@anuma/sdk 10.5.0 → 10.5.2

package/dist/react/index.cjs

@@ -21591,6 +21591,7 @@ init_logger();
 init_redactor();
 
 // src/lib/memory/entitySalience.ts
+init_patterns();
 var GENERIC_ENTITY_NAMES = /* @__PURE__ */ new Set([
   "today",
   "tomorrow",
@@ -21736,6 +21737,125 @@ var WEAK_MODIFIERS = /* @__PURE__ */ new Set([
   "and",
   "or"
 ]);
+var MAX_ENTITY_NAME_LENGTH = 120;
+var MAX_ENTITY_NAME_TOKENS = 11;
+var PREFIXED_HEX = /^0x[0-9a-f]{16,}$/i;
+var BARE_HEX = /^[0-9a-f]{32,}$/i;
+var BASE58_STRING = /^[1-9A-HJ-NP-Za-km-z]{32,88}$/;
+var UUID = /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i;
+var TRUNCATED_ID = /^[0-9a-z]{3,}(?:…|\.{2,3})[0-9a-z]{3,}$/i;
+var KEY_SEGMENT_MIN_LENGTH = 16;
+var LONG_DIGIT_RUN = /^\d{6,}$/;
+var CURRENCY_MARKED = /[$€£¥₹₿]\d|\d[$€£¥₹₿]/;
+var NUMERIC_TOKEN = /^[~≈+\-−]?(?:us\$|[$€£¥₹₿])?\d(?:[\d,.]*\d)?(?:k|mm|m|bn|b|t|x|%|[$€£¥₹₿])?$/i;
+var PLAIN_YEAR = /^(?:1\d{3}|20\d{2})$/;
+var SUFFIXED_NAME = /^\d*[1-9][kmb]$/i;
+var BARE_INTEGER = /^\d+$/;
+var TICKER = /^\$?[A-Z]{2,6}$/;
+var DETACHED_CURRENCY_SYMBOL = /(^|\s)([$€£¥₹₿])\s+(?=\d)|(\d)\s+([$€£¥₹₿])(?=\s|$)/gu;
+var LEADING_PUNCTUATION = /^[^\p{L}\p{N}$€£¥₹₿~≈+\-−]+/u;
+var TRAILING_KEPT_CHARACTER = /[\p{L}\p{N}$€£¥₹₿%]/u;
+var LEADING_CURRENCY_CODES = /* @__PURE__ */ new Set(["usd", "eur", "gbp", "jpy", "cad", "aud", "chf", "cny"]);
+var QUANTITY_UNIT_WORDS = /* @__PURE__ */ new Set([
+  "usd",
+  "usdc",
+  "usdt",
+  "eur",
+  "gbp",
+  "sol",
+  "eth",
+  "btc",
+  "dollar",
+  "dollars",
+  "euro",
+  "euros",
+  "buck",
+  "bucks",
+  "k",
+  "thousand",
+  "million",
+  "billion",
+  "trillion",
+  "share",
+  "shares",
+  "token",
+  "tokens",
+  "coin",
+  "coins",
+  "unit",
+  "units",
+  "position",
+  "balance",
+  "stake",
+  "holdings"
+]);
+var AMOUNT_WORDS = /* @__PURE__ */ new Set([...QUANTITY_UNIT_WORDS, "percent", "of"]);
+var NAME_PII_CATEGORIES = /* @__PURE__ */ new Set(["EMAIL", "PHONE", "SSN", "CREDIT_CARD"]);
+var NAME_PII_PATTERNS = PII_PATTERNS.filter(
+  (pattern) => NAME_PII_CATEGORIES.has(pattern.category) && pattern.context === void 0
+);
+var KEPT_PLACEHOLDER_CATEGORIES = /* @__PURE__ */ new Set(["PERSON", "LOCATION", "ORG"]);
+var PLACEHOLDER_TOKEN = /\b([A-Za-z]+(?:_[A-Za-z]+)*)_\d+\b/g;
+function trimPunctuation(token) {
+  const chars = Array.from(token.replace(LEADING_PUNCTUATION, ""));
+  let end = chars.length;
+  while (end > 0 && !TRAILING_KEPT_CHARACTER.test(chars[end - 1])) end--;
+  return chars.slice(0, end).join("");
+}
+function hasLetterAndDigit(text16) {
+  return /\p{L}/u.test(text16) && /\p{N}/u.test(text16);
+}
+function isIdentifierToken(token) {
+  if (PREFIXED_HEX.test(token) || BARE_HEX.test(token) || BASE58_STRING.test(token)) return true;
+  if (UUID.test(token) || LONG_DIGIT_RUN.test(token)) return true;
+  if (TRUNCATED_ID.test(token) && hasLetterAndDigit(token)) return true;
+  return token.split(/[-_.]/).some((segment) => segment.length >= KEY_SEGMENT_MIN_LENGTH && hasLetterAndDigit(segment));
+}
+function containsStructuredPii(name) {
+  return NAME_PII_PATTERNS.some(
+    (pattern) => (name.match(pattern.regex) ?? []).some((match) => pattern.validate?.(match) ?? true)
+  );
+}
+function isAmountUnitToken(token) {
+  return NUMERIC_TOKEN.test(token) || TICKER.test(token) || AMOUNT_WORDS.has(token.toLowerCase());
+}
+function isAmountWordOrNumber(token) {
+  return NUMERIC_TOKEN.test(token) || AMOUNT_WORDS.has(token.toLowerCase());
+}
+function isAmount(tokens) {
+  const start = LEADING_CURRENCY_CODES.has(tokens[0]?.toLowerCase() ?? "") ? 1 : 0;
+  const first = tokens[start];
+  if (first === void 0 || !NUMERIC_TOKEN.test(first)) return false;
+  const units = tokens.slice(start + 1);
+  if (start === 0 && units.length === 0)
+    return !PLAIN_YEAR.test(first) && !SUFFIXED_NAME.test(first);
+  if (!units.every(isAmountUnitToken)) return false;
+  return start === 1 || !BARE_INTEGER.test(first) || units.some(isAmountWordOrNumber);
+}
+function hasQuantityPhrase(tokens) {
+  return tokens.some(
+    (token, i) => NUMERIC_TOKEN.test(token) && QUANTITY_UNIT_WORDS.has(tokens[i + 1]?.toLowerCase() ?? "")
+  );
+}
+function isIdentifierOrAmountEntityName(name) {
+  const trimmed = name.trim();
+  if (trimmed.length > MAX_ENTITY_NAME_LENGTH) return true;
+  if (containsStructuredPii(trimmed)) return true;
+  const tokens = trimmed.replace(DETACHED_CURRENCY_SYMBOL, "$1$2$3$4").split(/\s+/).map(trimPunctuation).filter((token) => token.length > 0);
+  if (tokens.length === 0) return false;
+  if (tokens.length > MAX_ENTITY_NAME_TOKENS) return true;
+  if (tokens.some(isIdentifierToken)) return true;
+  if (tokens.some((token) => CURRENCY_MARKED.test(token))) return true;
+  return isAmount(tokens) || hasQuantityPhrase(tokens);
+}
+function restoreEntityName(name, redactor) {
+  for (const [token, category] of name.matchAll(PLACEHOLDER_TOKEN)) {
+    if (KEPT_PLACEHOLDER_CATEGORIES.has(category.toUpperCase())) continue;
+    if (redactor.restoreForStorage(token).text !== token) return null;
+  }
+  const restored = redactor.restoreForStorage(name);
+  return restored.unresolved || isIdentifierOrAmountEntityName(restored.text) ? null : restored.text;
+}
 
 // src/lib/memory/injectionClassifier.ts
 init_logger();
@@ -23589,12 +23709,16 @@ async function prepareVaultCandidates(query, vaultCtx, embeddingOptions, cache,
     const currentModel = embeddingOptions.model ?? DEFAULT_API_EMBEDDING_MODEL;
     const uncachedTexts = [];
     const uncachedIndices = [];
+    const resolvedVectors = /* @__PURE__ */ new Map();
     let staleReembedCount = 0;
     for (let i = 0; queryEmbedding.length > 0 && i < memories.length; i++) {
       const content = memories[i].content;
       const memoryId = memories[i].uniqueId;
       const cached = cachedRowVector(cache, memoryId, memories[i].updatedAt, content);
-      if (cached && cached.length === queryEmbedding.length) continue;
+      if (cached && cached.length === queryEmbedding.length) {
+        resolvedVectors.set(memoryId, cached);
+        continue;
+      }
       if (cached) cache.delete(memoryId);
       const storedModel = memories[i].embeddingModel;
       const modelCompatible = (storedModel ?? currentModel) === currentModel;
@@ -23602,13 +23726,9 @@ async function prepareVaultCandidates(query, vaultCtx, embeddingOptions, cache,
         try {
           const parsed = JSON.parse(memories[i].embedding);
           if (Array.isArray(parsed) && parsed.length === queryEmbedding.length) {
-            cacheRowVector(
-              cache,
-              memoryId,
-              Float32Array.from(parsed),
-              memories[i].updatedAt,
-              content
-            );
+            const vector2 = Float32Array.from(parsed);
+            resolvedVectors.set(memoryId, vector2);
+            cacheRowVector(cache, memoryId, vector2, memories[i].updatedAt, content);
             continue;
           }
         } catch {
@@ -23641,20 +23761,17 @@ async function prepareVaultCandidates(query, vaultCtx, embeddingOptions, cache,
       }
       if (newEmbeddings) {
         for (let j = 0; j < uncachedTexts.length; j++) {
-          cacheRowVector(
-            cache,
-            memories[uncachedIndices[j]].uniqueId,
-            Float32Array.from(newEmbeddings[j]),
-            memories[uncachedIndices[j]].updatedAt,
-            memories[uncachedIndices[j]].content
-          );
+          const memory = memories[uncachedIndices[j]];
+          const vector2 = Float32Array.from(newEmbeddings[j]);
+          resolvedVectors.set(memory.uniqueId, vector2);
+          cacheRowVector(cache, memory.uniqueId, vector2, memory.updatedAt, memory.content);
         }
       }
     }
     embeddedItems = memories.map((m) => ({
       id: m.uniqueId,
       content: m.content,
-      embedding: cache.get(m.uniqueId) ?? [],
+      embedding: resolvedVectors.get(m.uniqueId) ?? [],
       createdAt: m.createdAt,
       updatedAt: m.updatedAt,
       proofCount: m.proofCount,
@@ -24548,8 +24665,8 @@ function restoreCandidates(candidates, redactor, ownNames) {
         ...c,
         content: content.text,
         entities: c.entities.flatMap((e) => {
-          const restored = redactor.restoreForStorage(e.name);
-          return restored.unresolved ? [] : [{ ...e, name: restored.text }];
+          const name = restoreEntityName(e.name, redactor);
+          return name === null ? [] : [{ ...e, name }];
         })
       },
       unresolved: content.unresolved
@@ -24817,18 +24934,21 @@ function validateCandidates(parsed, validIds, userIds, fallbackSourceId, ownName
   }
   return out;
 }
+function isAcceptedEntityName(name) {
+  return name.length > 0 && !isGenericEntityName(name) && !isIdentifierOrAmountEntityName(name);
+}
 function parseEntities(raw) {
   const out = [];
   for (const e of raw) {
     if (typeof e === "string") {
       const name2 = e.trim();
-      if (name2.length > 0 && !isGenericEntityName(name2)) out.push({ name: name2 });
+      if (isAcceptedEntityName(name2)) out.push({ name: name2 });
       continue;
     }
     if (typeof e !== "object" || e === null) continue;
     const obj = e;
     const name = typeof obj.name === "string" ? obj.name.trim() : "";
-    if (name.length === 0 || isGenericEntityName(name)) continue;
+    if (!isAcceptedEntityName(name)) continue;
     const kind = typeof obj.kind === "string" && ENTITY_KINDS.includes(obj.kind) ? obj.kind : void 0;
     const weight = typeof obj.salience === "string" && Object.keys(TOPIC_SALIENCE_WEIGHTS).includes(obj.salience) ? TOPIC_SALIENCE_WEIGHTS[obj.salience] : void 0;
     out.push({
@@ -27018,6 +27138,7 @@ Each memory is a short statement about the user. For each memory, list the NAMED
 Include only NAMED entities, skip generic/common nouns. A topic must identify a specific thing the user could recognize by name; if it would be equally true of thousands of unrelated memories, leave it out:
 - SKIP bare generic nouns even when the memory capitalizes them or a calendar entry is titled that way \u2014 "Home", "Work", "Meeting", "Lunch", "Birthday", "Appointment", "Trip", "Today", "Friday", "March" are NOT topics.
 - KEEP the specific thing instead, when the memory names one: "Blue Bottle on Valencia" over "cafe", "Chicago Marathon" over "event", "Hollowpoint Labs" over "work".
+- SKIP identifiers and numbers \u2014 addresses, hashes, keys, account or order numbers, balances, prices, amounts and position sizes are NOT topics. Use the general subject instead: "Solana" or "trading" rather than "3.5 SOL" or an address.
 - Prefer 0 entities over a weak one. A memory whose only candidate is a generic noun has NO entities.
 ${ENTITY_KIND_GUIDELINES}
 
@@ -27123,8 +27244,8 @@ List each memory's named entities.`,
 }
 function restoreEntities(entities, redactor) {
   return entities.flatMap((e) => {
-    const restored = redactor.restoreForStorage(e.name);
-    return restored.unresolved ? [] : [{ ...e, name: restored.text }];
+    const name = restoreEntityName(e.name, redactor);
+    return name === null ? [] : [{ ...e, name }];
   });
 }
 function parseTopicResponse(parsed, validIds) {

package/dist/react/index.mjs

@@ -1,5 +1,6 @@
 import {
   INTERNAL_FLOW_MARKER,
+  PII_PATTERNS,
   PiiRedactor,
   callPortalJsonCompletion,
   consoleLogger,
@@ -19,7 +20,7 @@ import {
   taskTypeHeader,
   validateEndpointOverride,
   withInternalFlowMarker
-} from "./chunk-6ANL34M3.mjs";
+} from "./chunk-6QCUNXU6.mjs";
 import {
   BASE_URL,
   MCP_R2_DOMAIN,
@@ -19852,6 +19853,125 @@ var WEAK_MODIFIERS = /* @__PURE__ */ new Set([
   "and",
   "or"
 ]);
+var MAX_ENTITY_NAME_LENGTH = 120;
+var MAX_ENTITY_NAME_TOKENS = 11;
+var PREFIXED_HEX = /^0x[0-9a-f]{16,}$/i;
+var BARE_HEX = /^[0-9a-f]{32,}$/i;
+var BASE58_STRING = /^[1-9A-HJ-NP-Za-km-z]{32,88}$/;
+var UUID = /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i;
+var TRUNCATED_ID = /^[0-9a-z]{3,}(?:…|\.{2,3})[0-9a-z]{3,}$/i;
+var KEY_SEGMENT_MIN_LENGTH = 16;
+var LONG_DIGIT_RUN = /^\d{6,}$/;
+var CURRENCY_MARKED = /[$€£¥₹₿]\d|\d[$€£¥₹₿]/;
+var NUMERIC_TOKEN = /^[~≈+\-−]?(?:us\$|[$€£¥₹₿])?\d(?:[\d,.]*\d)?(?:k|mm|m|bn|b|t|x|%|[$€£¥₹₿])?$/i;
+var PLAIN_YEAR = /^(?:1\d{3}|20\d{2})$/;
+var SUFFIXED_NAME = /^\d*[1-9][kmb]$/i;
+var BARE_INTEGER = /^\d+$/;
+var TICKER = /^\$?[A-Z]{2,6}$/;
+var DETACHED_CURRENCY_SYMBOL = /(^|\s)([$€£¥₹₿])\s+(?=\d)|(\d)\s+([$€£¥₹₿])(?=\s|$)/gu;
+var LEADING_PUNCTUATION = /^[^\p{L}\p{N}$€£¥₹₿~≈+\-−]+/u;
+var TRAILING_KEPT_CHARACTER = /[\p{L}\p{N}$€£¥₹₿%]/u;
+var LEADING_CURRENCY_CODES = /* @__PURE__ */ new Set(["usd", "eur", "gbp", "jpy", "cad", "aud", "chf", "cny"]);
+var QUANTITY_UNIT_WORDS = /* @__PURE__ */ new Set([
+  "usd",
+  "usdc",
+  "usdt",
+  "eur",
+  "gbp",
+  "sol",
+  "eth",
+  "btc",
+  "dollar",
+  "dollars",
+  "euro",
+  "euros",
+  "buck",
+  "bucks",
+  "k",
+  "thousand",
+  "million",
+  "billion",
+  "trillion",
+  "share",
+  "shares",
+  "token",
+  "tokens",
+  "coin",
+  "coins",
+  "unit",
+  "units",
+  "position",
+  "balance",
+  "stake",
+  "holdings"
+]);
+var AMOUNT_WORDS = /* @__PURE__ */ new Set([...QUANTITY_UNIT_WORDS, "percent", "of"]);
+var NAME_PII_CATEGORIES = /* @__PURE__ */ new Set(["EMAIL", "PHONE", "SSN", "CREDIT_CARD"]);
+var NAME_PII_PATTERNS = PII_PATTERNS.filter(
+  (pattern) => NAME_PII_CATEGORIES.has(pattern.category) && pattern.context === void 0
+);
+var KEPT_PLACEHOLDER_CATEGORIES = /* @__PURE__ */ new Set(["PERSON", "LOCATION", "ORG"]);
+var PLACEHOLDER_TOKEN = /\b([A-Za-z]+(?:_[A-Za-z]+)*)_\d+\b/g;
+function trimPunctuation(token) {
+  const chars = Array.from(token.replace(LEADING_PUNCTUATION, ""));
+  let end = chars.length;
+  while (end > 0 && !TRAILING_KEPT_CHARACTER.test(chars[end - 1])) end--;
+  return chars.slice(0, end).join("");
+}
+function hasLetterAndDigit(text16) {
+  return /\p{L}/u.test(text16) && /\p{N}/u.test(text16);
+}
+function isIdentifierToken(token) {
+  if (PREFIXED_HEX.test(token) || BARE_HEX.test(token) || BASE58_STRING.test(token)) return true;
+  if (UUID.test(token) || LONG_DIGIT_RUN.test(token)) return true;
+  if (TRUNCATED_ID.test(token) && hasLetterAndDigit(token)) return true;
+  return token.split(/[-_.]/).some((segment) => segment.length >= KEY_SEGMENT_MIN_LENGTH && hasLetterAndDigit(segment));
+}
+function containsStructuredPii(name) {
+  return NAME_PII_PATTERNS.some(
+    (pattern) => (name.match(pattern.regex) ?? []).some((match) => pattern.validate?.(match) ?? true)
+  );
+}
+function isAmountUnitToken(token) {
+  return NUMERIC_TOKEN.test(token) || TICKER.test(token) || AMOUNT_WORDS.has(token.toLowerCase());
+}
+function isAmountWordOrNumber(token) {
+  return NUMERIC_TOKEN.test(token) || AMOUNT_WORDS.has(token.toLowerCase());
+}
+function isAmount(tokens) {
+  const start = LEADING_CURRENCY_CODES.has(tokens[0]?.toLowerCase() ?? "") ? 1 : 0;
+  const first = tokens[start];
+  if (first === void 0 || !NUMERIC_TOKEN.test(first)) return false;
+  const units = tokens.slice(start + 1);
+  if (start === 0 && units.length === 0)
+    return !PLAIN_YEAR.test(first) && !SUFFIXED_NAME.test(first);
+  if (!units.every(isAmountUnitToken)) return false;
+  return start === 1 || !BARE_INTEGER.test(first) || units.some(isAmountWordOrNumber);
+}
+function hasQuantityPhrase(tokens) {
+  return tokens.some(
+    (token, i) => NUMERIC_TOKEN.test(token) && QUANTITY_UNIT_WORDS.has(tokens[i + 1]?.toLowerCase() ?? "")
+  );
+}
+function isIdentifierOrAmountEntityName(name) {
+  const trimmed = name.trim();
+  if (trimmed.length > MAX_ENTITY_NAME_LENGTH) return true;
+  if (containsStructuredPii(trimmed)) return true;
+  const tokens = trimmed.replace(DETACHED_CURRENCY_SYMBOL, "$1$2$3$4").split(/\s+/).map(trimPunctuation).filter((token) => token.length > 0);
+  if (tokens.length === 0) return false;
+  if (tokens.length > MAX_ENTITY_NAME_TOKENS) return true;
+  if (tokens.some(isIdentifierToken)) return true;
+  if (tokens.some((token) => CURRENCY_MARKED.test(token))) return true;
+  return isAmount(tokens) || hasQuantityPhrase(tokens);
+}
+function restoreEntityName(name, redactor) {
+  for (const [token, category] of name.matchAll(PLACEHOLDER_TOKEN)) {
+    if (KEPT_PLACEHOLDER_CATEGORIES.has(category.toUpperCase())) continue;
+    if (redactor.restoreForStorage(token).text !== token) return null;
+  }
+  const restored = redactor.restoreForStorage(name);
+  return restored.unresolved || isIdentifierOrAmountEntityName(restored.text) ? null : restored.text;
+}
 
 // src/lib/memory/injectionClassifier.ts
 var DEFAULT_MODEL = "inclusionai/ling-2.6-flash";
@@ -21689,12 +21809,16 @@ async function prepareVaultCandidates(query, vaultCtx, embeddingOptions, cache,
     const currentModel = embeddingOptions.model ?? DEFAULT_API_EMBEDDING_MODEL;
     const uncachedTexts = [];
     const uncachedIndices = [];
+    const resolvedVectors = /* @__PURE__ */ new Map();
     let staleReembedCount = 0;
     for (let i = 0; queryEmbedding.length > 0 && i < memories.length; i++) {
       const content = memories[i].content;
       const memoryId = memories[i].uniqueId;
       const cached = cachedRowVector(cache, memoryId, memories[i].updatedAt, content);
-      if (cached && cached.length === queryEmbedding.length) continue;
+      if (cached && cached.length === queryEmbedding.length) {
+        resolvedVectors.set(memoryId, cached);
+        continue;
+      }
       if (cached) cache.delete(memoryId);
       const storedModel = memories[i].embeddingModel;
       const modelCompatible = (storedModel ?? currentModel) === currentModel;
@@ -21702,13 +21826,9 @@ async function prepareVaultCandidates(query, vaultCtx, embeddingOptions, cache,
         try {
           const parsed = JSON.parse(memories[i].embedding);
           if (Array.isArray(parsed) && parsed.length === queryEmbedding.length) {
-            cacheRowVector(
-              cache,
-              memoryId,
-              Float32Array.from(parsed),
-              memories[i].updatedAt,
-              content
-            );
+            const vector2 = Float32Array.from(parsed);
+            resolvedVectors.set(memoryId, vector2);
+            cacheRowVector(cache, memoryId, vector2, memories[i].updatedAt, content);
             continue;
           }
         } catch {
@@ -21741,20 +21861,17 @@ async function prepareVaultCandidates(query, vaultCtx, embeddingOptions, cache,
       }
       if (newEmbeddings) {
         for (let j = 0; j < uncachedTexts.length; j++) {
-          cacheRowVector(
-            cache,
-            memories[uncachedIndices[j]].uniqueId,
-            Float32Array.from(newEmbeddings[j]),
-            memories[uncachedIndices[j]].updatedAt,
-            memories[uncachedIndices[j]].content
-          );
+          const memory = memories[uncachedIndices[j]];
+          const vector2 = Float32Array.from(newEmbeddings[j]);
+          resolvedVectors.set(memory.uniqueId, vector2);
+          cacheRowVector(cache, memory.uniqueId, vector2, memory.updatedAt, memory.content);
         }
       }
     }
     embeddedItems = memories.map((m) => ({
       id: m.uniqueId,
       content: m.content,
-      embedding: cache.get(m.uniqueId) ?? [],
+      embedding: resolvedVectors.get(m.uniqueId) ?? [],
       createdAt: m.createdAt,
       updatedAt: m.updatedAt,
       proofCount: m.proofCount,
@@ -22354,7 +22471,7 @@ async function tryConsolidate(trimmed, ctx, options, prepared) {
     content: m.content,
     similarity: m.similarity
   }));
-  const { consolidateMemory: doConsolidate } = await import("./consolidate-JL4P6FXW.mjs");
+  const { consolidateMemory: doConsolidate } = await import("./consolidate-RASCVKD2.mjs");
   const decision = await doConsolidate(trimmed, candidates, consolidateOptions);
   if (decision.action === "create") return decision.fallbackReason ? null : { create: true };
   if (decision.action === "supersede" && decision.content) {
@@ -22647,8 +22764,8 @@ function restoreCandidates(candidates, redactor, ownNames) {
         ...c,
         content: content.text,
         entities: c.entities.flatMap((e) => {
-          const restored = redactor.restoreForStorage(e.name);
-          return restored.unresolved ? [] : [{ ...e, name: restored.text }];
+          const name = restoreEntityName(e.name, redactor);
+          return name === null ? [] : [{ ...e, name }];
         })
       },
       unresolved: content.unresolved
@@ -22916,18 +23033,21 @@ function validateCandidates(parsed, validIds, userIds, fallbackSourceId, ownName
   }
   return out;
 }
+function isAcceptedEntityName(name) {
+  return name.length > 0 && !isGenericEntityName(name) && !isIdentifierOrAmountEntityName(name);
+}
 function parseEntities(raw) {
   const out = [];
   for (const e of raw) {
     if (typeof e === "string") {
       const name2 = e.trim();
-      if (name2.length > 0 && !isGenericEntityName(name2)) out.push({ name: name2 });
+      if (isAcceptedEntityName(name2)) out.push({ name: name2 });
       continue;
     }
     if (typeof e !== "object" || e === null) continue;
     const obj = e;
     const name = typeof obj.name === "string" ? obj.name.trim() : "";
-    if (name.length === 0 || isGenericEntityName(name)) continue;
+    if (!isAcceptedEntityName(name)) continue;
     const kind = typeof obj.kind === "string" && ENTITY_KINDS.includes(obj.kind) ? obj.kind : void 0;
     const weight = typeof obj.salience === "string" && Object.keys(TOPIC_SALIENCE_WEIGHTS).includes(obj.salience) ? TOPIC_SALIENCE_WEIGHTS[obj.salience] : void 0;
     out.push({
@@ -25096,6 +25216,7 @@ Each memory is a short statement about the user. For each memory, list the NAMED
 Include only NAMED entities, skip generic/common nouns. A topic must identify a specific thing the user could recognize by name; if it would be equally true of thousands of unrelated memories, leave it out:
 - SKIP bare generic nouns even when the memory capitalizes them or a calendar entry is titled that way \u2014 "Home", "Work", "Meeting", "Lunch", "Birthday", "Appointment", "Trip", "Today", "Friday", "March" are NOT topics.
 - KEEP the specific thing instead, when the memory names one: "Blue Bottle on Valencia" over "cafe", "Chicago Marathon" over "event", "Hollowpoint Labs" over "work".
+- SKIP identifiers and numbers \u2014 addresses, hashes, keys, account or order numbers, balances, prices, amounts and position sizes are NOT topics. Use the general subject instead: "Solana" or "trading" rather than "3.5 SOL" or an address.
 - Prefer 0 entities over a weak one. A memory whose only candidate is a generic noun has NO entities.
 ${ENTITY_KIND_GUIDELINES}
 
@@ -25201,8 +25322,8 @@ List each memory's named entities.`,
 }
 function restoreEntities(entities, redactor) {
   return entities.flatMap((e) => {
-    const restored = redactor.restoreForStorage(e.name);
-    return restored.unresolved ? [] : [{ ...e, name: restored.text }];
+    const name = restoreEntityName(e.name, redactor);
+    return name === null ? [] : [{ ...e, name }];
   });
 }
 function parseTopicResponse(parsed, validIds) {

package/dist/server/{chunk-BIMEZT3O.mjs → chunk-JEY3L7WD.mjs}

@@ -1084,6 +1084,7 @@ function notifyConsolidationFallback(reason, onFallback, detail) {
 export {
   getLogger,
   taskTypeHeader,
+  PII_PATTERNS,
   resolvePiiRedactor,
   createStreamingDeAnonymizer,
   validateEndpointOverride,

package/dist/server/{consolidate-JZJYY3SZ.mjs → consolidate-5DFDLGSS.mjs}

@@ -4,7 +4,7 @@ import {
   notifyConsolidationFallback,
   requiresResponsesTransport,
   resolvePiiRedactor
-} from "./chunk-BIMEZT3O.mjs";
+} from "./chunk-JEY3L7WD.mjs";
 import "./chunk-3RPGXXPZ.mjs";
 
 // src/lib/memory/consolidate.ts