-				},
-			},
-		},
-	}
-
-	result, err := New().Extract(context.Background(), extractionRequestWithClient(client))
-	if err != nil {
-		t.Fatalf("Extract() error = %v, want nil", err)
-	}
-	var payload map[string][]map[string]any
-	if err := json.Unmarshal(result.Output.Payload.Content, &payload); err != nil {
-		t.Fatalf("Unmarshal(Content) error = %v, want nil", err)
-	}
-	ref := payload["spell_casts"][0]["source_refs"].([]any)[0].(map[string]any)
-	if ref["source_id"] != "session-alpha" || ref["start_unit_id"] != float64(99) || ref["end_unit_id"] != "" {
-		t.Fatalf("source ref = %#v, want source_id canonicalized without unit repair", ref)
-	}
-}
-
-func TestExtractDefensivelyCopiesRawContent(t *testing.T) {
-	client := &fakeSpellsLLMClient{
-		response: extractionResponse{
-			SpellCasts: []spellCastResponse{
-				{
-					Caster:               "Aria",
-					Spell:                "Cure Wounds",
-					Effect:               "Heals.",
-					NarrativeDescription: "Aria heals.",
-					SourceRefs:           responseSourceRefs("session-alpha", 1, 2),
-				},
-			},
-		},
-	}
-
-	result, err := New().Extract(context.Background(), extractionRequestWithClient(client))
-	if err != nil {
-		t.Fatalf("Extract() error = %v, want nil", err)
-	}
-	client.response.SpellCasts[0].SourceRefs[0].StartUnitID = shared.UnitRefFromInt(99)
-
-	var payload extractionResponse
-	if err := json.Unmarshal(result.Output.Payload.Content, &payload); err != nil {
-		t.Fatalf("Unmarshal(Content) error = %v, want nil", err)
-	}
-	if got := payload.SpellCasts[0].SourceRefs[0].StartUnitID.String(); got != "1" {
-		t.Fatalf("source ref start = %q, want copied 1", got)
-	}
-}
-
-func extractionRequestWithClient(client contracts.StructuredLLMClient) contracts.ExtractionRequest {
-	req := promptExtractionRequest()
-	req.LLMClient = client
-	req.SourceInput = spellChunkInput(req.Chunk)
-	req.SessionID = "session-123"
-	req.LLMProfile = "profile-spells"
-	return req
-}
-
-const spellTranscriptJSON = `{"id":"session-alpha","segments":[{"id":1,"text":"Aria raises her hand and casts Cure Wounds."}]}`
-
-func spellSourceInput() contracts.LLMInputMaterial {
-	return contracts.NewLLMInputMaterial("source", "application/json", []byte(spellTranscriptJSON), "sha256:transcript", "file:///session-alpha.json")
-}
-
-func spellChunkInput(chunk *source.Chunk) contracts.LLMInputMaterial {
-	return contracts.NewLLMInputMaterial("source", chunk.MediaType, chunk.Content, "sha256:chunk", "file:///session-alpha.json")
-}
-
-func emptyChunkRequest(req contracts.ExtractionRequest) contracts.ExtractionRequest {
-	req.Chunk = &source.Chunk{
-		ID:       req.Chunk.ID,
-		SourceID: req.Chunk.SourceID,
-		Index:    req.Chunk.Index,
-	}
-	return req
-}
-
-func mismatchedSourceInputRequest(req contracts.ExtractionRequest) contracts.ExtractionRequest {
-	req.SourceInput = spellSourceInput()
-	return req
-}
-
-type fakeSpellsLLMClient struct {
-	response extractionResponse
-	content  []byte
-	err      error
-	requests []contracts.StructuredCompletionRequest
-}
-
-func (client *fakeSpellsLLMClient) CompleteStructured(ctx 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")
-	}
-	*target = client.response
-	content := append([]byte(nil), client.content...)
-	if len(content) == 0 {
-		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()
-	req.Vars = cloneVars(req.Vars)
-	return req
-}
-
-func cloneVars(in map[string]any) map[string]any {
-	if len(in) == 0 {
-		return nil
-	}
-	out := make(map[string]any, len(in))
-	for key, value := range in {
-		out[key] = value
-	}
-	return out
-}

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

@@ -2,21 +2,19 @@ package spells
 
 import "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
 
-type SpellCast struct {
-	Caster               string `json:"caster"`
-	Spell                string `json:"spell"`
-	Effect               string `json:"effect"`
-	NarrativeDescription string `json:"narrative_description"`
-}
-
 type extractionResponse struct {
 	SpellCasts []spellCastResponse `json:"spell_casts"`
 }
 
 type spellCastResponse struct {
-	Caster               string                     `json:"caster"`
-	Spell                string                     `json:"spell"`
-	Effect               string                     `json:"effect"`
-	NarrativeDescription string                     `json:"narrative_description"`
-	SourceRefs           []shared.SourceRefResponse `json:"source_refs"`
+	Caster               string                   `json:"caster"`
+	Spell                string                   `json:"spell"`
+	Effect               string                   `json:"effect"`
+	NarrativeDescription string                   `json:"narrative_description"`
+	SourceRefs           []spellSourceRefResponse `json:"source_refs"`
+}
+
+type spellSourceRefResponse struct {
+	StartUnitID shared.UnitRef `json:"start_unit_id"`
+	EndUnitID   shared.UnitRef `json:"end_unit_id"`
 }

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

@@ -7,13 +7,14 @@ 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"
 )
 
-func TestNewReturnsExtractorWithMetadata(t *testing.T) {
-	extractor := New()
-	if extractor == nil {
-		t.Fatal("New() = nil, want extractor")
+func TestNewRequiresLLMClientAndReturnsExtractor(t *testing.T) {
+	if _, err := New(nil, Options{}); err == nil || !strings.Contains(err.Error(), "LLM client") {
+		t.Fatalf("New(nil) error = %v, want LLM client error", err)
 	}
+	extractor := newExtractor(t, &fakeSpellsLLMClient{})
 	if extractor.Key() != Key {
 		t.Fatalf("extractor.Key() = %q, want %q", extractor.Key(), Key)
 	}
@@ -31,6 +32,7 @@ func TestModuleSpec(t *testing.T) {
 		Provides: []string{
 			"dnd.spell_casts",
 		},
+		ArtifactKind: dnd.SpellListKind,
 		ReferenceSlots: []contracts.ReferenceSlot{
 			{
 				Name:               "glossary",
@@ -74,12 +76,11 @@ func TestRegisterMakesExtractorBuildable(t *testing.T) {
 		t.Fatalf("Register() error = %v, want nil", err)
 	}
 
-	extractor, err := registry.BuildLegacyRaw(Key)
-	if err != nil {
-		t.Fatalf("Build() error = %v, want nil", err)
+	if _, err := registry.BuildLegacyRaw(Key); err == nil || !strings.Contains(err.Error(), "legacy raw") {
+		t.Fatalf("BuildLegacyRaw() error = %v, want typed registration error", err)
 	}
-	if extractor.Key() != Key {
-		t.Fatalf("extractor.Key() = %q, want %q", extractor.Key(), Key)
+	if _, err := DecodeOptions(map[string]any{"unexpected": true}); err == nil || !strings.Contains(err.Error(), "unknown option") {
+		t.Fatalf("DecodeOptions() error = %v, want unknown option error", err)
 	}
 }
 
@@ -101,7 +102,7 @@ func TestRegisterStoresModuleSpec(t *testing.T) {
 }
 
 func TestRuntimeReferenceSlotsMatchModuleSpec(t *testing.T) {
-	extractor := New()
+	extractor := newExtractor(t, &fakeSpellsLLMClient{})
 	spec := ModuleSpec()
 
 	if !reflect.DeepEqual(extractor.ReferenceSlots(), spec.ReferenceSlots) {

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

@@ -15,6 +15,6 @@ func loadResponseSchema() (llm.ResponseSchema, error) {
 		ID:        ResponseSchemaID,
 		Version:   SchemaVersion,
 		Name:      ResponseSchemaName,
-		AssetPath: "assets/schemas/dnd_spells.v1.json",
+		AssetPath: "assets/schemas/dnd_spells_llm.v1.json",
 	})
 }

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

@@ -1,130 +1,62 @@
 package spells
 
 import (
+	"bytes"
 	"encoding/json"
 	"strings"
 	"testing"
 )
 
-func TestLoadResponseSchemaForSpells(t *testing.T) {
+func TestLoadResponseSchemaUsesExtractorOwnedLLMSchema(t *testing.T) {
 	schema, err := loadResponseSchema()
 	if err != nil {
 		t.Fatalf("loadResponseSchema() error = %v, want nil", err)
 	}
-	if schema.Key != ResponseSchemaKey {
-		t.Fatalf("schema.Key = %q, want %q", schema.Key, ResponseSchemaKey)
+	if schema.Key != ResponseSchemaKey || schema.ID != ResponseSchemaID || schema.Version != SchemaVersion || schema.Name != ResponseSchemaName {
+		t.Fatalf("schema identity = %#v, want maintained response identity", schema)
 	}
-	if schema.ID != ResponseSchemaID {
-		t.Fatalf("schema.ID = %q, want %q", schema.ID, ResponseSchemaID)
-	}
-	if schema.Version != SchemaVersion {
-		t.Fatalf("schema.Version = %q, want %q", schema.Version, SchemaVersion)
-	}
-	if schema.Name != ResponseSchemaName {
-		t.Fatalf("schema.Name = %q, want %q", schema.Name, ResponseSchemaName)
-	}
-	if !strings.HasPrefix(schema.SHA256, "sha256:") {
-		t.Fatalf("schema.SHA256 = %q, want sha256 prefix", schema.SHA256)
-	}
-	if !json.Valid(schema.JSONSchema) {
-		t.Fatalf("schema.JSONSchema is invalid JSON: %s", schema.JSONSchema)
+	if !strings.HasPrefix(schema.SHA256, "sha256:") || !json.Valid(schema.JSONSchema) {
+		t.Fatalf("schema metadata = %#v, want valid hashed JSON", schema)
 	}
 
 	var decoded map[string]any
 	if err := json.Unmarshal(schema.JSONSchema, &decoded); err != nil {
-		t.Fatalf("Unmarshal(schema.JSONSchema) error = %v, want nil", err)
+		t.Fatalf("Unmarshal(schema.JSONSchema) error = %v", err)
 	}
-	properties := decoded["properties"].(map[string]any)
-	spellCastProperties := properties["spell_casts"].(map[string]any)["items"].(map[string]any)["properties"].(map[string]any)
-	sourceRefProperties := spellCastProperties["source_refs"].(map[string]any)["items"].(map[string]any)["properties"].(map[string]any)
-	sourceRefRequired := spellCastProperties["source_refs"].(map[string]any)["items"].(map[string]any)["required"].([]any)
-	if !containsJSONField(sourceRefRequired, "source_id") {
-		t.Fatalf("canonical source refs required = %#v, want source_id", sourceRefRequired)
-	}
-	for _, field := range []string{"start_unit_id", "end_unit_id"} {
-		property := sourceRefProperties[field].(map[string]any)
-		if property["type"] != "integer" {
-			t.Fatalf("%s type = %#v, want integer", field, property["type"])
-		}
-		if property["minimum"] != float64(1) {
-			t.Fatalf("%s minimum = %#v, want 1", field, property["minimum"])
-		}
-	}
-}
-
-func TestLLMResponseSchemaOmitsSourceID(t *testing.T) {
-	raw, err := embeddedAssets.ReadFile("assets/schemas/dnd_spells_llm.v1.json")
-	if err != nil {
-		t.Fatalf("ReadFile(LLM schema) error = %v, want nil", err)
-	}
-	if !json.Valid(raw) {
-		t.Fatalf("LLM schema is invalid JSON: %s", raw)
-	}
-
-	var decoded map[string]any
-	if err := json.Unmarshal(raw, &decoded); err != nil {
-		t.Fatalf("Unmarshal(LLM schema) error = %v, want nil", err)
+	if decoded["$id"] != "notarius.dnd.spells.llm" {
+		t.Fatalf("LLM schema $id = %#v, want extractor transport schema", decoded["$id"])
 	}
 	properties := decoded["properties"].(map[string]any)
 	spellCastProperties := properties["spell_casts"].(map[string]any)["items"].(map[string]any)["properties"].(map[string]any)
 	sourceRefItems := spellCastProperties["source_refs"].(map[string]any)["items"].(map[string]any)
 	sourceRefProperties := sourceRefItems["properties"].(map[string]any)
-	sourceRefRequired := sourceRefItems["required"].([]any)
 	if _, ok := sourceRefProperties["source_id"]; ok {
-		t.Fatalf("LLM source ref schema contains source_id property: %#v", sourceRefProperties)
-	}
-	if containsJSONField(sourceRefRequired, "source_id") {
-		t.Fatalf("LLM source refs required = %#v, want no source_id", sourceRefRequired)
+		t.Fatalf("LLM source ref schema contains canonical source_id: %#v", sourceRefProperties)
 	}
 }
 
 func TestResponseSchemaJSONIsMutationSafe(t *testing.T) {
 	first, err := loadResponseSchema()
 	if err != nil {
-		t.Fatalf("loadResponseSchema() error = %v, want nil", err)
+		t.Fatalf("loadResponseSchema() error = %v", err)
 	}
 	first.JSONSchema[0] = '['
-
 	second, err := loadResponseSchema()
-	if err != nil {
-		t.Fatalf("loadResponseSchema() error = %v, want nil", err)
-	}
-	if !json.Valid(second.JSONSchema) {
-		t.Fatalf("schema JSON was mutated: %s", second.JSONSchema)
-	}
-	if len(second.JSONSchema) > 0 && second.JSONSchema[0] == '[' {
-		t.Fatalf("schema JSON did not use defensive copy")
+	if err != nil || !json.Valid(second.JSONSchema) || bytes.Equal(first.JSONSchema, second.JSONSchema) {
+		t.Fatalf("second schema = %s, %v; want defensive valid copy", second.JSONSchema, err)
 	}
 }
 
 func TestResponseSchemaDiagnosticsOmitRawSchema(t *testing.T) {
 	schema, err := loadResponseSchema()
 	if err != nil {
-		t.Fatalf("loadResponseSchema() error = %v, want nil", err)
+		t.Fatalf("loadResponseSchema() error = %v", err)
 	}
 	diagnostics := schema.DiagnosticsMap()
-
 	if diagnostics["key"] != ResponseSchemaKey {
-		t.Fatalf("diagnostics[key] = %#v, want %q", diagnostics["key"], ResponseSchemaKey)
-	}
-	for _, key := range []string{"id", "version", "name", "sha256"} {
-		if diagnostics[key] == "" {
-			t.Fatalf("diagnostics[%q] = %#v, want value", key, diagnostics[key])
-		}
+		t.Fatalf("diagnostics = %#v, want response key", diagnostics)
 	}
 	if _, ok := diagnostics["json_schema"]; ok {
-		t.Fatalf("diagnostics should omit raw schema content: %#v", diagnostics)
-	}
-	if _, ok := diagnostics["JSONSchema"]; ok {
-		t.Fatalf("diagnostics should omit raw schema content: %#v", diagnostics)
+		t.Fatalf("diagnostics include raw schema: %#v", diagnostics)
 	}
 }
-
-func containsJSONField(fields []any, want string) bool {
-	for _, field := range fields {
-		if field == want {
-			return true
-		}
-	}
-	return false
-}

internal/modules/dnd/extract/spells/scriptorium_assets_test.go

@@ -57,7 +57,7 @@ func TestScriptoriumPromptDiagnosticsOmitRawMaterials(t *testing.T) {
 	transcript := []byte(`{"secret":"source text"}`)
 	reference := "private party note"
 	prepared := prepareSpellsPrompt(t, transcript, "private player note", reference, " ")
-	metadata := New().ManifestMetadata()
+	metadata := newExtractor(t, &fakeSpellsLLMClient{}).ManifestMetadata()
 
 	payload, err := json.Marshal(map[string]any{
 		"prepared": map[string]any{

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

@@ -1,7 +1,9 @@
 package spells
 
 import (
+	"context"
 	"encoding/json"
+	"errors"
 	"testing"
 
 	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
@@ -9,7 +11,7 @@ import (
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/shared"
 )
 
-func promptExtractionRequest() contracts.ExtractionRequest {
+func promptExtractionRequest() contracts.TypedExtractionRequest {
 	doc := promptSourceDocument()
 	chunk := &source.Chunk{
 		ID:       "session-alpha:chunk:0",
@@ -25,7 +27,7 @@ func promptExtractionRequest() contracts.ExtractionRequest {
 		Units:     append([]source.SourceUnit(nil), doc.Units...),
 		Metadata:  map[string]any{"ignored": "chunk metadata"},
 	}
-	return contracts.ExtractionRequest{
+	return contracts.TypedExtractionRequest{
 		Source: doc,
 		Chunk:  chunk,
 	}
@@ -70,22 +72,90 @@ func mustJSON(t *testing.T, value any) string {
 	return string(encoded)
 }
 
-func responseSourceRefs(sourceID string, startUnitID int, endUnitID int) []shared.SourceRefResponse {
-	return []shared.SourceRefResponse{
+func responseSourceRefs(startUnitID int, endUnitID int) []spellSourceRefResponse {
+	return []spellSourceRefResponse{
 		{
-			SourceID:    sourceID,
 			StartUnitID: shared.UnitRefFromInt(startUnitID),
 			EndUnitID:   shared.UnitRefFromInt(endUnitID),
 		},
 	}
 }
 
-func responseSourceRefsInt(sourceID string, startUnitID int, endUnitID int) []shared.SourceRefResponse {
-	return []shared.SourceRefResponse{
-		{
-			SourceID:    sourceID,
-			StartUnitID: shared.UnitRefFromInt(startUnitID),
-			EndUnitID:   shared.UnitRefFromInt(endUnitID),
-		},
-	}
+const spellTranscriptJSON = `{"id":"session-alpha","segments":[{"id":1,"text":"Aria raises her hand and casts Cure Wounds."}]}`
+
+func spellSourceInput() contracts.LLMInputMaterial {
+	return contracts.NewLLMInputMaterial("source", "application/json", []byte(spellTranscriptJSON), "sha256:transcript", "file:///session-alpha.json")
+}
+
+func spellChunkInput(chunk *source.Chunk) contracts.LLMInputMaterial {
+	return contracts.NewLLMInputMaterial("source", chunk.MediaType, chunk.Content, "sha256:chunk", "file:///session-alpha.json")
+}
+
+func extractionRequest() contracts.TypedExtractionRequest {
+	req := promptExtractionRequest()
+	req.SourceInput = spellChunkInput(req.Chunk)
+	req.SessionID = "session-123"
+	req.LLMProfile = "profile-spells"
+	return req
+}
+
+func emptyChunkRequest(req contracts.TypedExtractionRequest) contracts.TypedExtractionRequest {
+	req.Chunk = &source.Chunk{ID: req.Chunk.ID, SourceID: req.Chunk.SourceID, Index: req.Chunk.Index}
+	return req
+}
+
+func mismatchedSourceInputRequest(req contracts.TypedExtractionRequest) contracts.TypedExtractionRequest {
+	req.SourceInput = spellSourceInput()
+	return req
+}
+
+func newExtractor(t *testing.T, client contracts.StructuredLLMClient) *Extractor {
+	t.Helper()
+	extractor, err := New(client, Options{})
+	if err != nil {
+		t.Fatalf("New() error = %v, want nil", err)
+	}
+	return extractor
+}
+
+type fakeSpellsLLMClient struct {
+	response extractionResponse
+	content  []byte
+	err      error
+	requests []contracts.StructuredCompletionRequest
+}
+
+func (client *fakeSpellsLLMClient) 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")
+	}
+	*target = client.response
+	content := append([]byte(nil), client.content...)
+	if len(content) == 0 {
+		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()
+	if len(req.Vars) == 0 {
+		req.Vars = nil
+		return req
+	}
+	vars := make(map[string]any, len(req.Vars))
+	for key, value := range req.Vars {
+		vars[key] = value
+	}
+	req.Vars = vars
+	return req
 }

internal/modules/dnd/register/register.go

@@ -7,6 +7,7 @@ import (
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/chunk/scenes"
+	spellcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/spells"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/spells"
 	spellshape "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/validate/spells/shape"
 	spellsourcerefs "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/validate/spells/source_refs"
@@ -24,8 +25,9 @@ func Register(registries pipeline.Registries, assets *llm.AssetRegistry) error {
 		name     string
 		register func() error
 	}{
+		{name: "spells codec", register: func() error { return pipeline.RegisterArtifactCodec(registries.ArtifactCodecs, spellcodec.New()) }},
 		{name: "scenes chunker", register: func() error { return scenes.Register(registries.Chunkers) }},
-		{name: "spells extractor", register: func() error { return spells.Register(registries.Extractors) }},
+		{name: "spells extractor", register: func() error { return spells.RegisterWithRawAdapter(registries.Extractors, spellcodec.New()) }},
 		{name: "spell shape validator", register: func() error { return spellshape.Register(registries.Validators) }},
 		{name: "spell source references validator", register: func() error { return spellsourcerefs.Register(registries.Validators) }},
 		{name: "spell source relatedness validator", register: func() error { return spellrelatedness.Register(registries.Validators) }},
@@ -57,6 +59,8 @@ func validateRegistries(registries pipeline.Registries, assets *llm.AssetRegistr
 	switch {
 	case registries.Chunkers == nil:
 		return fmt.Errorf("dnd registrar: chunker registry must not be nil")
+	case registries.ArtifactCodecs == nil:
+		return fmt.Errorf("dnd registrar: artifact codec registry must not be nil")
 	case registries.Extractors == nil:
 		return fmt.Errorf("dnd registrar: extractor registry must not be nil")
 	case registries.Validators == nil:

internal/modules/dnd/register/register_test.go

@@ -7,6 +7,7 @@ import (
 	"strings"
 	"testing"
 
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/llm"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/spells"
@@ -20,6 +21,9 @@ func TestRegisterAddsDNDFamily(t *testing.T) {
 	}
 	assertKeys(t, "chunkers", registries.Chunkers.RegisteredKeys(), []string{"dnd/scenes"})
 	assertKeys(t, "extractors", registries.Extractors.RegisteredKeys(), []string{"dnd/spells"})
+	if got := registries.ArtifactCodecs.RegisteredKinds(); !reflect.DeepEqual(got, []contracts.ArtifactKind{"dnd/spell-list"}) {
+		t.Fatalf("artifact codec kinds = %#v, want dnd/spell-list", got)
+	}
 	assertKeys(t, "validators", registries.Validators.RegisteredKeys(), []string{
 		"extract/dnd/spells/shape",
 		"extract/dnd/spells/source_refs",
@@ -51,7 +55,6 @@ func TestRegisterAddsDNDFamily(t *testing.T) {
 	})
 	assertAssetNames(t, assets.SchemaFS, []string{
 		"dnd_scenes.v1.json",
-		"dnd_spells.v1.json",
 		"dnd_spells_llm.v1.json",
 	})
 }
@@ -63,6 +66,7 @@ func TestRegisterRejectsMissingDNDDependenciesBeforeMutation(t *testing.T) {
 		wantErr string
 	}{
 		{name: "chunkers", remove: func(r *pipeline.Registries, _ **llm.AssetRegistry) { r.Chunkers = nil }, wantErr: "chunker registry"},
+		{name: "artifact codecs", remove: func(r *pipeline.Registries, _ **llm.AssetRegistry) { r.ArtifactCodecs = nil }, wantErr: "artifact codec registry"},
 		{name: "extractors", remove: func(r *pipeline.Registries, _ **llm.AssetRegistry) { r.Extractors = nil }, wantErr: "extractor registry"},
 		{name: "validators", remove: func(r *pipeline.Registries, _ **llm.AssetRegistry) { r.Validators = nil }, wantErr: "validator registry"},
 		{name: "validator chains", remove: func(r *pipeline.Registries, _ **llm.AssetRegistry) { r.ValidatorChains = nil }, wantErr: "validator chain registry"},
@@ -91,7 +95,7 @@ func TestRegisterReportsDuplicateDNDRegistration(t *testing.T) {
 		t.Fatalf("first Register() error = %v, want nil", err)
 	}
 	err := Register(registries, assets)
-	if err == nil || !strings.Contains(err.Error(), "register dnd scenes chunker") || !strings.Contains(err.Error(), "already registered") {
+	if err == nil || !strings.Contains(err.Error(), "register dnd spells codec") || !strings.Contains(err.Error(), "already registered") {
 		t.Fatalf("second Register() error = %v, want contextual duplicate error", err)
 	}
 }

internal/modules/dnd/types.go

@@ -0,0 +1,21 @@
+// Package dnd owns the canonical in-process artifact types for the D&D domain.
+package dnd
+
+import (
+	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
+	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
+)
+
+const SpellListKind contracts.ArtifactKind = "dnd/spell-list"
+
+type SpellList struct {
+	SpellCasts []SpellCast `json:"spell_casts"`
+}
+
+type SpellCast struct {
+	Caster               string             `json:"caster"`
+	Spell                string             `json:"spell"`
+	Effect               string             `json:"effect"`
+	NarrativeDescription string             `json:"narrative_description"`
+	SourceRefs           []source.SourceRef `json:"source_refs"`
+}

internal/modules/integration/dnd_spells_config_test.go

@@ -10,6 +10,7 @@ import (
 	"gitea.maximumdirect.net/eric/notarius/internal/core/source"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
 	"gitea.maximumdirect.net/eric/notarius/internal/framework/pipeline"
+	spellcodec "gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/codec/spells"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/dnd/extract/spells"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/generic/merge/appendorder"
 	"gitea.maximumdirect.net/eric/notarius/internal/modules/generic/normalize/noop"
@@ -152,6 +153,7 @@ func dndSpellsTestCatalog(t *testing.T, specs dndSpellsCatalogSpecs) pipeline.Mo
 	inputs := pipeline.NewInputAdapterRegistry()
 	chunkers := pipeline.NewChunkerRegistry()
 	extractors := pipeline.NewExtractorRegistry()
+	codecs := pipeline.NewArtifactCodecRegistry()
 	mergers := pipeline.NewMergerRegistry()
 	normalizers := pipeline.NewNormalizerRegistry()
 	outputs := pipeline.NewOutputEncoderRegistry()
@@ -177,13 +179,20 @@ func dndSpellsTestCatalog(t *testing.T, specs dndSpellsCatalogSpecs) pipeline.Mo
 	}
 
 	if specs.extractor.Key == "" {
-		if err := spells.Register(extractors); err != nil {
+		codec := spellcodec.New()
+		if err := pipeline.RegisterArtifactCodec(codecs, codec); err != nil {
+			t.Fatalf("register dnd spells codec: %v", err)
+		}
+		if err := spells.RegisterWithRawAdapter(extractors, codec); err != nil {
 			t.Fatalf("register dnd spells extractor: %v", err)
 		}
-	} else if err := extractors.RegisterLegacyRawWithSpec(specs.extractor, func() (contracts.LegacyRawExtractor, error) {
-		return spells.New(), nil
-	}); err != nil {
-		t.Fatalf("register dnd spells extractor override: %v", err)
+	} else {
+		specs.extractor.ArtifactKind = ""
+		if err := extractors.RegisterLegacyRawWithSpec(specs.extractor, func() (contracts.LegacyRawExtractor, error) {
+			return configLegacyExtractor{key: specs.extractor.Key}, nil
+		}); err != nil {
+			t.Fatalf("register dnd spells extractor override: %v", err)
+		}
 	}
 
 	if err := mergers.RegisterLegacyRawWithSpec(pipeline.ModuleSpec{
@@ -215,7 +224,7 @@ func dndSpellsTestCatalog(t *testing.T, specs dndSpellsCatalogSpecs) pipeline.Mo
 	return pipeline.ModuleCatalog{
 		Inputs:          inputs,
 		Chunkers:        chunkers,
-		ArtifactCodecs:  pipeline.NewArtifactCodecRegistry(),
+		ArtifactCodecs:  codecs,
 		Extractors:      extractors,
 		Mergers:         mergers,
 		Normalizers:     normalizers,
@@ -224,6 +233,14 @@ func dndSpellsTestCatalog(t *testing.T, specs dndSpellsCatalogSpecs) pipeline.Mo
 	}
 }
 
+type configLegacyExtractor struct{ key string }
+
+func (extractor configLegacyExtractor) Key() string                     { return extractor.key }
+func (configLegacyExtractor) ReferenceSlots() []contracts.ReferenceSlot { return nil }
+func (configLegacyExtractor) Extract(context.Context, contracts.ExtractionRequest) (contracts.ExtractionResult, error) {
+	return contracts.ExtractionResult{}, nil
+}
+
 func dndSpellsChunkerSpec() pipeline.ModuleSpec {
 	return pipeline.ModuleSpec{
 		Key:      "fake/chunk",