comisai 1.0.44 → 1.0.46

package/node_modules/@comis/memory/dist/sqlite-mental-model-store.d.ts

@@ -1,12 +1,12 @@
 /**
  * SqliteMentalModelStore: the SOLE adapter for the segregated
- * `MentalModelStorePort` (@comis/core, the v2.31 Mental Model doc store
- * generalized from the v2.26 Verified Learning WS2 / SKILL-01 procedural store).
+ * `MentalModelStorePort` (@comis/core) — the Mental Model doc store that holds
+ * learned skill/profile/topic docs.
  * It owns ALL the `mental_models` SQL — the idempotent `admit()` upsert (one row
  * per admitted doc), the scoped `get`/`list(scope, kind?)` reads, and the
  * `promote`/`demote`/`evict` lifecycle transitions.
  *
- * ## Idempotency (SKILL-01 / T-201-09)
+ * ## Idempotency
  *
  * `admit()` derives the row `id` as a deterministic sha256 hash of the UNIQUE
  * tuple `(tenant_id, agent_id, kind, topic_key, name)` in CODE before insert, AND
@@ -16,23 +16,23 @@
  * `UNIQUE (tenant_id, agent_id, kind, topic_key, name)` backstop catches it
  * regardless.
  *
- * ## Trust ceiling (SEC-01 / T-201-05) — the keystone
+ * ## Trust ceiling — the keystone
  *
  * A learned doc can NEVER be `system`. The store writes the LITERAL `'learned'`
  * for `trust_level` on EVERY admit (it never reads a caller-supplied trust), and
  * the DB `CHECK (trust_level IN ('learned'))` rejects any other value at insert
  * time — belt (code coercion) AND suspenders (DB constraint).
  *
- * ## Isolation is the load-bearing security boundary (SEC-01 / T-201-06)
+ * ## Isolation is the load-bearing security boundary
  *
  * Comis runs many agents in one DB. EVERY statement filters on
  * `(tenant_id, agent_id)` — parameterized — and the table keys/indexes lead on
  * those columns, so a doc under one (tenant, agent) is NEVER visible to a read
  * under another even when `name` is byte-identical. An UNRESOLVED
  * `(tenant, agent)` scope (empty id) fails-closed with `err(...)` — it NEVER
- * widens to a shared/global pool (the `get_current_schema()` leak vector,
- * T-201-07). The optional `kind` filter on `list()` is an ADDITIONAL `AND`, never
- * a replacement for the scope filter.
+ * widens to a shared/global pool (the `get_current_schema()` leak vector). The
+ * optional `kind` filter on `list()` is an ADDITIONAL `AND`, never a replacement
+ * for the scope filter.
  *
  * ## Soft lifecycle (never hard-delete)
  *
@@ -46,7 +46,7 @@
  * The persisted JSON columns (`source_traj_ids` etc.) are parsed with a
  * graceful-degrade `safeParse` (corrupt JSON → empty list, never a throw). Logs
  * carry counts/ids + metadata only — never doc bodies/descriptions
- * (§2.7 / T-201-10).
+ * (AGENTS.md §2.7).
  *
  * @module
  */
@@ -70,7 +70,7 @@ export interface MentalModelStoreDeps {
  * handle. The handle's lifecycle (open/close, pragmas) is owned by the caller
  * (the memory adapter) — this factory neither opens nor closes it. Built
  * UNCONDITIONALLY (no model/IO cost, like every dormant store); the per-agent
- * enable flag gates the daemon-side `admit`/`get`/`list` call (Plan 07), not
+ * enable flag gates the daemon-side `admit`/`get`/`list` call, not
  * construction.
  */
 export declare function createSqliteMentalModelStore(deps: MentalModelStoreDeps): MentalModelStorePort;

package/node_modules/@comis/memory/dist/sqlite-mental-model-store.js

@@ -1,13 +1,13 @@
 // SPDX-License-Identifier: Apache-2.0
 /**
  * SqliteMentalModelStore: the SOLE adapter for the segregated
- * `MentalModelStorePort` (@comis/core, the v2.31 Mental Model doc store
- * generalized from the v2.26 Verified Learning WS2 / SKILL-01 procedural store).
+ * `MentalModelStorePort` (@comis/core) — the Mental Model doc store that holds
+ * learned skill/profile/topic docs.
  * It owns ALL the `mental_models` SQL — the idempotent `admit()` upsert (one row
  * per admitted doc), the scoped `get`/`list(scope, kind?)` reads, and the
  * `promote`/`demote`/`evict` lifecycle transitions.
  *
- * ## Idempotency (SKILL-01 / T-201-09)
+ * ## Idempotency
  *
  * `admit()` derives the row `id` as a deterministic sha256 hash of the UNIQUE
  * tuple `(tenant_id, agent_id, kind, topic_key, name)` in CODE before insert, AND
@@ -17,23 +17,23 @@
  * `UNIQUE (tenant_id, agent_id, kind, topic_key, name)` backstop catches it
  * regardless.
  *
- * ## Trust ceiling (SEC-01 / T-201-05) — the keystone
+ * ## Trust ceiling — the keystone
  *
  * A learned doc can NEVER be `system`. The store writes the LITERAL `'learned'`
  * for `trust_level` on EVERY admit (it never reads a caller-supplied trust), and
  * the DB `CHECK (trust_level IN ('learned'))` rejects any other value at insert
  * time — belt (code coercion) AND suspenders (DB constraint).
  *
- * ## Isolation is the load-bearing security boundary (SEC-01 / T-201-06)
+ * ## Isolation is the load-bearing security boundary
  *
  * Comis runs many agents in one DB. EVERY statement filters on
  * `(tenant_id, agent_id)` — parameterized — and the table keys/indexes lead on
  * those columns, so a doc under one (tenant, agent) is NEVER visible to a read
  * under another even when `name` is byte-identical. An UNRESOLVED
  * `(tenant, agent)` scope (empty id) fails-closed with `err(...)` — it NEVER
- * widens to a shared/global pool (the `get_current_schema()` leak vector,
- * T-201-07). The optional `kind` filter on `list()` is an ADDITIONAL `AND`, never
- * a replacement for the scope filter.
+ * widens to a shared/global pool (the `get_current_schema()` leak vector). The
+ * optional `kind` filter on `list()` is an ADDITIONAL `AND`, never a replacement
+ * for the scope filter.
  *
  * ## Soft lifecycle (never hard-delete)
  *
@@ -47,7 +47,7 @@
  * The persisted JSON columns (`source_traj_ids` etc.) are parsed with a
  * graceful-degrade `safeParse` (corrupt JSON → empty list, never a throw). Logs
  * carry counts/ids + metadata only — never doc bodies/descriptions
- * (§2.7 / T-201-10).
+ * (AGENTS.md §2.7).
  *
  * @module
  */
@@ -74,10 +74,10 @@ function parseIdList(raw) {
         return [];
     }
 }
-// Lenient parser for the structured_body AST column (the v2.31 Reflection
+// Lenient parser for the structured_body AST column (the Reflection
 // section-list). The shape mirrors @comis/core's StructuredBody — a NULL column,
 // non-JSON text, or a payload that does not match the shape degrades to
-// `undefined` (NOT a throw — T-223-09: a corrupt AST row must not crash recall;
+// `undefined` (NOT a throw: a corrupt AST row must not crash recall;
 // the doc is treated as "no AST" and re-synthesized).
 const StructuredBodySchema = z.object({
     sections: z.array(z.object({ id: z.string(), heading: z.string(), body: z.string() })),
@@ -98,8 +98,27 @@ function parseStructuredBody(raw) {
         return undefined;
     }
 }
-// FOLD-01 (Phase 225): the canonical `mental_models.history` JSON shape — an
-// ordered (oldest-first) array of prior bodies, mirroring the 224
+/**
+ * Parse the nullable JSON-encoded `required_tools` column into the tool-NAME set,
+ * or `undefined` when the column is NULL (a user-intent skill / profile / topic) or
+ * holds corrupt JSON (degrade-to-absent, mirroring {@link parseStructuredBody} — a
+ * corrupt column must never throw a read). This is the read-side of the write-side
+ * bind: the surface uses `requiredTools !== undefined` as the procedure-doc
+ * discriminator, so an absent/corrupt column reads as "not a procedure doc".
+ */
+function parseRequiredTools(raw) {
+    if (raw === null)
+        return undefined;
+    try {
+        const parsed = StringArraySchema.safeParse(JSON.parse(raw));
+        return parsed.success ? parsed.data : undefined;
+    }
+    catch {
+        return undefined;
+    }
+}
+// The canonical `mental_models.history` JSON shape — an
+// ordered (oldest-first) array of prior bodies, mirroring the
 // `SqliteMemoryAdapter` `SupersedeHistorySchema` byte-for-byte
 // ({ previousContent, changedAt }). A strictObject so a malformed/legacy column
 // degrades to "absent" (→ a fresh array in supersede / undefined in get) instead
@@ -108,7 +127,7 @@ const HistorySchema = z.array(z.strictObject({ previousContent: z.string(), chan
 /**
  * Parse a nullable JSON-encoded `history` column into the typed prior-body array,
  * or `undefined` when the column is NULL / corrupt (degrade-to-absent, mirrors
- * `parseStructuredBody` / the 224 `parseHistoryColumn`). supersede() then starts a
+ * `parseStructuredBody` / the adapter's `parseHistoryColumn`). supersede() then starts a
  * fresh array, so a damaged column self-heals on the next correction.
  */
 function parseHistoryColumn(raw) {
@@ -138,6 +157,10 @@ function mentalModelId(s) {
 }
 /** Map a parsed `mental_models` row to the domain {@link MentalModel}. */
 function rowToMentalModel(row) {
+    // Read-side of the write-side bind: the surface discriminator. Spread in ONLY
+    // when the column parses to a value (NULL / corrupt ⇒ the field is absent), so a
+    // user-intent skill / profile / topic reads `requiredTools` undefined.
+    const requiredTools = parseRequiredTools(row.required_tools);
     return {
         id: row.id,
         name: row.name,
@@ -153,13 +176,15 @@ function rowToMentalModel(row) {
         confidence: row.confidence,
         mutating: row.mutating === 1,
         sourceTrajIds: parseIdList(row.source_traj_ids),
-        // The Reflection section-AST (REFLECT-04) — lenient parse so delta-ops read
+        // The Reflection section-AST — lenient parse so delta-ops read
         // the prior doc; undefined when the column is NULL or holds garbage (the
-        // doc is then treated as new — synthesize fresh, A6).
+        // doc is then treated as new — synthesize fresh).
         structuredBody: parseStructuredBody(row.structured_body),
-        // The FOLD-01 supersede trail (Phase 225) — lenient parse; undefined when the
-        // column is NULL (never superseded) or corrupt (degrade-to-absent, T-223-09).
+        // The supersede trail — lenient parse; undefined when the
+        // column is NULL (never superseded) or corrupt (degrade-to-absent).
         history: parseHistoryColumn(row.history),
+        // The procedure-doc tool footprint (advisory); absent for a user-intent skill.
+        ...(requiredTools !== undefined ? { requiredTools } : {}),
         createdAt: row.created_at,
     };
 }
@@ -168,7 +193,7 @@ function rowToMentalModel(row) {
  * handle. The handle's lifecycle (open/close, pragmas) is owned by the caller
  * (the memory adapter) — this factory neither opens nor closes it. Built
  * UNCONDITIONALLY (no model/IO cost, like every dormant store); the per-agent
- * enable flag gates the daemon-side `admit`/`get`/`list` call (Plan 07), not
+ * enable flag gates the daemon-side `admit`/`get`/`list` call, not
  * construction.
  */
 export function createSqliteMentalModelStore(deps) {
@@ -176,28 +201,33 @@ export function createSqliteMentalModelStore(deps) {
     // --- Prepared statements (parameterized; reused across calls) ---
     // Idempotent admit keyed on the deterministic id (= hash of the
     // (tenant_id, agent_id, kind, topic_key, name) UNIQUE tuple). A re-admit upserts
-    // the SAME row. trust_level is the LITERAL 'learned' (the SEC-01 keystone —
+    // the SAME row. trust_level is the LITERAL 'learned' (the trust-ceiling keystone —
     // never a bound caller value); all other columns are bound `?` params (never
     // string-built). kind/topic_key are bound (default 'skill'/'' at the call site).
-    // structured_body (the v2.31 Reflection section-AST, REFLECT-04) is bound as a
+    // structured_body (the Reflection section-AST) is bound as a
     // JSON `?` (NULL when the caller omits it) and updated in lockstep with body on
-    // conflict. `history` stays the DB default NULL (Phase 224 owns supersession).
-    // The dropped `scripts` column was the literal NULL — its removal is a no-op.
+    // conflict. required_tools / params_schema are the DETERMINISTIC advisory metadata
+    // the procedure run derives from the AUDITED descriptor — bound `?` (NULL for the
+    // user-intent skill path) and updated in lockstep with body too; NEVER the
+    // trust_level (that stays the 'learned' LITERAL — the trust keystone). `history`
+    // stays the DB default NULL on admit (supersede() owns the trail). The dropped
+    // `scripts` column was the literal NULL — its removal is a no-op (no learned code).
     const insertStmt = db.prepare("INSERT INTO mental_models " +
         "(id, tenant_id, agent_id, kind, topic_key, name, description, body, structured_body, required_tools, params_schema, " +
         " trust_level, state, proof_count, confidence, strength, source_traj_ids, validation_result, " +
         " mutating, pinned, validated_at, created_at, updated_at, evicted_at) " +
-        "VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, NULL, NULL, 'learned', 'candidate', ?, ?, ?, ?, NULL, ?, 0, NULL, ?, NULL, NULL) " +
+        "VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, 'learned', 'candidate', ?, ?, ?, ?, NULL, ?, 0, NULL, ?, NULL, NULL) " +
         "ON CONFLICT(id) DO UPDATE SET " +
         "description = excluded.description, body = excluded.body, " +
-        "structured_body = excluded.structured_body, proof_count = excluded.proof_count, " +
+        "structured_body = excluded.structured_body, required_tools = excluded.required_tools, " +
+        "params_schema = excluded.params_schema, proof_count = excluded.proof_count, " +
         "confidence = excluded.confidence, strength = excluded.strength, " +
         "source_traj_ids = excluded.source_traj_ids, mutating = excluded.mutating, " +
         // A re-admit of a previously-evicted doc resurrects it (clears evicted_at,
         // resets state to candidate) — the replay-stable path.
         "state = 'candidate', evicted_at = NULL, updated_at = excluded.created_at");
     // Scoped reads — the `tenant_id = ? AND agent_id = ?` filter is the
-    // load-bearing isolation boundary (SEC-01); every value is a bound `?` param.
+    // load-bearing isolation boundary; every value is a bound `?` param.
     // SELECT_COLS drops `scripts` and adds kind/topic_key/structured_body/history —
     // kept in lockstep with the MentalModelRowSchema strictObject (a drift throws).
     const SELECT_COLS = "id, name, description, kind, topic_key, trust_level, state, body, structured_body, history, " +
@@ -216,17 +246,17 @@ export function createSqliteMentalModelStore(deps) {
     // Lifecycle transitions — all scoped to (tenant, agent) AND id (bound params).
     // promote: proof_count bumps on EVERY call, but the candidate→active flip is
     // GATED on the caller's proof bar — `proof_count + 1 >= promoteAtProofCount`
-    // (D2 / T-202-04: a single attributed success must NOT mint an active
+    // (a single attributed success must NOT mint an active
     // doc). The threshold is the FIRST bound `?` (promote runs its own bind
     // path, not the shared runTransition); an already-active doc keeps bumping
     // proof_count but the `state = 'candidate'` guard means its state never moves.
     const promoteStmt = db.prepare("UPDATE mental_models SET proof_count = proof_count + 1, " +
         "state = CASE WHEN state = 'candidate' AND proof_count + 1 >= ? THEN 'active' ELSE state END, " +
         "updated_at = ? WHERE tenant_id = ? AND agent_id = ? AND id = ?");
-    // promoteByName keys on the NAME (not the re-derived id) — SKILL-04: a reflected
+    // promoteByName keys on the NAME (not the re-derived id) — a reflected
     // doc is admitted WITH a non-empty topicKey, so re-deriving the id with an assumed
     // `topicKey:''` MISSES it. The reflection job names a doc `skill-<full-topicKey>`
-    // (WR-01 — the FULL topicKey, not a 16-char truncation), so the name EMBEDS the
+    // (the FULL topicKey, not a 16-char truncation), so the name EMBEDS the
     // unique topic_key and is therefore unique per (tenant, agent, kind) — a name-keyed
     // UPDATE resolves the SAME single row get() does, with no truncation-collision risk.
     // Identical proof-bar CASE to promoteStmt; only the WHERE key differs (name vs id).
@@ -234,7 +264,7 @@ export function createSqliteMentalModelStore(deps) {
         "state = CASE WHEN state = 'candidate' AND proof_count + 1 >= ? THEN 'active' ELSE state END, " +
         "updated_at = ? WHERE tenant_id = ? AND agent_id = ? AND name = ?");
     // demote: step state back toward stale on a verified failure (soft, monotone).
-    // WR-06: the WHERE pins `state IN ('active','candidate')` — the ONLY states a demote
+    // The WHERE pins `state IN ('active','candidate')` — the ONLY states a demote
     // moves. A terminal-state row (stale/archived) therefore matches 0 rows, so the
     // `updated_at` rewrite never inflates `info.changes` into a phantom transition and
     // demoteByName's `changed` reflects a REAL state delta (not a no-op write). The CASE
@@ -243,8 +273,8 @@ export function createSqliteMentalModelStore(deps) {
         "state = CASE WHEN state = 'active' THEN 'stale' WHEN state = 'candidate' THEN 'stale' ELSE state END, " +
         "strength = CASE WHEN strength > 0 THEN strength - 1 ELSE strength END, " +
         "updated_at = ? WHERE tenant_id = ? AND agent_id = ? AND id = ? AND state IN ('active', 'candidate')");
-    // demoteByName keys on the NAME (not the re-derived id) — SKILL-04, the mirror of
-    // promoteByNameStmt. Same monotone state step + the WR-06 `state IN
+    // demoteByName keys on the NAME (not the re-derived id) — the mirror of
+    // promoteByNameStmt. Same monotone state step + the `state IN
     // ('active','candidate')` terminal-state guard; only the WHERE key differs.
     const demoteByNameStmt = db.prepare("UPDATE mental_models SET " +
         "state = CASE WHEN state = 'active' THEN 'stale' WHEN state = 'candidate' THEN 'stale' ELSE state END, " +
@@ -286,8 +316,13 @@ export function createSqliteMentalModelStore(deps) {
                 });
                 insertStmt.run(id, scope.tenantId, scope.agentId, kind, topicKey, input.name, input.description, input.body, 
                 // structured_body: JSON-stringify the AST, or NULL when omitted. Updated
-                // in lockstep with body on the idempotent ON CONFLICT upsert (REFLECT-04).
-                input.structuredBody !== undefined ? JSON.stringify(input.structuredBody) : null, input.proofCount, input.confidence, 
+                // in lockstep with body on the idempotent ON CONFLICT upsert.
+                input.structuredBody !== undefined ? JSON.stringify(input.structuredBody) : null, 
+                // required_tools / params_schema: the DETERMINISTIC advisory metadata (bound only on
+                // the procedure run — NULL for the user-intent skill path). required_tools is the
+                // JSON-encoded content-free tool-NAME set; params_schema is the fixed content-free
+                // string. NEVER the trust_level (the 'learned' LITERAL above is the trust keystone).
+                input.requiredTools !== undefined ? JSON.stringify([...input.requiredTools]) : null, input.paramsSchema !== undefined ? input.paramsSchema : null, input.proofCount, input.confidence, 
                 // strength seeds from confidence (a verified candidate enters with a
                 // positive strength budget that demote() draws down).
                 input.confidence, input.sourceTrajIds.length > 0 ? JSON.stringify([...input.sourceTrajIds]) : null, input.mutating ? 1 : 0, input.createdAt);
@@ -371,12 +406,10 @@ export function createSqliteMentalModelStore(deps) {
         },
         async promoteByName(name, scope, promoteAtProofCount) {
             // Resolve the row by `(tenant, agent, name)` — the SAME key get() resolves by.
-            // SKILL-04 fix: the reuse loop holds only the skill NAME (ATTR-01), and a
-            // reflected doc is admitted WITH a non-empty topicKey, so re-deriving the id
-            // with a hardcoded `topicKey:''` (the prior behavior) MISSED any reflected doc
-            // and `changed` was always false — the entire reflect→reuse→promote loop was
-            // silently dead on its real input. The reflection name embeds the FULL topicKey
-            // (`skill-<full-topicKey>`, WR-01), so the name is unique per (tenant, agent,
+            // The reuse loop holds only the skill NAME, and a reflected doc is admitted WITH a
+            // non-empty topicKey, so re-deriving the id with a hardcoded `topicKey:''` would MISS
+            // any reflected doc (and `changed` would always be false). The reflection name embeds
+            // the FULL topicKey (`skill-<full-topicKey>`), so the name is unique per (tenant, agent,
             // kind) — a name-keyed UPDATE finds the SINGLE row get() resolves by (no 16-char
             // truncation collision). Reports rows-changed so a 0-row write (an unknown/
             // evicted name) stays detectable.
@@ -398,11 +431,11 @@ export function createSqliteMentalModelStore(deps) {
             }
         },
         async demoteByName(name, scope) {
-            // Resolve the row by `(tenant, agent, name)` — the mirror of promoteByName's
-            // SKILL-04 fix. Re-deriving the id with `topicKey:''` missed any reflected doc
+            // Resolve the row by `(tenant, agent, name)` — the mirror of promoteByName.
+            // Re-deriving the id with `topicKey:''` would miss any reflected doc
             // (non-empty topicKey); the reflection name embeds the FULL topicKey
-            // (`skill-<full-topicKey>`, WR-01), so it is unique per (tenant, agent, kind).
-            // The WR-06 terminal-state guard lives in the statement, so an unknown/evicted/
+            // (`skill-<full-topicKey>`), so it is unique per (tenant, agent, kind).
+            // The terminal-state guard lives in the statement, so an unknown/evicted/
             // stale name yields 0 rows → changed:false (never a phantom transition).
             const rejected = rejectUnresolvedScope(scope);
             if (rejected)
@@ -422,19 +455,19 @@ export function createSqliteMentalModelStore(deps) {
             }
         },
         async supersede(input, scope, now) {
-            // FOLD-01 (Phase 225): a profile/topic CORRECTION → UPDATE body (+ structured_body
-            // when supplied) and APPEND the prior body to `history` ({previousContent,
-            // changedAt}); the row is UPDATEd, never DELETEd (deletion stays reserved for the
-            // evict() security path). Mirrors SqliteMemoryAdapter.supersede (224 FORGET-04).
+            // A profile/topic CORRECTION → UPDATE body (+ structured_body when supplied) and
+            // APPEND the prior body to `history` ({previousContent, changedAt}); the row is
+            // UPDATEd, never DELETEd (deletion stays reserved for the evict() security path).
+            // Mirrors SqliteMemoryAdapter.supersede.
             const rejected = rejectUnresolvedScope(scope);
             if (rejected)
                 return rejected;
             const startMs = systemNowMs();
-            // INV-5 redaction firewall on the untrusted correction, BEFORE the txn — a
+            // Redaction firewall on the untrusted correction, BEFORE the txn — a
             // CRITICAL classification (dangerous command / secret egress) is REJECTED and
             // never persisted. A `warn` is permitted: a learned doc's trust is the fixed
             // 'learned' (the supersede never touches trust_level), so only `critical` blocks
-            // (mirror 224 sqlite-memory-adapter.ts:556 — a learned body already passed
+            // (mirror sqlite-memory-adapter.ts:556 — a learned body already passed
             // validateLearnedDocBody in the engine; this is the store-side belt).
             const verdict = validateMemoryWrite(input.body);
             if (verdict.severity === "critical") {
@@ -449,7 +482,7 @@ export function createSqliteMentalModelStore(deps) {
                 return err(new Error("mental-model supersede: body failed redaction validation"));
             }
             try {
-                // Scoped statements — the (tenant, agent, name) filter is the SEC-01 isolation
+                // Scoped statements — the (tenant, agent, name) filter is the isolation
                 // boundary; every value a bound `?` (NEVER concatenated). The doc is keyed by
                 // name within scope (the same key get() resolves by), `evicted_at IS NULL` so a
                 // correction never silently revives a soft-evicted doc.
@@ -457,7 +490,7 @@ export function createSqliteMentalModelStore(deps) {
                     "WHERE tenant_id = ? AND agent_id = ? AND name = ? AND evicted_at IS NULL");
                 const updateBody = db.prepare("UPDATE mental_models SET body = ?, structured_body = ?, history = ?, updated_at = ? " +
                     "WHERE tenant_id = ? AND agent_id = ? AND name = ? AND evicted_at IS NULL");
-                // The revise unit — ONE synchronous transaction (mirror the 224 supersede).
+                // The revise unit — ONE synchronous transaction (mirror the adapter's supersede).
                 // better-sqlite3 auto-ROLLBACKs on ANY throw, so SELECT-incumbent →
                 // history-append → UPDATE is atomic; a parse fault THROWS → ROLLBACK (caught
                 // below → err). The decided branch is returned for the metadata log.
@@ -469,7 +502,7 @@ export function createSqliteMentalModelStore(deps) {
                         return "not-found";
                     // Append the prior body (oldest-first), then update. History is appended
                     // REGARDLESS of whether body changed — a correction is the durable signal
-                    // (mirror 224). The mental_models_au/_tri_au WHEN-guarded triggers re-sync the
+                    // (mirror the adapter's supersede). The mental_models_au/_tri_au WHEN-guarded triggers re-sync the
                     // FTS/trigram twins on the body UPDATE (schema-mental-models.ts:226,252) — no
                     // NEW trigger work.
                     const prior = [

package/node_modules/@comis/memory/dist/sqlite-outcome-store.d.ts

@@ -1,11 +1,11 @@
 /**
  * SqliteOutcomeStore: the SOLE adapter for the segregated `OutcomeSignalPort`
- * (@comis/core, v2.26 Verified Learning WS1). It owns ALL the `outcome_events`
+ * (@comis/core). It owns ALL the `outcome_events`
  * SQL — the idempotent `observe()` write (one raw observation per signal source),
  * the scoped `resolve()` read+fusion (precedence-first then confidence,
  * fail-closed `unknown`), and the age-based `prune()`.
  *
- * ## Idempotency (OUTCOME-01 / T-198-09)
+ * ## Idempotency
  *
  * `observe()` derives the row `id` as a deterministic sha256 hash of the UNIQUE
  * tuple `(tenant_id, agent_id, trajectory_id, source, observed_at)` in CODE
@@ -14,7 +14,7 @@
  * upsert the same primary key even if the row was deleted, and the `UNIQUE`
  * backstop catches it regardless.
  *
- * ## Isolation is the load-bearing security boundary (SEC-01 / T-198-05)
+ * ## Isolation is the load-bearing security boundary
  *
  * Comis runs many agents in one DB. EVERY statement (both `observe` and
  * `resolve`) filters on `(tenant_id, agent_id)` — parameterized — and the table
@@ -22,7 +22,7 @@
  * visible to a read under another even when `trajectory_id` is byte-identical. An
  * UNRESOLVED `(tenant, agent)` scope (empty id) on `resolve()` fails-closed with
  * `err(...)` — it NEVER widens to a shared/global pool (the hindsight
- * `get_current_schema()` leak vector, design §9).
+ * `get_current_schema()` leak vector).
  *
  * ## Untrusted input
  *