BYK · BYK · May 2, 2026 · May 2, 2026
diff --git a/AGENTS.md b/AGENTS.md
@@ -22,7 +22,7 @@
 * **Monorepo structure: @loreai/core + opencode-lore packages with Bun workspaces**: Bun workspace monorepo, 3 packages: \`packages/core/\` (\`@loreai/core\`, runtime-agnostic, esbuild → \`dist/node/\` + \`dist/bun/\`), \`packages/opencode/\` (\`opencode-lore\`, ships raw TS — build is no-op echo so \`bun --filter '\*' build\` works uniformly; OpenCode loader runs TS under Bun), \`packages/pi/\` (\`@loreai/pi\`, ~132KB esbuild bundle). Root \`package.json\` is private with \`workspaces: \["packages/\*"]\` but MUST have \`main\`/\`exports\` pointing to \`./packages/opencode/src/index.ts\` — trampoline required because OpenCode's \`file:///\` plugin loader resolves from repo root; without it plugin loading silently fails. Declarations via \`tsc -p tsconfig.build.json\` into \`dist/types/\`, copied to both target dirs for barrel re-exports. Tests: \`bun test\` from root with preload at \`packages/core/test/setup.ts\`. Workspace \`package.json\` \`main\`/\`types\` MUST point to built \`dist/\` (Node can't run \`.ts\`); workspace-internal consumers use tsconfig \`paths\` mapping \`@loreai/core\` → \`../core/src/index.ts\`. Build tsconfig MUST NOT include this paths mapping (TS6059).
 
 <!-- lore:019dbbb8-fbcd-7205-89de-e6197a32c6a7 -->
-* **OpenCode built-in compaction fully disabled — Lore owns all context management**: OpenCode built-in compaction fully disabled — Lore owns all context management. \`config\` hook sets \`cfg.compaction = { auto: false, prune: false }\`. Lore overrides manual \`/compact\` via \`experimental.session.compacting\`: (a) chunked distillation via \`backgroundDistill\`, (b) \`loadForSession\` (archived=0 post-F2) injected as \`output.context\`, (c) \`findPreviousCompactSummary(client, sessionID)\` walks newest-first for assistant messages with truthy \`info.summary\`, joins text parts with \`\n\n\`, (d) \`buildCompactPrompt({ hasDistillations, knowledge, previousSummary })\` emits \`\<previous-summary>\` anchor before SUMMARY\_TEMPLATE when present. Pi's \`session\_before\_compact\` is deterministic substitution, no anchor. If upstream adds compaction not gated by \`compaction.auto\`, Lore breaks silently. Hooks Lore registers (6 in \`packages/opencode/src/index.ts\`): \`config\`, \`event\` (message.updated/session.error/session.idle), \`experimental.chat.system.transform\` (LTM), \`experimental.chat.messages.transform\` (gradient), \`experimental.session.compacting\`, \`tool\` (recall). Any rename of these breaks Lore silently.
+* **OpenCode built-in compaction fully disabled — Lore owns all context management**: OpenCode built-in compaction fully disabled — Lore owns all context management. \`config\` hook sets \`cfg.compaction = { auto: false, prune: false }\`. Lore overrides manual \`/compact\` via \`experimental.session.compacting\`: chunked distillation via \`backgroundDistill\`, \`loadForSession\` (archived=0 post-F2) injected as \`output.context\`, \`findPreviousCompactSummary\` walks newest-first for assistant messages with truthy \`info.summary\`, \`buildCompactPrompt\` emits \`\<previous-summary>\` anchor before SUMMARY\_TEMPLATE. Pi's \`session\_before\_compact\` is deterministic substitution, no anchor. 6 hooks Lore registers in \`packages/opencode/src/index.ts\`: \`config\`, \`event\` (message.updated/session.error/session.idle), \`experimental.chat.system.transform\` (LTM), \`experimental.chat.messages.transform\` (gradient), \`experimental.session.compacting\`, \`tool\` (recall). Any rename of these or upstream compaction not gated by \`compaction.auto\` breaks Lore silently.
 
 <!-- lore:019d9af0-d691-77c7-9a8b-cc5a21037b0c -->
 * **SQLite #db/driver subpath import for Bun/Node dual-runtime**: Core uses Node subpath imports (\`#db/driver\` in package.json) to resolve \`bun:sqlite\` or \`node:sqlite\` at runtime. \`driver.bun.ts\` re-exports \`Database\` from \`bun:sqlite\` + \`sha256\` via \`node:crypto\`. \`driver.node.ts\` extends \`DatabaseSync\` from \`node:sqlite\` with \`.query(sql)\` shim using WeakMap statement caching — API parity with \`bun:sqlite\`. Tests run under Bun; esbuild bundles use \`conditions: \["node"|"bun"]\`. API differences: \`.query()\` vs \`.prepare()\`, \`{create:true}\` Bun-only, \`.transaction(fn)\` Bun-only (use manual BEGIN/COMMIT/ROLLBACK for cross-runtime). FTS5/pragmas identical. \`node:sqlite\` stable in Node 22.5+, no native addons.
@@ -37,6 +37,9 @@
 <!-- lore:019c91d6-04af-7334-8374-e8bbf14cb43d -->
 * **Calibration used DB message count instead of transformed window count — caused layer 0 false passthrough**: Gradient/calibration traps: (1) Calibration must use transformed window count via \`getLastTransformedCount()\`, not DB count — delta≈1 → layer 0 passthrough → overflow. (2) \`actualInput\` must include \`cache.write\` — cold-cache otherwise falls to layer 0. (3) Trailing pure-text assistant messages cause Anthropic prefill errors; drop loop must run at ALL layers (layer 0 shares ref with output). Never drop messages with tool parts (\`hasToolParts\`) — infinite loop. (4) Unregistered projects get zero context management → stuck compaction loops; recovery deletes messages after last good assistant message.
 
+<!-- lore:019de7f5-4a14-7bb6-ad30-714cb662bf0a -->
+* **Craft release: stale release branch must be deleted before re-triggering**: If \`release.yml\` was triggered before a needed PR merged, the resulting \`release/X.Y.Z\` branch and publish issue contain a stale snapshot. To re-cut: (1) \`gh issue close \<publish-issue-num>\`, (2) \`git push origin --delete release/X.Y.Z\`, (3) re-run \`gh workflow run release.yml -f version=auto\`. Craft will re-use the same version number if no new conventional-commit \`feat:\` landed since, or bump minor if a \`feat:\` is now included. Then accept the new publish issue. Publish workflow auto-closes the issue on success.
+
 <!-- lore:019c8f4f-67ca-7212-a8c4-8a75b230ceea -->
 * **Test DB isolation via LORE\_DB\_PATH and Bun test preload**: Lore test suite uses isolated temp DB via \`packages/core/test/setup.ts\` preload (\`bunfig.toml\` at repo root: \`preload = \["./packages/core/test/setup.ts"]\`). Preload sets \`LORE\_DB\_PATH\` to \`mkdtempSync\` path before any imports of \`src/db.ts\`; \`afterAll\` cleans up. \`src/db.ts\` checks \`LORE\_DB\_PATH\` first. \`agents-file.test.ts\` needs \`beforeEach\` cleanup for intra-file isolation and \`TEST\_UUIDS\` cleanup in \`afterAll\` (shared with \`ltm.test.ts\`). Tests covering OpenCode-specific code (plugin init, recovery functions) live in \`packages/opencode/test/\`. Driver-level tests in \`packages/core/test/db-driver.test.ts\`.
 
@@ -45,6 +48,9 @@
 
 ### Pattern
 
+<!-- lore:019de7fb-feb8-7ea2-b341-2c5d2baf36ed -->
+* **Context health diagnostics: C\_norm + R\_compression**: \`packages/core/src/temporal.ts\` exports \`temporalCnorm(timestamps, now?)\` — normalized variance of relative-existence weights ∈\[0,1]; 0=uniform, 1=concentrated in past. \`packages/core/src/distillation.ts\` exports \`compressionRatio(distilledTokens, sourceTokens)\` returning \`k/√N\`; <1.0 = aggressive/likely-lossy. Both logged per-segment in \`distillSegment\` via \`log.info\` (gated by LORE\_DEBUG=1 — currently invisible without env var). Heuristics adapted from D7x7z49/llm-context-idea (temporal clustering + square-root boundary). Credited in README 'Standing on the shoulders of'. Phase 4 (meta-threshold tuning from R\_compression) deferred pending real data — favor storing metrics as nullable columns on \`distillations\` table over log scraping.
+
 <!-- lore:019dc6bd-8425-7452-af95-9d0cf4c95f5c -->
 * **F2 metaDistill anchoring + loadForSession archive default**: F2 metaDistill anchoring: \`metaDistill\` (\`packages/core/src/distillation.ts\`) anchors on prior gen>0 meta via \`latestMetaObservations(projectPath, sessionID)\` and only consolidates NEW gen-0 since. \`recursiveUser()\` emits \`\<previous-meta-summary>\` block when anchored; byte-identical when absent. Generation chain: \`Math.max(existing.gen, priorMeta.gen) + 1\` — \`loadGen0\` only returns gen=0, must fold in \`priorMeta.generation\`. Threshold: anchored ≥1 new gen-0; first-round ≥3. \`loadForSession\` defaults to \`archived=0\`. storeDistillation→archive isn't transactional; mid-crash leaves stale meta + un-archived gen-0s, next run re-consolidates.
 
@@ -58,7 +64,7 @@
 * **Layer 4 token-budget tail (F7)**: Gradient layer 4 token-budget tail (F7): \`gradient.ts\` layer 4 ('nuclear') replaced fixed \`slice(-3)\` with \`tailBudget = clamp(usable \* 0.25, 2\_000, 8\_000)\`. Walks backward from \`currentTurnStart()\` accumulating \`estimateMessage()\` tokens until exhausted. Current turn (last user + subsequent assistants) ALWAYS included even if it exceeds budget — terminal layer must always return. Tool parts NOT stripped (would cause infinite tool-call loop). Distilled prefix unchanged. Variable scoping: \`transformInner()\` declares \`const turnStart\` inside layer 3; layer 4 must use a different name (e.g. \`nuclearTurnStart\`).
 
 <!-- lore:019cb050-ef48-7cbe-8e58-802f17c34591 -->
-* **Lore logging: LORE\_DEBUG gating for info/warn, always-on for errors**: \`packages/core/src/log.ts\`: \`log.info()\`/\`log.warn()\` suppressed unless \`LORE\_DEBUG=1|true\`; \`log.error()\` always emits. All write to stderr with \`\[lore]\` prefix. Exists because OpenCode TUI renders all stderr as red error text — routine status messages were alarming users. Use \`log.info()\` for status, \`log.warn()\` for non-actionable oddities, \`log.error()\` only in catch blocks for real failures. Never use \`console.error\` directly. LORE\_DEBUG also gates per-turn gradient diagnostics (layer, tokens, cap, prefix hash, system prompt hash) added for cache-bust investigation — invaluable for identifying byte-identity breaks but too noisy for default mode.
+* **Lore logging: LORE\_DEBUG gating for info/warn, always-on for errors**: \`packages/core/src/log.ts\`: \`log.info()\`/\`log.warn()\` suppressed unless \`LORE\_DEBUG=1|true\`; \`log.error()\` always emits to stderr with \`\[lore]\` prefix. Exists because OpenCode TUI renders all stderr as red error text. Use \`log.info()\` for status, \`log.warn()\` for non-actionable oddities, \`log.error()\` only in catch blocks. Never use \`console.error\` directly. LORE\_DEBUG also gates per-turn gradient diagnostics (layer, tokens, cap, prefix hash, system prompt hash) for cache-bust investigation. TRAP: any diagnostic added via \`log.info()\` is invisible by default — for metrics that must be observable without env-var gating, write to DB columns rather than logs.
 
 <!-- lore:019cb12a-c957-7e24-b3f5-6869f3429d13 -->
 * **Lore release process: craft + issue-label publish**: Release process (craft + issue-label publish): publishes 4 tarballs (\`@loreai/core\`, \`@loreai/opencode\`, \`@loreai/pi\`, \`opencode-lore\` legacy mirror). CI: \`bun pm pack\` each, extract, swap \`name\` via jq, repack. \`npm version --workspaces\` fails EUNSUPPORTEDPROTOCOL on \`workspace:\*\` — workaround: \`preReleaseCommand: scripts/bump-version.sh\` rewrites \`version\` via jq AND patches \`bun.lock\` workspace version fields via awk. CRITICAL: \`bun pm pack\` rewrites \`workspace:\*\` from \`bun.lock\`, NOT package.json — without lockfile patch, tarballs ship stale cross-workspace deps (ETARGET). \`actions/setup-node@v4\` MUST set \`registry-url: https://registry.npmjs.org\` or OIDC fails ENEEDAUTH.
@@ -73,7 +79,10 @@
 * **Recall logic extracted to core, thin tool wrappers per host**: Recall search+format logic lives in \`packages/core/src/recall.ts\` as host-agnostic \`runRecall({projectPath, sessionID, query, scope, llm, knowledgeEnabled, searchConfig})\` returning a formatted markdown string. Host adapters wrap it in their tool framework: \`packages/opencode/src/reflect.ts\` uses \`tool({args, execute})\` with OpenCode's Zod-ish schema; \`packages/pi/src/reflect.ts\` uses \`pi.registerTool({parameters: Type.Object(...), execute})\` with Typebox. Both adapters are ~75 lines — all BM25/FTS5/RRF fusion, vector search, cross-project discovery, lat.md section search stays in core. Pattern applies to any future host (ACP, CLI): keep logic in \`@loreai/core\`, write a thin tool-framework wrapper per host.
 
 <!-- lore:019dc5f4-9ecc-777a-bf51-855749f86e2a -->
-* **SQLite migration runner: per-statement atomic, not transactional**: SQLite migration runner per-statement atomic, not transactional: \`packages/core/src/db.ts\` \`migrate()\` calls \`database.exec(MIGRATIONS\[i])\` with no \`BEGIN\`/\`COMMIT\` wrapper. Multi-statement migrations crash mid-run leave partial progress — write idempotent SQL (e.g. F3b's WHERE clause matches 0 rows after first run). VACUUM special-cased at \`VACUUM\_MIGRATION\_INDEX = 2\` (cannot run in transaction). FTS5 content-table-backed: per-row UPDATE auto-triggers reindex via \`temporal\_fts\_update\`; bulk reindex via \`INSERT INTO temporal\_fts(temporal\_fts) VALUES('rebuild')\`.
+* **SQLite migration runner: per-statement atomic, not transactional**: SQLite migration runner (\`packages/core/src/db.ts\`) is per-statement atomic, NOT transactional: \`migrate()\` calls \`database.exec(MIGRATIONS\[i])\` with no BEGIN/COMMIT — write idempotent SQL. Traps: (1) \`ALTER TABLE ADD COLUMN\` is NOT idempotent — duplicate-column error aborts later statements (migration 7 killed \`kv\_meta\` on re-run). Fix: catch duplicate-column errors and re-exec remaining statements; \`recoverMissingObjects()\` runs every startup to idempotently \`CREATE TABLE IF NOT EXISTS\` critical tables (structural only, no backfill — add new critical tables here). (2) \`db()\` must NOT assign \`instance\` until \`migrate()\` succeeds. VACUUM special-cased at \`VACUUM\_MIGRATION\_INDEX = 2\`. FTS5 content-table-backed: per-row UPDATE auto-reindexes; bulk via \`INSERT INTO temporal\_fts(temporal\_fts) VALUES('rebuild')\`.
+
+<!-- lore:019de7fb-fec3-75a5-ae26-88193fc59abf -->
+* **Time-gap-aware detectSegments + recall recency RRF**: \`detectSegments\` (\`packages/core/src/distillation.ts\`) prefers splitting at the largest inter-message time gap when that gap is ≥3× the median gap; falls back to count-based splitting for uniform timestamps (preserves legacy behavior). Min segment size 3 — tiny trailing segments still merged into previous. Exported for tests. Recall pipeline (\`packages/core/src/recall.ts:runRecall\`) adds a recency-sorted list of temporal results to RRF fusion alongside the BM25 list, both keyed \`t:\<id>\` so RRF naturally boosts items appearing in both — no new thresholds. Pure additive; no schema changes; existing BM25-only behavior preserved when recency list is empty.
 
 ### Preference
 

diff --git a/packages/core/src/db.ts b/packages/core/src/db.ts
@@ -2,7 +2,7 @@ import { Database } from "#db/driver";
 import { join, dirname } from "path";
 import { mkdirSync } from "fs";
 
-const SCHEMA_VERSION = 11;
+const SCHEMA_VERSION = 12;
 
 const MIGRATIONS: string[] = [
   `
@@ -333,6 +333,22 @@ const MIGRATIONS: string[] = [
   WHERE content LIKE '%' || char(10) || '[tool:%'
      OR content LIKE '%' || char(10) || '[reasoning] %';
   `,
+  `
+  -- Version 12: Context health diagnostic columns on distillations.
+  --
+  -- r_compression: k/√N where k = distilled token count, N = source token
+  -- count. Values < 1.0 signal likely lossy compression. NULL for rows
+  -- created before this migration or for meta-distillations (gen > 0)
+  -- where the metric is not computed.
+  --
+  -- c_norm: normalized variance of relative-existence weights over source
+  -- message timestamps. Range [0, 1]; 0 = uniform distribution, 1 = attention
+  -- dominated by distant past. NULL for pre-migration rows or meta-distillations.
+  --
+  -- Both columns are nullable REALs — cheap to add, no backfill needed.
+  ALTER TABLE distillations ADD COLUMN r_compression REAL;
+  ALTER TABLE distillations ADD COLUMN c_norm REAL;
+  `,
 ];
 
 function dataDir() {