audrey 1.0.0 → 1.0.2

package/dist/src/validate.js

@@ -2,18 +2,21 @@ import { generateId } from './ulid.js';
 import { safeJsonParse } from './utils.js';
 import { buildContradictionDetectionPrompt } from './prompts.js';
 const REINFORCEMENT_THRESHOLD = 0.85;
-const CONTRADICTION_THRESHOLD = 0.60;
+const CONTRADICTION_THRESHOLD = 0.6;
 export async function validateMemory(db, embeddingProvider, episode, options = {}) {
     const { threshold = REINFORCEMENT_THRESHOLD, contradictionThreshold = CONTRADICTION_THRESHOLD, llmProvider, embeddingVector, embeddingBuffer, } = options;
-    const episodeBuffer = embeddingBuffer ?? embeddingProvider.vectorToBuffer(embeddingVector ?? await embeddingProvider.embed(episode.content));
-    const nearestSemantic = db.prepare(`
+    const episodeBuffer = embeddingBuffer ??
+        embeddingProvider.vectorToBuffer(embeddingVector ?? (await embeddingProvider.embed(episode.content)));
+    const nearestSemantic = db
+        .prepare(`
     SELECT s.*, (1.0 - v.distance) AS similarity
     FROM vec_semantics v
     JOIN semantics s ON s.id = v.id
     WHERE v.embedding MATCH ?
       AND k = 1
       AND (v.state = 'active' OR v.state = 'context_dependent')
-  `).get(episodeBuffer);
+  `)
+        .get(episodeBuffer);
     let bestMatch = null;
     let bestSimilarity = 0;
     if (nearestSemantic) {
@@ -24,7 +27,9 @@ export async function validateMemory(db, embeddingProvider, episode, options = {
         const matchId = bestMatch.id;
         const reinforce = db.transaction(() => {
             // Re-read evidence inside the transaction to avoid lost updates under concurrency.
-            const current = db.prepare('SELECT evidence_episode_ids FROM semantics WHERE id = ?').get(matchId);
+            const current = db
+                .prepare('SELECT evidence_episode_ids FROM semantics WHERE id = ?')
+                .get(matchId);
             const existing = safeJsonParse(current?.evidence_episode_ids ?? null, []);
             const wasAdded = !existing.includes(episode.id);
             if (wasAdded) {
@@ -67,7 +72,7 @@ export async function validateMemory(db, embeddingProvider, episode, options = {
             conditions: candidate.conditions &&
                 typeof candidate.conditions === 'object' &&
                 !Array.isArray(candidate.conditions) &&
-                Object.values(candidate.conditions).every((v) => typeof v === 'string')
+                Object.values(candidate.conditions).every(v => typeof v === 'string')
                 ? candidate.conditions
                 : undefined,
             explanation: typeof candidate.explanation === 'string' ? candidate.explanation : undefined,
@@ -75,7 +80,11 @@ export async function validateMemory(db, embeddingProvider, episode, options = {
         if (verdict.contradicts) {
             const matchId = bestMatch.id;
             const resolution = verdict.resolution === 'context_dependent'
-                ? { type: 'context_dependent', conditions: verdict.conditions, explanation: verdict.explanation }
+                ? {
+                    type: 'context_dependent',
+                    conditions: verdict.conditions,
+                    explanation: verdict.explanation,
+                }
                 : verdict.resolution
                     ? { type: verdict.resolution, explanation: verdict.explanation }
                     : null;
@@ -86,8 +95,7 @@ export async function validateMemory(db, embeddingProvider, episode, options = {
                     db.prepare("UPDATE semantics SET state = 'disputed' WHERE id = ?").run(matchId);
                 }
                 else if (verdict.resolution === 'context_dependent' && verdict.conditions) {
-                    db.prepare("UPDATE semantics SET state = 'context_dependent', conditions = ? WHERE id = ?")
-                        .run(JSON.stringify(verdict.conditions), matchId);
+                    db.prepare("UPDATE semantics SET state = 'context_dependent', conditions = ? WHERE id = ?").run(JSON.stringify(verdict.conditions), matchId);
                 }
             });
             recordContradiction();
@@ -107,7 +115,9 @@ function computeSourceDiversity(db, evidenceIds, currentEpisode) {
     sourceTypes.add(currentEpisode.source);
     if (evidenceIds.length > 0) {
         const placeholders = evidenceIds.map(() => '?').join(',');
-        const rows = db.prepare(`SELECT DISTINCT source FROM episodes WHERE id IN (${placeholders})`).all(...evidenceIds);
+        const rows = db
+            .prepare(`SELECT DISTINCT source FROM episodes WHERE id IN (${placeholders})`)
+            .all(...evidenceIds);
         for (const row of rows) {
             sourceTypes.add(row.source);
         }

package/dist/src/validate.js.map

@@ -1 +1 @@
-{"version":3,"file":"validate.js","sourceRoot":"","sources":["../../src/validate.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,iCAAiC,EAAE,MAAM,cAAc,CAAC;AAEjE,MAAM,uBAAuB,GAAG,IAAI,CAAC;AACrC,MAAM,uBAAuB,GAAG,IAAI,CAAC;AAkBrC,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,EAAqB,EACrB,iBAAoC,EACpC,OAAwD,EACxD,UAMI,EAAE;IAEN,MAAM,EACJ,SAAS,GAAG,uBAAuB,EACnC,sBAAsB,GAAG,uBAAuB,EAChD,WAAW,EACX,eAAe,EACf,eAAe,GAChB,GAAG,OAAO,CAAC;IAEZ,MAAM,aAAa,GAAG,eAAe,IAAI,iBAAiB,CAAC,cAAc,CACvE,eAAe,IAAI,MAAM,iBAAiB,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,CAClE,CAAC;IAEF,MAAM,eAAe,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;GAOlC,CAAC,CAAC,GAAG,CAAC,aAAa,CAAuC,CAAC;IAE5D,IAAI,SAAS,GAAkC,IAAI,CAAC;IACpD,IAAI,cAAc,GAAG,CAAC,CAAC;IAEvB,IAAI,eAAe,EAAE,CAAC;QACpB,SAAS,GAAG,eAAe,CAAC;QAC5B,cAAc,GAAG,eAAe,CAAC,UAAU,CAAC;IAC9C,CAAC;IAED,IAAI,SAAS,IAAI,cAAc,IAAI,SAAS,EAAE,CAAC;QAC7C,MAAM,OAAO,GAAG,SAAS,CAAC,EAAE,CAAC;QAC7B,MAAM,SAAS,GAAG,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE;YACpC,mFAAmF;YACnF,MAAM,OAAO,GAAG,EAAE,CAAC,OAAO,CACxB,yDAAyD,CAC1D,CAAC,GAAG,CAAC,OAAO,CAAwD,CAAC;YACtE,MAAM,QAAQ,GAAG,aAAa,CAC5B,OAAO,EAAE,oBAAoB,IAAI,IAAI,EACrC,EAAE,CACH,CAAC;YACF,MAAM,QAAQ,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAChD,IAAI,QAAQ,EAAE,CAAC;gBACb,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAC5B,CAAC;YACD,MAAM,SAAS,GAAG,sBAAsB,CAAC,EAAE,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;YAChE,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACrC,yEAAyE;YACzE,qEAAqE;YACrE,EAAE,CAAC,OAAO,CAAC;;;;;;;;OAQV,CAAC,CAAC,GAAG,CACJ,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAChB,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,EACxB,QAAQ,CAAC,MAAM,EACf,SAAS,EACT,GAAG,EACH,OAAO,CACR,CAAC;QACJ,CAAC,CAAC,CAAC;QACH,SAAS,EAAE,CAAC;QAEZ,OAAO;YACL,MAAM,EAAE,YAAY;YACpB,UAAU,EAAE,OAAO;YACnB,UAAU,EAAE,cAAc;SAC3B,CAAC;IACJ,CAAC;IAED,IAAI,SAAS,IAAI,cAAc,IAAI,sBAAsB,IAAI,WAAW,EAAE,CAAC;QACzE,MAAM,QAAQ,GAAG,iCAAiC,CAAC,OAAO,CAAC,OAAO,EAAE,SAAS,CAAC,OAAO,CAAC,CAAC;QACvF,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAC7C,IAAI,CAAC,GAAG,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,kDAAkD,CAAC,CAAC;QACtE,CAAC;QACD,MAAM,SAAS,GAAG,GAA8B,CAAC;QACjD,IAAI,OAAO,SAAS,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;YAC/C,MAAM,IAAI,KAAK,CAAC,gEAAgE,CAAC,CAAC;QACpF,CAAC;QACD,MAAM,OAAO,GAKT;YACF,WAAW,EAAE,SAAS,CAAC,WAAW;YAClC,UAAU,EAAE,OAAO,SAAS,CAAC,UAAU,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,CAAC,SAAS;YACvF,UAAU,EACR,SAAS,CAAC,UAAU;gBACpB,OAAO,SAAS,CAAC,UAAU,KAAK,QAAQ;gBACxC,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,UAAU,CAAC;gBACpC,MAAM,CAAC,MAAM,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,CAAC;gBACrE,CAAC,CAAE,SAAS,CAAC,UAAqC;gBAClD,CAAC,CAAC,SAAS;YACf,WAAW,EAAE,OAAO,SAAS,CAAC,WAAW,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,WAAW,CAAC,CAAC,CAAC,SAAS;SAC3F,CAAC;QAEF,IAAI,OAAO,CAAC,WAAW,EAAE,CAAC;YACxB,MAAM,OAAO,GAAG,SAAS,CAAC,EAAE,CAAC;YAC7B,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,KAAK,mBAAmB;gBAC3D,CAAC,CAAC,EAAE,IAAI,EAAE,mBAAmB,EAAE,UAAU,EAAE,OAAO,CAAC,UAAU,EAAE,WAAW,EAAE,OAAO,CAAC,WAAW,EAAE;gBACjG,CAAC,CAAC,OAAO,CAAC,UAAU;oBAClB,CAAC,CAAC,EAAE,IAAI,EAAE,OAAO,CAAC,UAAU,EAAE,WAAW,EAAE,OAAO,CAAC,WAAW,EAAE;oBAChE,CAAC,CAAC,IAAI,CAAC;YAEX,IAAI,eAAe,GAAG,EAAE,CAAC;YACzB,MAAM,mBAAmB,GAAG,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE;gBAC9C,eAAe,GAAG,mBAAmB,CACnC,EAAE,EACF,OAAO,EACP,UAAU,EACV,OAAO,CAAC,EAAE,EACV,UAAU,EACV,UAAU,CACX,CAAC;gBACF,IAAI,OAAO,CAAC,UAAU,KAAK,UAAU,EAAE,CAAC;oBACtC,EAAE,CAAC,OAAO,CAAC,sDAAsD,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;gBAClF,CAAC;qBAAM,IAAI,OAAO,CAAC,UAAU,KAAK,mBAAmB,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;oBAC5E,EAAE,CAAC,OAAO,CAAC,+EAA+E,CAAC;yBACxF,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,OAAO,CAAC,CAAC;gBACtD,CAAC;YACH,CAAC,CAAC,CAAC;YACH,mBAAmB,EAAE,CAAC;YAEtB,OAAO;gBACL,MAAM,EAAE,eAAe;gBACvB,eAAe;gBACf,UAAU,EAAE,OAAO;gBACnB,UAAU,EAAE,cAAc;gBAC1B,UAAU,EAAE,OAAO,CAAC,UAAU,IAAI,IAAI;aACvC,CAAC;QACJ,CAAC;IACH,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;AAC5B,CAAC;AAED,SAAS,sBAAsB,CAC7B,EAAqB,EACrB,WAAqB,EACrB,cAAkC;IAElC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,WAAW,CAAC,GAAG,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC;IAEvC,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3B,MAAM,YAAY,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC1D,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CACrB,qDAAqD,YAAY,GAAG,CACrE,CAAC,GAAG,CAAC,GAAG,WAAW,CAAgB,CAAC;QACrC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QAC9B,CAAC;IACH,CAAC;IAED,OAAO,WAAW,CAAC,IAAI,CAAC;AAC1B,CAAC;AAED,MAAM,UAAU,mBAAmB,CACjC,EAAqB,EACrB,QAAgB,EAChB,UAAkB,EAClB,QAAgB,EAChB,UAAkB,EAClB,UAAyB;IAEzB,MAAM,EAAE,GAAG,UAAU,EAAE,CAAC;IACxB,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAErC,MAAM,KAAK,GAAG,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/C,MAAM,UAAU,GAAG,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC;IAC3C,MAAM,cAAc,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IAEtE,EAAE,CAAC,OAAO,CAAC;;;;GAIV,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,KAAK,EAAE,cAAc,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC;IAE/F,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,EAAqB,EAAE,eAAuB,EAAE,aAAqB;IACvG,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACrC,EAAE,CAAC,OAAO,CAAC;;;;;;GAMV,CAAC,CAAC,GAAG,CAAC,aAAa,EAAE,GAAG,EAAE,eAAe,CAAC,CAAC;AAC9C,CAAC"}
+{"version":3,"file":"validate.js","sourceRoot":"","sources":["../../src/validate.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,iCAAiC,EAAE,MAAM,cAAc,CAAC;AAEjE,MAAM,uBAAuB,GAAG,IAAI,CAAC;AACrC,MAAM,uBAAuB,GAAG,GAAG,CAAC;AAkBpC,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,EAAqB,EACrB,iBAAoC,EACpC,OAAwD,EACxD,UAMI,EAAE;IAEN,MAAM,EACJ,SAAS,GAAG,uBAAuB,EACnC,sBAAsB,GAAG,uBAAuB,EAChD,WAAW,EACX,eAAe,EACf,eAAe,GAChB,GAAG,OAAO,CAAC;IAEZ,MAAM,aAAa,GACjB,eAAe;QACf,iBAAiB,CAAC,cAAc,CAC9B,eAAe,IAAI,CAAC,MAAM,iBAAiB,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CACpE,CAAC;IAEJ,MAAM,eAAe,GAAG,EAAE;SACvB,OAAO,CACN;;;;;;;GAOH,CACE;SACA,GAAG,CAAC,aAAa,CAAuC,CAAC;IAE5D,IAAI,SAAS,GAAkC,IAAI,CAAC;IACpD,IAAI,cAAc,GAAG,CAAC,CAAC;IAEvB,IAAI,eAAe,EAAE,CAAC;QACpB,SAAS,GAAG,eAAe,CAAC;QAC5B,cAAc,GAAG,eAAe,CAAC,UAAU,CAAC;IAC9C,CAAC;IAED,IAAI,SAAS,IAAI,cAAc,IAAI,SAAS,EAAE,CAAC;QAC7C,MAAM,OAAO,GAAG,SAAS,CAAC,EAAE,CAAC;QAC7B,MAAM,SAAS,GAAG,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE;YACpC,mFAAmF;YACnF,MAAM,OAAO,GAAG,EAAE;iBACf,OAAO,CAAC,yDAAyD,CAAC;iBAClE,GAAG,CAAC,OAAO,CAAwD,CAAC;YACvE,MAAM,QAAQ,GAAG,aAAa,CAAW,OAAO,EAAE,oBAAoB,IAAI,IAAI,EAAE,EAAE,CAAC,CAAC;YACpF,MAAM,QAAQ,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAChD,IAAI,QAAQ,EAAE,CAAC;gBACb,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YAC5B,CAAC;YACD,MAAM,SAAS,GAAG,sBAAsB,CAAC,EAAE,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;YAChE,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACrC,yEAAyE;YACzE,qEAAqE;YACrE,EAAE,CAAC,OAAO,CACR;;;;;;;;OAQD,CACA,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,QAAQ,CAAC,MAAM,EAAE,SAAS,EAAE,GAAG,EAAE,OAAO,CAAC,CAAC;QAC9F,CAAC,CAAC,CAAC;QACH,SAAS,EAAE,CAAC;QAEZ,OAAO;YACL,MAAM,EAAE,YAAY;YACpB,UAAU,EAAE,OAAO;YACnB,UAAU,EAAE,cAAc;SAC3B,CAAC;IACJ,CAAC;IAED,IAAI,SAAS,IAAI,cAAc,IAAI,sBAAsB,IAAI,WAAW,EAAE,CAAC;QACzE,MAAM,QAAQ,GAAG,iCAAiC,CAAC,OAAO,CAAC,OAAO,EAAE,SAAS,CAAC,OAAO,CAAC,CAAC;QACvF,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAC7C,IAAI,CAAC,GAAG,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,kDAAkD,CAAC,CAAC;QACtE,CAAC;QACD,MAAM,SAAS,GAAG,GAA8B,CAAC;QACjD,IAAI,OAAO,SAAS,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;YAC/C,MAAM,IAAI,KAAK,CAAC,gEAAgE,CAAC,CAAC;QACpF,CAAC;QACD,MAAM,OAAO,GAKT;YACF,WAAW,EAAE,SAAS,CAAC,WAAW;YAClC,UAAU,EAAE,OAAO,SAAS,CAAC,UAAU,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,CAAC,SAAS;YACvF,UAAU,EACR,SAAS,CAAC,UAAU;gBACpB,OAAO,SAAS,CAAC,UAAU,KAAK,QAAQ;gBACxC,CAAC,KAAK,CAAC,OAAO,CAAC,SAAS,CAAC,UAAU,CAAC;gBACpC,MAAM,CAAC,MAAM,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,CAAC;gBACnE,CAAC,CAAE,SAAS,CAAC,UAAqC;gBAClD,CAAC,CAAC,SAAS;YACf,WAAW,EAAE,OAAO,SAAS,CAAC,WAAW,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,WAAW,CAAC,CAAC,CAAC,SAAS;SAC3F,CAAC;QAEF,IAAI,OAAO,CAAC,WAAW,EAAE,CAAC;YACxB,MAAM,OAAO,GAAG,SAAS,CAAC,EAAE,CAAC;YAC7B,MAAM,UAAU,GACd,OAAO,CAAC,UAAU,KAAK,mBAAmB;gBACxC,CAAC,CAAC;oBACE,IAAI,EAAE,mBAAmB;oBACzB,UAAU,EAAE,OAAO,CAAC,UAAU;oBAC9B,WAAW,EAAE,OAAO,CAAC,WAAW;iBACjC;gBACH,CAAC,CAAC,OAAO,CAAC,UAAU;oBAClB,CAAC,CAAC,EAAE,IAAI,EAAE,OAAO,CAAC,UAAU,EAAE,WAAW,EAAE,OAAO,CAAC,WAAW,EAAE;oBAChE,CAAC,CAAC,IAAI,CAAC;YAEb,IAAI,eAAe,GAAG,EAAE,CAAC;YACzB,MAAM,mBAAmB,GAAG,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE;gBAC9C,eAAe,GAAG,mBAAmB,CACnC,EAAE,EACF,OAAO,EACP,UAAU,EACV,OAAO,CAAC,EAAE,EACV,UAAU,EACV,UAAU,CACX,CAAC;gBACF,IAAI,OAAO,CAAC,UAAU,KAAK,UAAU,EAAE,CAAC;oBACtC,EAAE,CAAC,OAAO,CAAC,sDAAsD,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;gBAClF,CAAC;qBAAM,IAAI,OAAO,CAAC,UAAU,KAAK,mBAAmB,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;oBAC5E,EAAE,CAAC,OAAO,CACR,+EAA+E,CAChF,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,OAAO,CAAC,CAAC;gBACrD,CAAC;YACH,CAAC,CAAC,CAAC;YACH,mBAAmB,EAAE,CAAC;YAEtB,OAAO;gBACL,MAAM,EAAE,eAAe;gBACvB,eAAe;gBACf,UAAU,EAAE,OAAO;gBACnB,UAAU,EAAE,cAAc;gBAC1B,UAAU,EAAE,OAAO,CAAC,UAAU,IAAI,IAAI;aACvC,CAAC;QACJ,CAAC;IACH,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;AAC5B,CAAC;AAED,SAAS,sBAAsB,CAC7B,EAAqB,EACrB,WAAqB,EACrB,cAAkC;IAElC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,WAAW,CAAC,GAAG,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC;IAEvC,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3B,MAAM,YAAY,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC1D,MAAM,IAAI,GAAG,EAAE;aACZ,OAAO,CAAC,qDAAqD,YAAY,GAAG,CAAC;aAC7E,GAAG,CAAC,GAAG,WAAW,CAAgB,CAAC;QACtC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QAC9B,CAAC;IACH,CAAC;IAED,OAAO,WAAW,CAAC,IAAI,CAAC;AAC1B,CAAC;AAED,MAAM,UAAU,mBAAmB,CACjC,EAAqB,EACrB,QAAgB,EAChB,UAAkB,EAClB,QAAgB,EAChB,UAAkB,EAClB,UAAyB;IAEzB,MAAM,EAAE,GAAG,UAAU,EAAE,CAAC;IACxB,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAErC,MAAM,KAAK,GAAG,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/C,MAAM,UAAU,GAAG,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC;IAC3C,MAAM,cAAc,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IAEtE,EAAE,CAAC,OAAO,CACR;;;;GAID,CACA,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,KAAK,EAAE,cAAc,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC;IAE9F,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,MAAM,UAAU,mBAAmB,CACjC,EAAqB,EACrB,eAAuB,EACvB,aAAqB;IAErB,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACrC,EAAE,CAAC,OAAO,CACR;;;;;;GAMD,CACA,CAAC,GAAG,CAAC,aAAa,EAAE,GAAG,EAAE,eAAe,CAAC,CAAC;AAC7C,CAAC"}

package/docs/paper/07-evaluation.md

@@ -24,13 +24,13 @@ These numbers measure Audrey's local call path under an in-process mock embeddin
 
 ## Behavioral Regression Result
 
-The current `benchmarks/output/summary.json` was generated on 2026-05-13T08:33:24.917Z with command `node benchmarks/run.js --provider mock --dimensions 64` (Ledger: E24). It reports:
+The current `benchmarks/output/summary.json` was generated on 2026-05-29T03:45:32.997Z with command `node benchmarks/run.js --provider mock --dimensions 64` (Ledger: E24). It reports:
 
 | System | Score Percent | Pass Rate | Average Duration Ms |
 |---|---:|---:|---:|
-| Audrey | 100 | 100 | 15.083333333333334 |
-| Vector Only | 41.66666666666667 | 25 | 0.25 |
-| Keyword + Recency | 41.66666666666667 | 25 | 0.5 |
+| Audrey | 100 | 100 | 15.416666666666666 |
+| Vector Only | 41.66666666666667 | 25 | 0.3333333333333333 |
+| Keyword + Recency | 41.66666666666667 | 25 | 0.6666666666666666 |
 | Recent Window | 37.5 | 25 | 0 |
 
 This output is a regression-gate result. The baselines are toy local baselines used to catch retrieval and lifecycle regressions in the Audrey codebase. They are not external systems, not tuned competitor implementations, and not GuardBench baselines (Ledger: E23-E24). The current suite covers retrieval and operation families such as information extraction, knowledge updates, multi-session reasoning, conflict resolution, procedural learning, privacy boundary, overwrite, delete-and-abstain, semantic merge, and procedural merge (Ledger: E23-E24).
@@ -55,7 +55,7 @@ It reports local adapters only, not external-system comparisons (Ledger: E46):
 | Evidence recall | 100% |
 | Redaction leaks | 0 |
 | Recall-degradation detection | 100% |
-| Guard latency p50 / p95 | 3.214 ms / 21.395 ms |
+| Guard latency p50 / p95 | 2.916 ms / 21.17 ms |
 | Published artifact raw-secret leaks | 0 |
 | Audrey Guard decision accuracy | 100% |
 | No-memory decision accuracy | 10% |

package/docs/paper/audrey-paper-v1.md

@@ -344,7 +344,7 @@ The deterministic demo, `audrey demo --scenario repeated-failure`, constructs a
 
 The current paper version has two implemented empirical anchors. First, `benchmarks/snapshots/perf-0.22.2.json` reports canonical local performance under the mock-provider methodology: generated on 2026-05-01 from git SHA `e2e821b`, using mock 64-dimensional in-process embeddings, hybrid recall limit 5, and corpus sizes 100, 1,000, and 5,000 on Node 25.5.0 with a 24-core Ryzen 9 7900X3D and 62.9 GB RAM (Ledger: E20). Under that methodology, hybrid recall p95 is 1.82 ms, 2.364 ms, and 3.417 ms for those three sizes, and encode p95 is 0.589 ms, 2.147 ms, and 1.838 ms (Ledger: E21-E22).
 
-Second, `bench:memory:check` is wired into the release gate and enforces retrieval/lifecycle benchmark guardrails against weak local baselines (Ledger: E23). The current checked-in output reports a 2026-05-08 mock-provider run in which Audrey scores 100% with 100% pass rate, while the strongest listed local baselines score 41.67% with 25% pass rate in that output (Ledger: E24). These numbers support regression-gate honesty; they do not replace GuardBench.
+Second, `bench:memory:check` is wired into the release gate as an **internal regression suite**, not a competitive benchmark. It exists to catch retrieval/lifecycle regressions in Audrey itself. The suite includes hand-tuned weak local baselines (vector-only, keyword-plus-recency, recent-window) whose role is to anchor a relative pass margin — they are not stand-ins for production memory systems, and their scores should not be cited as comparative claims about any external system (Ledger: E23). The current checked-in output reports a 2026-05-08 mock-provider run in which Audrey scores 100% with 100% pass rate and the listed local stub baselines score in the 25-42% range in that same run (Ledger: E24). These numbers support regression-gate honesty inside this repository; they are not cross-system results and they do not replace GuardBench.
 
 The README benchmark table currently differs from the canonical JSON snapshot, so the paper quotes only the JSON snapshot and tracks the README correction as a follow-up (Ledger: E28).
 
@@ -895,13 +895,13 @@ These numbers measure Audrey's local call path under an in-process mock embeddin
 
 ### Behavioral Regression Result
 
-The current `benchmarks/output/summary.json` was generated on 2026-05-13T08:33:24.917Z with command `node benchmarks/run.js --provider mock --dimensions 64` (Ledger: E24). It reports:
+The current `benchmarks/output/summary.json` was generated on 2026-05-29T03:45:32.997Z with command `node benchmarks/run.js --provider mock --dimensions 64` (Ledger: E24). It reports:
 
 | System | Score Percent | Pass Rate | Average Duration Ms |
 |---|---:|---:|---:|
-| Audrey | 100 | 100 | 15.083333333333334 |
-| Vector Only | 41.66666666666667 | 25 | 0.25 |
-| Keyword + Recency | 41.66666666666667 | 25 | 0.5 |
+| Audrey | 100 | 100 | 15.416666666666666 |
+| Vector Only | 41.66666666666667 | 25 | 0.3333333333333333 |
+| Keyword + Recency | 41.66666666666667 | 25 | 0.6666666666666666 |
 | Recent Window | 37.5 | 25 | 0 |
 
 This output is a regression-gate result. The baselines are toy local baselines used to catch retrieval and lifecycle regressions in the Audrey codebase. They are not external systems, not tuned competitor implementations, and not GuardBench baselines (Ledger: E23-E24). The current suite covers retrieval and operation families such as information extraction, knowledge updates, multi-session reasoning, conflict resolution, procedural learning, privacy boundary, overwrite, delete-and-abstain, semantic merge, and procedural merge (Ledger: E23-E24).
@@ -924,7 +924,7 @@ It reports local adapters only, not external-system comparisons (Ledger: E46):
 | Evidence recall | 100% |
 | Redaction leaks | 0 |
 | Recall-degradation detection | 100% |
-| Guard latency p50 / p95 | 3.214 ms / 21.395 ms |
+| Guard latency p50 / p95 | 2.916 ms / 21.17 ms |
 | Published artifact raw-secret leaks | 0 |
 | Audrey Guard decision accuracy | 100% |
 | No-memory decision accuracy | 10% |

package/docs/paper/evidence-ledger.md

@@ -49,7 +49,7 @@ Every implementation claim in the paper should point to one or more ledger IDs i
 | E43 - Audrey exposes a Claude Code hook generator, guarded settings apply path, and hook-mode Guard command: `hook-config claude-code` emits hooks, `hook-config claude-code --apply --scope project|user` merges them into Claude Code settings with backup/idempotence, `guard --hook --fail-on-warn` consumes PreToolUse JSON and returns `hookSpecificOutput.permissionDecision`, and `observe-tool` records post-tool events. Codex hook wiring remains pending on a stable host hook surface. | Hook integration boundary | README.md; mcp-server/index.ts; tests/mcp-server.test.js | Yes, focused Vitest and CLI hook smoke passed on 2026-05-12 |
 | E44 - Audrey preflight events now persist `preflight_evidence_ids` and `audrey_guard_action_key`; `memory_validate` accepts optional `preflight_event_id`, action key, and evidence ids, persists them on the validation audit event, and rejects validation lineage when the memory id was not evidence for that preflight. | Validation lineage implementation | src/action-key.ts; src/controller.ts; src/preflight.ts; src/audrey.ts; mcp-server/index.ts; tests/controller.test.js | Yes, focused Vitest passed on 2026-05-12 |
 | E45 - Preflight risk scoring uses a fixed severity map (`info=0.1`, `low=0.25`, `medium=0.55`, `high=0.85`), sorts warnings by severity, and strict mode blocks on high-severity warnings; the scoring path does not consume validation feedback. | Fixed risk scoring boundary | src/preflight.ts:6-60,291-299,332-338; src/feedback.ts:3-18,70-163 | Yes, 2026-05-08 |
-| E46 - `benchmarks/guardbench.js` runs ten local comparative GuardBench scenarios across Audrey Guard, no-memory, recent-window, vector-only, and FTS-only adapters and writes `benchmarks/output/guardbench-summary.json`, `benchmarks/output/guardbench-manifest.json`, and `benchmarks/output/guardbench-raw.json`; the latest local run has Audrey Guard passing 10/10 scenarios with 100% prevention rate, 0% false-block rate, 100% evidence recall, zero decision-output redaction leaks, zero published artifact raw-secret leaks, 100% recall-degradation detection, 100% decision accuracy, and 3.214ms/21.395ms p50/p95 guard latency under the mock-provider methodology. Baseline decision accuracy was no-memory 10%, recent-window 60%, vector-only 40%, and FTS-only 10%, with 0% full-contract pass rate for each baseline. | GuardBench local comparative results | benchmarks/guardbench.js; benchmarks/output/guardbench-summary.json; benchmarks/output/guardbench-manifest.json; benchmarks/output/guardbench-raw.json; package.json | Yes, `npm run bench:guard:check` passed on 2026-05-13 |
+| E46 - `benchmarks/guardbench.js` runs ten local comparative GuardBench scenarios across Audrey Guard, no-memory, recent-window, vector-only, and FTS-only adapters and writes `benchmarks/output/guardbench-summary.json`, `benchmarks/output/guardbench-manifest.json`, and `benchmarks/output/guardbench-raw.json`; the latest local run has Audrey Guard passing 10/10 scenarios with 100% prevention rate, 0% false-block rate, 100% evidence recall, zero decision-output redaction leaks, zero published artifact raw-secret leaks, 100% recall-degradation detection, 100% decision accuracy, and 2.916ms/21.17ms p50/p95 guard latency under the mock-provider methodology. Baseline decision accuracy was no-memory 10%, recent-window 60%, vector-only 40%, and FTS-only 10%, with 0% full-contract pass rate for each baseline. | GuardBench local comparative results | benchmarks/guardbench.js; benchmarks/output/guardbench-summary.json; benchmarks/output/guardbench-manifest.json; benchmarks/output/guardbench-raw.json; package.json | Yes, `npm run bench:guard:check` passed on 2026-05-13 |
 | E47 - GuardBench accepts external ESM adapters through `--adapter`, supports `default`, `adapter`, or `createGuardBenchAdapter()` exports, withholds `expectedDecision` and `requiredEvidence` during adapter execution, then scores adapter output against the same full-contract decision/evidence/redaction checks. | GuardBench external adapter contract | benchmarks/guardbench.js; tests/guardbench.test.js; package.json | Yes, `node scripts/run-vitest.mjs run tests/guardbench.test.js` passed on 2026-05-12 |
 | E48 - Audrey ships a Mem0 Platform GuardBench adapter that uses the current Mem0 REST shape: V3 async memory add with event polling, V2 filtered memory search, and user-entity cleanup. It requires runtime `MEM0_API_KEY` and is not run by default. | First external-system GuardBench adapter | benchmarks/adapters/mem0-platform.mjs; tests/guardbench.test.js; README.md | Import/contract and mocked REST-flow tests passed on 2026-05-12; live Mem0 run not yet executed |
 | E49 - GuardBench ships a credential-free example external adapter and a `bench:guard:adapter-smoke` script so the adapter loader can be exercised through the real CLI path without external credentials. | External adapter smoke path | benchmarks/adapters/example-allow.mjs; package.json; README.md; tests/guardbench.test.js | Yes, `npm run bench:guard:adapter-smoke` passed on 2026-05-12 |

package/docs/paper/output/arxiv/arxiv-manifest.json

@@ -1,11 +1,11 @@
 {
   "schemaVersion": "1.0.0",
   "suite": "Audrey arXiv source package",
-  "generatedAt": "2026-05-13T23:33:56.985Z",
+  "generatedAt": "2026-05-29T03:45:42.097Z",
   "sourceMarkdown": "docs/paper/audrey-paper-v1.md",
   "publicationPack": "docs/paper/publication-pack.json",
   "sourceHashes": {
-    "sourceMarkdown": "3f9c807f0e0e82b0f19018b83900b62efecc17ff9bcee5b493a5d4fe0528ad7a",
+    "sourceMarkdown": "f2afc1cda24b1ba91cf39429e7836ead97c7a1c815b294b3d623bb20d8f5e7e6",
     "publicationPack": "a1a523d5938faea72be568b843ac3890e61cea6070b0cfa46acf22ad3d2fb974",
     "referencesBib": "c0bfcaf7bfe37d6933c812e46352be8a95397eaa430a0f5bc94037600a53f654"
   },
@@ -13,8 +13,8 @@
     {
       "path": "main.tex",
       "source": "docs/paper/audrey-paper-v1.md",
-      "bytes": 122247,
-      "sha256": "b0122e625380ad9a6aff78e2acbd5984b558c054962f8774f77fee70e6588d06"
+      "bytes": 122667,
+      "sha256": "e3ee98ea8c523e8f394b8fbbc73e206f0d6126b7349df5669206d4d48d9feea6"
     },
     {
       "path": "references.bib",

package/docs/paper/output/arxiv/main.tex

@@ -385,7 +385,7 @@ The deterministic demo, \texttt{audrey demo --scenario repeated-failure}, constr
 
 The current paper version has two implemented empirical anchors. First, \texttt{benchmarks/snapshots/perf-0.22.2.json} reports canonical local performance under the mock-provider methodology: generated on 2026-05-01 from git SHA \texttt{e2e821b}, using mock 64-dimensional in-process embeddings, hybrid recall limit 5, and corpus sizes 100, 1,000, and 5,000 on Node 25.5.0 with a 24-core Ryzen 9 7900X3D and 62.9 GB RAM (Ledger: E20). Under that methodology, hybrid recall p95 is 1.82 ms, 2.364 ms, and 3.417 ms for those three sizes, and encode p95 is 0.589 ms, 2.147 ms, and 1.838 ms (Ledger: E21-E22).
 
-Second, \texttt{bench:memory:check} is wired into the release gate and enforces retrieval/lifecycle benchmark guardrails against weak local baselines (Ledger: E23). The current checked-in output reports a 2026-05-08 mock-provider run in which Audrey scores 100\% with 100\% pass rate, while the strongest listed local baselines score 41.67\% with 25\% pass rate in that output (Ledger: E24). These numbers support regression-gate honesty; they do not replace GuardBench.
+Second, \texttt{bench:memory:check} is wired into the release gate as an **internal regression suite**, not a competitive benchmark. It exists to catch retrieval/lifecycle regressions in Audrey itself. The suite includes hand-tuned weak local baselines (vector-only, keyword-plus-recency, recent-window) whose role is to anchor a relative pass margin — they are not stand-ins for production memory systems, and their scores should not be cited as comparative claims about any external system (Ledger: E23). The current checked-in output reports a 2026-05-08 mock-provider run in which Audrey scores 100\% with 100\% pass rate and the listed local stub baselines score in the 25-42\% range in that same run (Ledger: E24). These numbers support regression-gate honesty inside this repository; they are not cross-system results and they do not replace GuardBench.
 
 The README benchmark table currently differs from the canonical JSON snapshot, so the paper quotes only the JSON snapshot and tracks the README correction as a follow-up (Ledger: E28).
 
@@ -729,13 +729,13 @@ These numbers measure Audrey's local call path under an in-process mock embeddin
 
 \subsection{Behavioral Regression Result}
 
-The current \texttt{benchmarks/output/summary.json} was generated on 2026-05-13T08:33:24.917Z with command \texttt{node benchmarks/run.js --provider mock --dimensions 64} (Ledger: E24). It reports:
+The current \texttt{benchmarks/output/summary.json} was generated on 2026-05-29T03:45:32.997Z with command \texttt{node benchmarks/run.js --provider mock --dimensions 64} (Ledger: E24). It reports:
 
 \begin{verbatim}
 | System | Score Percent | Pass Rate | Average Duration Ms |
-| Audrey | 100 | 100 | 15.083333333333334 |
-| Vector Only | 41.66666666666667 | 25 | 0.25 |
-| Keyword + Recency | 41.66666666666667 | 25 | 0.5 |
+| Audrey | 100 | 100 | 15.416666666666666 |
+| Vector Only | 41.66666666666667 | 25 | 0.3333333333333333 |
+| Keyword + Recency | 41.66666666666667 | 25 | 0.6666666666666666 |
 | Recent Window | 37.5 | 25 | 0 |
 \end{verbatim}
 
@@ -759,7 +759,7 @@ It reports local adapters only, not external-system comparisons (Ledger: E46):
 | Evidence recall | 100% |
 | Redaction leaks | 0 |
 | Recall-degradation detection | 100% |
-| Guard latency p50 / p95 | 3.214 ms / 21.395 ms |
+| Guard latency p50 / p95 | 2.916 ms / 21.17 ms |
 | Published artifact raw-secret leaks | 0 |
 | Audrey Guard decision accuracy | 100% |
 | No-memory decision accuracy | 10% |

package/docs/paper/output/arxiv-compile-report.json

@@ -1,13 +1,13 @@
 {
   "schemaVersion": "1.0.0",
   "suite": "Audrey arXiv compile check",
-  "generatedAt": "2026-05-13T23:33:57.318Z",
+  "generatedAt": "2026-05-29T03:45:42.412Z",
   "source": {
     "sourceDir": "docs/paper/output/arxiv",
     "manifest": "docs/paper/output/arxiv/arxiv-manifest.json",
-    "manifestSha256": "386121a5f6234f56a370c5c54bde4ff12fc462921e7e5770d6968a4710271df8",
+    "manifestSha256": "6364c368755a4188d5b2deea9e3fe80201f5fa55658e6ae570141e54c5293bc6",
     "mainTex": "docs/paper/output/arxiv/main.tex",
-    "mainTexSha256": "b0122e625380ad9a6aff78e2acbd5984b558c054962f8774f77fee70e6588d06",
+    "mainTexSha256": "e3ee98ea8c523e8f394b8fbbc73e206f0d6126b7349df5669206d4d48d9feea6",
     "referencesBib": "docs/paper/output/arxiv/references.bib",
     "referencesBibSha256": "c0bfcaf7bfe37d6933c812e46352be8a95397eaa430a0f5bc94037600a53f654"
   },

package/docs/paper/output/submission-bundle/README.md

@@ -52,7 +52,7 @@ npx audrey guard --tool Bash "npm run deploy"
 Expected first-run shape:
 
 ```text
-Audrey Doctor v1.0.0
+Audrey Doctor v1.0.2
 Store health: not initialized
 Verdict: ready
 ```
@@ -94,7 +94,7 @@ and writes a timestamped backup before changing a non-empty file. The generated
 and `PostToolUseFailure` hooks record redacted tool traces. Verify the active
 hook set inside Claude Code with `/hooks`.
 
-All local MCP paths default to local embeddings and one shared SQLite-backed memory directory. Use `AUDREY_DATA_DIR` to isolate projects, tenants, or host identities.
+All local MCP paths default to local embeddings and one shared SQLite-backed memory directory. **Set a distinct `AUDREY_DATA_DIR` per tenant, agent identity, or concurrent host.** SQLite uses WAL mode without an advisory lock, so two processes sharing a directory will contend on writes. Isolation is a hard requirement for multi-agent setups, not a recommendation.
 
 Installer-generated host config does not include provider API keys by default. Prefer setting `ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, `GOOGLE_API_KEY`, or `GEMINI_API_KEY` in the host runtime environment; use `npx audrey install --include-secrets` only if you explicitly accept argv/config exposure.
 
@@ -296,10 +296,23 @@ output shapes are validated by JSON schemas under `benchmarks/schemas/`.
 
 Latest local result in this checkout: 10/10 scenarios passed, 100% prevention
 rate, 0% false-block rate, 0 raw secret leaks, 0 published artifact leaks in
-the raw-secret sweep, and 3.214ms / 21.395ms
-p50/p95 guard latency under the mock-provider methodology. Local baseline
-decision accuracy was: no-memory 10%, recent-window 60%, vector-only 40%, and
-FTS-only 10%; none passed the full GuardBench decision-plus-evidence contract.
+the raw-secret sweep, and 2.916ms / 21.17ms
+p50/p95 guard latency under the mock-provider methodology.
+
+**Methodology caveats, on purpose.** All numbers above are produced against
+the in-process mock 64-dim embedding provider documented in the run's
+`provenance` block. They characterize Audrey's controller and SQLite path,
+not real-provider end-to-end latency or production false-positive rates. The
+100% prevention rate is over the 5 GuardBench scenarios that expect a
+`block` decision (the suite is 10 scenarios total, mixed across allow / warn
+/ block). Local baseline decision accuracy was: no-memory 10%, recent-window
+60%, vector-only 40%, and FTS-only 10%; none of the local baselines passed
+the GuardBench decision-plus-evidence contract, which since v1.0.1 requires
+the correct decision plus at least one returned evidence id for `block` /
+`warn` scenarios (no longer Audrey-specific lineage phrasing — see
+`CHANGELOG.md#101---2026-05-15`). External-system numbers for Mem0 and Zep
+are explicitly out of scope for this Stage-A artifact; live credentialed
+runs land in a v2 paper after raw evidence bundles publish.
 
 ```bash
 npm run bench:guard
@@ -517,8 +530,19 @@ The Node sidecar defaults to `127.0.0.1:7437`. The Docker image intentionally bi
 
 ## Development
 
+Developer setup runs from source, not from the published tarball, so `npm run build` is required before any CLI subcommand resolves:
+
 ```bash
 npm ci
+npm run build
+npm run lint     # ESLint (type-checked typescript-eslint); CI requires it clean
+npm run format   # Prettier; use `npm run format:check` to verify without writing
+npm test
+```
+
+Once built, the `Quick Start` commands work against the local `dist/` output. Code style and types are enforced: `npm run lint` and `npm run format:check` run in CI (Ubuntu + Windows) and in every release gate, so the baseline cannot regress. The full release gate runs everything CI runs:
+
+```bash
 npm run release:gate
 python -m unittest discover -s python/tests -v
 npm run python:release:check

package/docs/paper/output/submission-bundle/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json

@@ -1,7 +1,7 @@
 {
   "schemaVersion": "1.0.0",
   "suite": "GuardBench adapter self-test",
-  "generatedAt": "2026-05-13T23:33:55.959Z",
+  "generatedAt": "2026-05-29T03:45:40.969Z",
   "ok": true,
   "adapter": {
     "name": "Example Allow Adapter",
@@ -15,21 +15,21 @@
     "requestedAdapter": "Example Allow Adapter",
     "scenarios": 10,
     "expectedScenarios": 10,
-    "fullContractPassRate": 0,
+    "fullContractPassRate": 0.1,
     "decisionAccuracy": 0.1,
     "redactionLeaks": 0,
     "failures": []
   },
   "score": {
     "scenarios": 10,
-    "fullContractPassRate": 0,
+    "fullContractPassRate": 0.1,
     "decisionAccuracy": 0.1,
-    "evidenceRecall": 0,
+    "evidenceRecall": 0.1,
     "redactionLeaks": 0,
     "latency": {
-      "p50Ms": 0.012,
-      "p95Ms": 0.049,
-      "maxMs": 0.049
+      "p50Ms": 0.01,
+      "p95Ms": 0.043,
+      "maxMs": 0.043
     }
   },
   "contract": {

package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-dry-run.json

@@ -1,7 +1,7 @@
 {
   "schemaVersion": "1.0.0",
   "suite": "GuardBench external adapter dry-run matrix",
-  "generatedAt": "2026-05-13T23:33:56.533Z",
+  "generatedAt": "2026-05-29T03:45:41.522Z",
   "ok": true,
   "registry": "benchmarks/adapters/registry.json",
   "outRoot": "benchmarks/output/external",

package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-evidence.json

@@ -1,7 +1,7 @@
 {
   "schemaVersion": "1.0.0",
   "suite": "GuardBench external evidence verification",
-  "generatedAt": "2026-05-13T23:33:56.821Z",
+  "generatedAt": "2026-05-29T03:45:41.794Z",
   "ok": true,
   "allowPending": true,
   "registry": "benchmarks/adapters/registry.json",

package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-conformance-card.json

@@ -1,7 +1,7 @@
 {
   "schemaVersion": "1.0.0",
   "suite": "GuardBench conformance card",
-  "generatedAt": "2026-05-13T23:33:51.583Z",
+  "generatedAt": "2026-05-29T03:45:36.958Z",
   "sourceDir": "benchmarks/output",
   "manifestVersion": "0.2.0",
   "suiteId": "guardbench-local-comparative",
@@ -25,9 +25,9 @@
     "evidenceRecall": 1,
     "redactionLeaks": 0,
     "latency": {
-      "p50Ms": 3.097,
-      "p95Ms": 29.711,
-      "maxMs": 29.711
+      "p50Ms": 2.916,
+      "p95Ms": 21.17,
+      "maxMs": 21.17
     }
   },
   "conformance": {
@@ -39,21 +39,21 @@
   "integrity": {
     "artifactHashes": {
       "guardbench-manifest.json": "57636ce19fdaa6e50fc3fc961d9e499a9f43632f588c713a9fefe8e8a6fa724c",
-      "guardbench-summary.json": "2a6d5ee83cce2502135fb0442ef8cd3f2679fdc38c84207612c22a800a7a113a",
-      "guardbench-raw.json": "c5b9c68cf946478fbfba617f17717e05ea3e01301089de19153d59e77e674bc6"
+      "guardbench-summary.json": "e8669cd6c80dc3dc849b3c4fcc473ea706eb3a760bced69682d0dc2396b2e233",
+      "guardbench-raw.json": "15b39fd1a65709a89455fbfcaf815daf364b204fa526d5065cc12fcaed281d28"
     },
     "externalRunMetadataHash": null
   },
   "provenance": {
-    "generatedAt": "2026-05-13T23:33:51.221Z",
-    "gitSha": "970752172441967c3ede79562eca69b08efb1f12",
+    "generatedAt": "2026-05-29T03:45:36.607Z",
+    "gitSha": "ceed2f51b615175c8bb412b96b5e5a501561189f",
     "gitDirty": false,
-    "node": "v24.14.1",
-    "v8": "13.6.233.17-node.44",
+    "node": "v24.16.0",
+    "v8": "13.6.233.17-node.49",
     "platform": "linux",
     "arch": "x64",
-    "osRelease": "6.17.0-1010-azure",
-    "cpuModel": "AMD EPYC 7763 64-Core Processor",
+    "osRelease": "6.17.0-1015-azure",
+    "cpuModel": "AMD EPYC 9V74 80-Core Processor",
     "cpuCount": 4,
     "totalMemoryGb": 15.61,
     "embeddingProvider": "mock",