diff --git a/gitnexus/src/core/group/PIPELINE.md b/gitnexus/src/core/group/PIPELINE.md
new file mode 100644
index 000000000..7730b48e7
--- /dev/null
+++ b/gitnexus/src/core/group/PIPELINE.md
@@ -0,0 +1,139 @@
+# Group Analysis Pipeline
+
+Flow chart of the cross-repo contract extraction + matching pipeline.
+This covers what runs **inside this PR** (extractors + manifest) and
+the downstream handoff to the bridge storage (PR #795) and
+cross-impact query (PR #606).
+
+## High-level overview
+
+```mermaid
+flowchart TD
+  A[group.yaml] --> B[GroupConfig parser]
+  B --> C{For each repo<br/>in group}
+  C --> D[Per-repo LadybugDB<br/>indexed by main pipeline]
+
+  D --> E1[TopicExtractor]
+  D --> E2[HttpRouteExtractor]
+  D --> E3[GrpcExtractor]
+
+  E1 --> F[ExtractedContract array<br/>per repo]
+  E2 --> F
+  E3 --> F
+
+  B --> M[ManifestExtractor]
+  M --> G[Manifest contracts<br/>+ cross-links]
+
+  F --> H[Contract matching<br/>exact + wildcard]
+  G --> H
+
+  H --> I[(bridge.lbug<br/>#795)]
+
+  I --> J[runGroupImpact<br/>#606]
+  J --> K[CrossRepoImpact]
+```
+
+## Per-repo extractor pipeline
+
+Each extractor under `src/core/group/extractors/` follows the same
+two-strategy shape:
+
+```mermaid
+flowchart TD
+  R[RepoHandle + CypherExecutor<br/>for this repo] --> S{Graph-assisted<br/>Strategy A<br/>available?}
+
+  S -->|yes| A1[Cypher query against<br/>per-repo LadybugDB]
+  A1 --> A2{non-empty<br/>result?}
+  A2 -->|yes| OUT[ExtractedContract array]
+  A2 -->|no| B1
+
+  S -->|no| B1[Source-scan Strategy B]
+  B1 --> B2[glob repo source files]
+  B2 --> B3{ext in registry?}
+  B3 -->|yes| B4[Per-language plugin<br/>scan parsed tree]
+  B3 -->|no| SKIP[skip file]
+  B4 --> OUT
+
+  SKIP --> B2
+```
+
+**Strategy A** (graph-assisted) uses Cypher over edges already produced
+by the main ingestion pipeline:
+- HTTP: `HANDLES_ROUTE` / `FETCHES` edges from `(File)-[]->(Route)`
+- topic: none (pipeline doesn't yet produce topic nodes — Strategy B only)
+- gRPC: none (Strategy B + proto map only)
+
+**Strategy B** (source-scan) is 100% tree-sitter based after this PR.
+Each `*-patterns/<lang>.ts` plugin owns its grammar + S-expression
+queries; the top-level orchestrator imports neither.
+
+## Plugin architecture
+
+```mermaid
+flowchart LR
+  O[Orchestrator<br/>topic|http|grpc-extractor.ts] --> REG[REGISTRY<br/>*-patterns/index.ts]
+  REG --> P1[java.ts<br/>tree-sitter-java]
+  REG --> P2[go.ts<br/>tree-sitter-go]
+  REG --> P3[python.ts<br/>tree-sitter-python]
+  REG --> P4[node.ts<br/>JS + TS + TSX]
+  REG --> P5[php.ts<br/>tree-sitter-php<br/>HTTP only]
+  REG --> P6[proto.ts<br/>tree-sitter-proto<br/>gRPC only, optional]
+
+  P1 --> SCAN[tree-sitter-scanner.ts<br/>compilePatterns + runCompiledPatterns]
+  P2 --> SCAN
+  P3 --> SCAN
+  P4 --> SCAN
+  P5 --> SCAN
+  P6 --> SCAN
+
+  SCAN --> DET[Detection objects<br/>TopicMeta / HttpDetection / GrpcDetection]
+  DET --> O
+  O --> CT[ExtractedContract array]
+```
+
+The orchestrator never imports a grammar. Adding a new language /
+framework = drop one file in `*-patterns/`, register it in
+`index.ts`. No orchestrator edits required.
+
+## Manifest extraction
+
+```mermaid
+flowchart TD
+  Y[group.yaml links] --> ME[ManifestExtractor]
+  ME --> LOOP{for each link}
+  LOOP --> RES[resolveSymbol<br/>label-scoped Cypher]
+  RES --> OK{found?}
+  OK -->|yes| REF[real symbol uid + ref]
+  OK -->|no| SYN[synthetic uid<br/>manifest::repo::cid]
+
+  REF --> EMIT[emit provider + consumer<br/>Contract objects<br/>+ CrossLink]
+  SYN --> EMIT
+
+  EMIT --> BRIDGE[(bridge.lbug<br/>#795)]
+```
+
+Label-scoped queries in `resolveSymbol` keep accidental cross-matches
+out:
+- `topic` → `(n:Function|Method|Class|Interface)`
+- `grpc` method → `(n:Function|Method)`, service → `(n:Class|Interface)`
+- `lib` → `(n:Package|Module)`
+
+## Cross-impact query (PR #606)
+
+```mermaid
+flowchart TD
+  U[User changes symbol S<br/>in repo R] --> LI[Local impact engine<br/>per-repo uid expansion]
+  LI --> IDS[Affected uid set]
+
+  IDS --> BR[Bridge query<br/>MATCH Contract WHERE uid IN ids]
+  BR --> CL[CrossLink traversal]
+  CL --> OTHER[Matching contract in<br/>other repo]
+
+  OTHER --> FE[Fan-out impact<br/>to consuming repo]
+  FE --> OUT[CrossRepoImpact<br/>per affected repo]
+```
+
+The bridge stores every extracted contract keyed by `symbolUid`.
+Manifest-sourced contracts use the synthetic uid form so both sides
+of the `(local impact) ↔ (bridge query)` join derive the same uid
+without coordinating through any shared state.
diff --git a/gitnexus/src/core/group/extractors/manifest-extractor.ts b/gitnexus/src/core/group/extractors/manifest-extractor.ts
index 4cfcd5478..29c8f9b21 100644
--- a/gitnexus/src/core/group/extractors/manifest-extractor.ts
+++ b/gitnexus/src/core/group/extractors/manifest-extractor.ts
@@ -121,9 +121,12 @@ export class ManifestExtractor {
     // match "/suborders", and a gRPC manifest entry in a repo with any
     // .proto file would attach to a random proto symbol.
     //
-    // If resolveSymbol returns null, the extractor creates a contract with
-    // an empty symbolUid/ref — cross-impact still works via name-based
-    // matching through the `hint` path in runGroupImpact.
+    // If resolveSymbol returns null, the extractor falls back to a
+    // deterministic synthetic uid via `manifestSymbolUid(repo, contractId)`
+    // (see the function's docstring for why synthetic rather than empty).
+    // Cross-impact still works: the bridge query joins on the synthetic
+    // uid, and the local impact engine derives the same uid for the
+    // unresolved symbol — name-based hints are the additional safety net.
     try {
       let rows: Record<string, unknown>[];
       if (link.type === 'http') {
@@ -219,6 +222,29 @@ export class ManifestExtractor {
     return null;
   }
 
+  /**
+   * Build a canonical contract id for a manifest link.
+   *
+   * HTTP is the only type with two valid forms:
+   *   - Explicit method: `"GET::/api/orders"` → `"http::GET::/api/orders"`
+   *     (matches exactly against `HttpRouteExtractor` provider/consumer
+   *     contracts, which are also keyed by `http::<METHOD>::<path>`).
+   *   - Method-agnostic: `"/api/orders"` → `"http::*::/api/orders"`
+   *     — the `*` is a wildcard and is intended to match any concrete
+   *     HTTP method on that path. Wildcard-aware matching is the
+   *     responsibility of the sync / cross-impact layer (see #793);
+   *     downstream code should treat `http::*::<path>` as matching
+   *     every `http::<METHOD>::<path>` for the same path.
+   *
+   * Recommend the explicit-method form in group.yaml whenever the
+   * manifest author knows the method — it round-trips through exact
+   * equality matching without requiring wildcard logic downstream.
+   *
+   * NOTE on exhaustiveness: the switch covers every current
+   * `ContractType` variant and falls through to a `never` assertion so
+   * TypeScript fails the build if a new variant is added without a
+   * corresponding case.
+   */
   private buildContractId(type: ContractType, contract: string): string {
     switch (type) {
       case 'http': {
@@ -233,6 +259,10 @@ export class ManifestExtractor {
         return `lib::${contract}`;
       case 'custom':
         return `custom::${contract}`;
+      default: {
+        const _exhaustive: never = type;
+        throw new Error(`Unhandled ContractType: ${String(_exhaustive)}`);
+      }
     }
   }
 }