docs: sync reference docs for v0.42.78.0 — autocut floor, cross-source links, hook-timeout pattern

- KEY_FILES.md: autocut.ts entry gains the weak-top floor (minTopScore 0.5,
  [0,1] range, out-of-range ignored, acmts= cache-key part, raw-logit
  reranker exemption gated on the effective per-call model, per-call
  reranker-override cache skip) + the four new test pins; rerank.ts entry
  gains rerankerEmitsNormalizedScores; link-extraction.ts entry gains
  isCrossSourceLinksEnabled (env > DB > default false, DB-extract-paths-only
  scope); extract.ts entry gains resolveCandidateSources' three return
  shapes + skipped_missing_target/skipped_cross_source counters riding the
  returns and --json summaries.
- README.md: multi-source cross-source wikilink opt-in
  (link_resolution.cross_source) next to the global_basename opt-in.
- docs/TESTING.md: canonical PGLite block carries the 60_000 beforeAll hook
  timeout (bun's 5s default regularly trips on the full migration chain).
- TODOS.md: missing blank line before the WAL-repair section header.
- llms.txt / llms-full.txt regenerated (build:llms, same commit as the doc
  edits per the freshness gate).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

README.md

@@ -264,7 +264,7 @@ The whole loop is described in [`docs/architecture/topologies.md`](docs/architec
 
 **Hybrid search.** Vector (HNSW on pgvector) + BM25 keyword + reciprocal-rank fusion + source-tier boost + intent-aware query rewriting. Three named search modes (`conservative`, `balanced`, `tokenmax`) bundle the cost/quality knobs into a single config key. Live cost/recall comparisons in [`docs/eval/SEARCH_MODE_METHODOLOGY.md`](docs/eval/SEARCH_MODE_METHODOLOGY.md). Default: `balanced` with ZeroEntropy reranker on. Per-query graph signals notice when a top result is a hub for THAT query (adjacency boost), is corroborated across team brains (cross-source boost), or is being crowded out by weak chunks from a chatty session (session demote). Run `gbrain search "<query>" --explain` to see per-stage attribution: base score, every boost that fired, what it multiplied. `gbrain doctor` ships a `graph_signals_coverage` check; `gbrain search stats` shows fire counts and failure breakdowns. Vector retrieval pools the best chunk per page, so a page surfaces on its strongest evidence instead of losing to a neighbor on one weak chunk. Queries that match a page's title phrase or a declared free-text alias (`gbrain reindex --aliases` backfills existing pages) get boosted to the page they name. Every result carries an `evidence` tag (why it matched) and a `create_safety` hint (`exists` / `probable` / `unknown`) so an agent decides whether a page already exists instead of guessing from a raw score. `gbrain search diagnose "<query>" --target <slug>` traces which retrieval layer surfaces (or misses) a page.
 
-**Self-wiring knowledge graph.** Every `put_page` extracts entity refs from markdown/wikilinks/typed-link syntax and writes edges with zero LLM calls. Typed edges (`attended`, `works_at`, `invested_in`, `founded`, `advises`, `mentions`, …). Multi-hop traversal via `gbrain graph-query`. The graph is what produces the +31.4 P@5 lift over vector-only RAG. **Obsidian-style vaults:** bare `[[note-name]]` wikilinks that point across folders — you wrote `[[struktura]]` but the page lives at `projects/struktura.md` — resolve by basename once you opt in with `gbrain config set link_resolution.global_basename true`. Off by default; `gbrain doctor` tells you how many edges you'd gain before you flip it. See [migrating an Obsidian vault](INSTALL_FOR_AGENTS.md#step-45-wire-the-knowledge-graph).
+**Self-wiring knowledge graph.** Every `put_page` extracts entity refs from markdown/wikilinks/typed-link syntax and writes edges with zero LLM calls. Typed edges (`attended`, `works_at`, `invested_in`, `founded`, `advises`, `mentions`, …). Multi-hop traversal via `gbrain graph-query`. The graph is what produces the +31.4 P@5 lift over vector-only RAG. **Obsidian-style vaults:** bare `[[note-name]]` wikilinks that point across folders — you wrote `[[struktura]]` but the page lives at `projects/struktura.md` — resolve by basename once you opt in with `gbrain config set link_resolution.global_basename true`. Off by default; `gbrain doctor` tells you how many edges you'd gain before you flip it. See [migrating an Obsidian vault](INSTALL_FOR_AGENTS.md#step-45-wire-the-knowledge-graph). **Multi-source brains:** a wikilink whose target page lives in a *different source* creates an edge once you opt in with `gbrain config set link_resolution.cross_source true` (then run `gbrain extract links --source db`); with it off (the default), the extract summary counts the skipped candidates — cross-source links never vanish silently.
 
 **Job queue (Minions).** BullMQ-shaped, Postgres-native job queue. Durable subagents (LLM tool loops that survive crashes via two-phase pending→done persistence), shell jobs with audit, child jobs with cascading timeouts, rate leases for outbound providers, attachments via S3/Supabase storage. Replaces "spawn subagent as fire-and-forget Promise" with something that recovers from anything.
 

TODOS.md

@@ -36,6 +36,7 @@
   or the autocut floor misapplies. Prefer `scoreScale: 'unit' | 'logit'` on
   reranker recipes, consulted at the hybrid.ts gate. (Kept verbatim from the
   reviewed #3131 re-land; changing the seam belongs to its own change.)
+
 ## WAL-repair wave follow-ups (#223/#1670/#2575)
 
 - [ ] **P2 — gate auto-repair on an unclean-shutdown marker (adversarial F7).** The classifier

docs/TESTING.md

@@ -108,7 +108,7 @@ beforeAll(async () => {
   engine = new PGLiteEngine();
   await engine.connect({});
   await engine.initSchema();
-});
+}, 60_000); // PGLite full-migration-chain init needs breathing room (house pattern)
 
 afterAll(async () => {
   await engine.disconnect();
@@ -119,7 +119,7 @@ beforeEach(async () => {
 });
 ```
 
-Why this exact shape: `beforeAll` creates a single engine per file (PGLite WASM cold-start + initSchema is ~20s); `beforeEach` truncates user data via `resetPgliteState` ("two orders of magnitude faster" than fresh-engine-per-test); `afterAll` disconnects so the engine doesn't leak across file boundaries within a shard process.
+Why this exact shape: `beforeAll` creates a single engine per file (PGLite WASM cold-start + initSchema is ~20s); the explicit `60_000` hook timeout matters — bun's default is 5s, which the full migration chain regularly exceeds on cold start, and a timed-out `beforeAll` fails every test in the file with a misleading error; `beforeEach` truncates user data via `resetPgliteState` ("two orders of magnitude faster" than fresh-engine-per-test); `afterAll` disconnects so the engine doesn't leak across file boundaries within a shard process.
 
 #### `withEnv` pattern (R1 fix)
 

llms-full.txt

@@ -1778,7 +1778,7 @@ The whole loop is described in [`docs/architecture/topologies.md`](docs/architec
 
 **Hybrid search.** Vector (HNSW on pgvector) + BM25 keyword + reciprocal-rank fusion + source-tier boost + intent-aware query rewriting. Three named search modes (`conservative`, `balanced`, `tokenmax`) bundle the cost/quality knobs into a single config key. Live cost/recall comparisons in [`docs/eval/SEARCH_MODE_METHODOLOGY.md`](docs/eval/SEARCH_MODE_METHODOLOGY.md). Default: `balanced` with ZeroEntropy reranker on. Per-query graph signals notice when a top result is a hub for THAT query (adjacency boost), is corroborated across team brains (cross-source boost), or is being crowded out by weak chunks from a chatty session (session demote). Run `gbrain search "<query>" --explain` to see per-stage attribution: base score, every boost that fired, what it multiplied. `gbrain doctor` ships a `graph_signals_coverage` check; `gbrain search stats` shows fire counts and failure breakdowns. Vector retrieval pools the best chunk per page, so a page surfaces on its strongest evidence instead of losing to a neighbor on one weak chunk. Queries that match a page's title phrase or a declared free-text alias (`gbrain reindex --aliases` backfills existing pages) get boosted to the page they name. Every result carries an `evidence` tag (why it matched) and a `create_safety` hint (`exists` / `probable` / `unknown`) so an agent decides whether a page already exists instead of guessing from a raw score. `gbrain search diagnose "<query>" --target <slug>` traces which retrieval layer surfaces (or misses) a page.
 
-**Self-wiring knowledge graph.** Every `put_page` extracts entity refs from markdown/wikilinks/typed-link syntax and writes edges with zero LLM calls. Typed edges (`attended`, `works_at`, `invested_in`, `founded`, `advises`, `mentions`, …). Multi-hop traversal via `gbrain graph-query`. The graph is what produces the +31.4 P@5 lift over vector-only RAG. **Obsidian-style vaults:** bare `[[note-name]]` wikilinks that point across folders — you wrote `[[struktura]]` but the page lives at `projects/struktura.md` — resolve by basename once you opt in with `gbrain config set link_resolution.global_basename true`. Off by default; `gbrain doctor` tells you how many edges you'd gain before you flip it. See [migrating an Obsidian vault](INSTALL_FOR_AGENTS.md#step-45-wire-the-knowledge-graph).
+**Self-wiring knowledge graph.** Every `put_page` extracts entity refs from markdown/wikilinks/typed-link syntax and writes edges with zero LLM calls. Typed edges (`attended`, `works_at`, `invested_in`, `founded`, `advises`, `mentions`, …). Multi-hop traversal via `gbrain graph-query`. The graph is what produces the +31.4 P@5 lift over vector-only RAG. **Obsidian-style vaults:** bare `[[note-name]]` wikilinks that point across folders — you wrote `[[struktura]]` but the page lives at `projects/struktura.md` — resolve by basename once you opt in with `gbrain config set link_resolution.global_basename true`. Off by default; `gbrain doctor` tells you how many edges you'd gain before you flip it. See [migrating an Obsidian vault](INSTALL_FOR_AGENTS.md#step-45-wire-the-knowledge-graph). **Multi-source brains:** a wikilink whose target page lives in a *different source* creates an edge once you opt in with `gbrain config set link_resolution.cross_source true` (then run `gbrain extract links --source db`); with it off (the default), the extract summary counts the skipped candidates — cross-source links never vanish silently.
 
 **Job queue (Minions).** BullMQ-shaped, Postgres-native job queue. Durable subagents (LLM tool loops that survive crashes via two-phase pending→done persistence), shell jobs with audit, child jobs with cascading timeouts, rate leases for outbound providers, attachments via S3/Supabase storage. Replaces "spawn subagent as fire-and-forget Promise" with something that recovers from anything.