gsd-pi 2.74.0 → 2.75.0-dev.fd2382b9f

package/src/resources/extensions/gsd/gsd-db.ts

@@ -4,6 +4,21 @@
 //
 // Exposes a unified sync API for decisions and requirements storage.
 // Schema is initialized on first open with WAL mode for file-backed DBs.
+//
+// ─── Single-writer invariant ─────────────────────────────────────────────
+// This file is the ONLY place in the codebase that issues write SQL
+// (INSERT / UPDATE / DELETE / REPLACE / BEGIN-COMMIT transactions) against
+// the engine database at `.gsd/gsd.db`. All other modules must call the
+// typed wrappers exported here. The structural test
+// `tests/single-writer-invariant.test.ts` fails CI if a new bypass appears.
+//
+// `_getAdapter()` is retained for read-only SELECTs in query modules
+// (context-store, memory-store queries, doctor checks, projections).
+// Do NOT use it for writes — add a wrapper here instead.
+//
+// The separate `.gsd/unit-claims.db` managed by `unit-ownership.ts` is an
+// intentionally independent store for cross-worktree claim races and is
+// excluded from this invariant.
 
 import { createRequire } from "node:module";
 import { existsSync, copyFileSync, mkdirSync, realpathSync } from "node:fs";
@@ -12,6 +27,10 @@ import type { Decision, Requirement, GateRow, GateId, GateScope, GateStatus, Gat
 import { GSDError, GSD_STALE_STATE } from "./errors.js";
 import { getGateIdsForTurn, type OwnerTurn } from "./gate-registry.js";
 import { logError, logWarning } from "./workflow-logger.js";
+// Type-only import to avoid a circular runtime dep. The runtime side of
+// workflow-manifest.ts depends on this file, but the StateManifest type is
+// pure structure with no runtime coupling.
+import type { StateManifest } from "./workflow-manifest.js";
 
 const _require = createRequire(import.meta.url);
 
@@ -35,8 +54,8 @@ let loadAttempted = false;
 
 function suppressSqliteWarning(): void {
   const origEmit = process.emit;
-  // @ts-expect-error overriding process.emit for warning filter
-  process.emit = function (event: string, ...args: unknown[]): boolean {
+  // Override via loose cast: Node's overloaded emit signature is not directly assignable.
+  (process as any).emit = function (event: string, ...args: unknown[]): boolean {
     if (
       event === "warning" &&
       args[0] &&
@@ -161,7 +180,7 @@ function openRawDb(path: string): unknown {
   return new Database(path);
 }
 
-const SCHEMA_VERSION = 14;
+const SCHEMA_VERSION = 17;
 
 function indexExists(db: DbAdapter, name: string): boolean {
   return !!db.prepare(
@@ -216,6 +235,7 @@ function initSchema(db: DbAdapter, fileBacked: boolean): void {
         rationale TEXT NOT NULL DEFAULT '',
         revisable TEXT NOT NULL DEFAULT '',
         made_by TEXT NOT NULL DEFAULT 'agent',
+        source TEXT NOT NULL DEFAULT 'discussion', -- ADR-011 P2: 'discussion' | 'planning' | 'escalation'
         superseded_by TEXT DEFAULT NULL
       )
     `);
@@ -315,6 +335,8 @@ function initSchema(db: DbAdapter, fileBacked: boolean): void {
         observability_impact TEXT NOT NULL DEFAULT '',
         sequence INTEGER DEFAULT 0, -- Ordering hint: tools may set this to control execution order
         replan_triggered_at TEXT DEFAULT NULL,
+        is_sketch INTEGER NOT NULL DEFAULT 0, -- ADR-011: 1 = slice is a sketch awaiting refinement
+        sketch_scope TEXT NOT NULL DEFAULT '', -- ADR-011: 2-3 sentence rough scope from plan-milestone
         PRIMARY KEY (milestone_id, id),
         FOREIGN KEY (milestone_id) REFERENCES milestones(id)
       )
@@ -333,6 +355,11 @@ function initSchema(db: DbAdapter, fileBacked: boolean): void {
         duration TEXT NOT NULL DEFAULT '',
         completed_at TEXT DEFAULT NULL,
         blocker_discovered INTEGER DEFAULT 0,
+        blocker_source TEXT NOT NULL DEFAULT '', -- ADR-011 P2: provenance for blocker_discovered (e.g. 'reject-escalation')
+        escalation_pending INTEGER NOT NULL DEFAULT 0, -- ADR-011 P2: pause-on-escalation flag
+        escalation_awaiting_review INTEGER NOT NULL DEFAULT 0, -- ADR-011 P2: artifact exists but continueWithDefault=true (no pause)
+        escalation_artifact_path TEXT DEFAULT NULL, -- ADR-011 P2: path to T##-ESCALATION.json
+        escalation_override_applied_at TEXT DEFAULT NULL, -- ADR-011 P2: DB claim lock for idempotent override injection
         deviations TEXT NOT NULL DEFAULT '',
         known_issues TEXT NOT NULL DEFAULT '',
         key_files TEXT NOT NULL DEFAULT '[]',
@@ -424,6 +451,70 @@ function initSchema(db: DbAdapter, fileBacked: boolean): void {
       )
     `);
 
+    db.exec(`
+      CREATE TABLE IF NOT EXISTS gate_runs (
+        id INTEGER PRIMARY KEY AUTOINCREMENT,
+        trace_id TEXT NOT NULL,
+        turn_id TEXT NOT NULL,
+        gate_id TEXT NOT NULL,
+        gate_type TEXT NOT NULL DEFAULT '',
+        unit_type TEXT DEFAULT NULL,
+        unit_id TEXT DEFAULT NULL,
+        milestone_id TEXT DEFAULT NULL,
+        slice_id TEXT DEFAULT NULL,
+        task_id TEXT DEFAULT NULL,
+        outcome TEXT NOT NULL DEFAULT 'pass',
+        failure_class TEXT NOT NULL DEFAULT 'none',
+        rationale TEXT NOT NULL DEFAULT '',
+        findings TEXT NOT NULL DEFAULT '',
+        attempt INTEGER NOT NULL DEFAULT 1,
+        max_attempts INTEGER NOT NULL DEFAULT 1,
+        retryable INTEGER NOT NULL DEFAULT 0,
+        evaluated_at TEXT NOT NULL DEFAULT ''
+      )
+    `);
+
+    db.exec(`
+      CREATE TABLE IF NOT EXISTS turn_git_transactions (
+        trace_id TEXT NOT NULL,
+        turn_id TEXT NOT NULL,
+        unit_type TEXT DEFAULT NULL,
+        unit_id TEXT DEFAULT NULL,
+        stage TEXT NOT NULL DEFAULT 'turn-start',
+        action TEXT NOT NULL DEFAULT 'status-only',
+        push INTEGER NOT NULL DEFAULT 0,
+        status TEXT NOT NULL DEFAULT 'ok',
+        error TEXT DEFAULT NULL,
+        metadata_json TEXT NOT NULL DEFAULT '{}',
+        updated_at TEXT NOT NULL DEFAULT '',
+        PRIMARY KEY (trace_id, turn_id, stage)
+      )
+    `);
+
+    db.exec(`
+      CREATE TABLE IF NOT EXISTS audit_events (
+        event_id TEXT PRIMARY KEY,
+        trace_id TEXT NOT NULL,
+        turn_id TEXT DEFAULT NULL,
+        caused_by TEXT DEFAULT NULL,
+        category TEXT NOT NULL,
+        type TEXT NOT NULL,
+        ts TEXT NOT NULL,
+        payload_json TEXT NOT NULL DEFAULT '{}'
+      )
+    `);
+
+    db.exec(`
+      CREATE TABLE IF NOT EXISTS audit_turn_index (
+        trace_id TEXT NOT NULL,
+        turn_id TEXT NOT NULL,
+        first_ts TEXT NOT NULL,
+        last_ts TEXT NOT NULL,
+        event_count INTEGER NOT NULL DEFAULT 0,
+        PRIMARY KEY (trace_id, turn_id)
+      )
+    `);
+
     db.exec("CREATE INDEX IF NOT EXISTS idx_memories_active ON memories(superseded_by)");
     db.exec("CREATE INDEX IF NOT EXISTS idx_replan_history_milestone ON replan_history(milestone_id, created_at)");
 
@@ -437,6 +528,14 @@ function initSchema(db: DbAdapter, fileBacked: boolean): void {
 
     // v14 index — slice dependency lookups
     db.exec("CREATE INDEX IF NOT EXISTS idx_slice_deps_target ON slice_dependencies(milestone_id, depends_on_slice_id)");
+    db.exec("CREATE INDEX IF NOT EXISTS idx_gate_runs_turn ON gate_runs(trace_id, turn_id)");
+    db.exec("CREATE INDEX IF NOT EXISTS idx_gate_runs_lookup ON gate_runs(milestone_id, slice_id, task_id, gate_id)");
+    db.exec("CREATE INDEX IF NOT EXISTS idx_turn_git_tx_turn ON turn_git_transactions(trace_id, turn_id)");
+    db.exec("CREATE INDEX IF NOT EXISTS idx_audit_events_trace ON audit_events(trace_id, ts)");
+    db.exec("CREATE INDEX IF NOT EXISTS idx_audit_events_turn ON audit_events(trace_id, turn_id, ts)");
+    // ADR-011 Phase 2 — also created by the v17 migration; fresh installs
+    // skip migrations so the index must be created here too.
+    db.exec("CREATE INDEX IF NOT EXISTS idx_tasks_escalation_pending ON tasks(milestone_id, slice_id, escalation_pending)");
 
     db.exec(`CREATE VIEW IF NOT EXISTS active_decisions AS SELECT * FROM decisions WHERE superseded_by IS NULL`);
     db.exec(`CREATE VIEW IF NOT EXISTS active_requirements AS SELECT * FROM requirements WHERE superseded_by IS NULL`);
@@ -791,6 +890,104 @@ function migrateSchema(db: DbAdapter): void {
       });
     }
 
+    if (currentVersion < 15) {
+      db.exec(`
+        CREATE TABLE IF NOT EXISTS gate_runs (
+          id INTEGER PRIMARY KEY AUTOINCREMENT,
+          trace_id TEXT NOT NULL,
+          turn_id TEXT NOT NULL,
+          gate_id TEXT NOT NULL,
+          gate_type TEXT NOT NULL DEFAULT '',
+          unit_type TEXT DEFAULT NULL,
+          unit_id TEXT DEFAULT NULL,
+          milestone_id TEXT DEFAULT NULL,
+          slice_id TEXT DEFAULT NULL,
+          task_id TEXT DEFAULT NULL,
+          outcome TEXT NOT NULL DEFAULT 'pass',
+          failure_class TEXT NOT NULL DEFAULT 'none',
+          rationale TEXT NOT NULL DEFAULT '',
+          findings TEXT NOT NULL DEFAULT '',
+          attempt INTEGER NOT NULL DEFAULT 1,
+          max_attempts INTEGER NOT NULL DEFAULT 1,
+          retryable INTEGER NOT NULL DEFAULT 0,
+          evaluated_at TEXT NOT NULL DEFAULT ''
+        )
+      `);
+      db.exec(`
+        CREATE TABLE IF NOT EXISTS turn_git_transactions (
+          trace_id TEXT NOT NULL,
+          turn_id TEXT NOT NULL,
+          unit_type TEXT DEFAULT NULL,
+          unit_id TEXT DEFAULT NULL,
+          stage TEXT NOT NULL DEFAULT 'turn-start',
+          action TEXT NOT NULL DEFAULT 'status-only',
+          push INTEGER NOT NULL DEFAULT 0,
+          status TEXT NOT NULL DEFAULT 'ok',
+          error TEXT DEFAULT NULL,
+          metadata_json TEXT NOT NULL DEFAULT '{}',
+          updated_at TEXT NOT NULL DEFAULT '',
+          PRIMARY KEY (trace_id, turn_id, stage)
+        )
+      `);
+      db.exec(`
+        CREATE TABLE IF NOT EXISTS audit_events (
+          event_id TEXT PRIMARY KEY,
+          trace_id TEXT NOT NULL,
+          turn_id TEXT DEFAULT NULL,
+          caused_by TEXT DEFAULT NULL,
+          category TEXT NOT NULL,
+          type TEXT NOT NULL,
+          ts TEXT NOT NULL,
+          payload_json TEXT NOT NULL DEFAULT '{}'
+        )
+      `);
+      db.exec(`
+        CREATE TABLE IF NOT EXISTS audit_turn_index (
+          trace_id TEXT NOT NULL,
+          turn_id TEXT NOT NULL,
+          first_ts TEXT NOT NULL,
+          last_ts TEXT NOT NULL,
+          event_count INTEGER NOT NULL DEFAULT 0,
+          PRIMARY KEY (trace_id, turn_id)
+        )
+      `);
+      db.exec("CREATE INDEX IF NOT EXISTS idx_gate_runs_turn ON gate_runs(trace_id, turn_id)");
+      db.exec("CREATE INDEX IF NOT EXISTS idx_gate_runs_lookup ON gate_runs(milestone_id, slice_id, task_id, gate_id)");
+      db.exec("CREATE INDEX IF NOT EXISTS idx_turn_git_tx_turn ON turn_git_transactions(trace_id, turn_id)");
+      db.exec("CREATE INDEX IF NOT EXISTS idx_audit_events_trace ON audit_events(trace_id, ts)");
+      db.exec("CREATE INDEX IF NOT EXISTS idx_audit_events_turn ON audit_events(trace_id, turn_id, ts)");
+      db.prepare("INSERT INTO schema_version (version, applied_at) VALUES (:version, :applied_at)").run({
+        ":version": 15,
+        ":applied_at": new Date().toISOString(),
+      });
+    }
+
+    if (currentVersion < 16) {
+      // ADR-011 Phase 1: sketch-then-refine progressive planning — sketch columns on slices.
+      ensureColumn(db, "slices", "is_sketch", `ALTER TABLE slices ADD COLUMN is_sketch INTEGER NOT NULL DEFAULT 0`);
+      ensureColumn(db, "slices", "sketch_scope", `ALTER TABLE slices ADD COLUMN sketch_scope TEXT NOT NULL DEFAULT ''`);
+      // ADR-011 Phase 2: decisions can now be sourced from escalation resolutions.
+      ensureColumn(db, "decisions", "source", `ALTER TABLE decisions ADD COLUMN source TEXT NOT NULL DEFAULT 'discussion'`);
+      db.prepare("INSERT INTO schema_version (version, applied_at) VALUES (:version, :applied_at)").run({
+        ":version": 16,
+        ":applied_at": new Date().toISOString(),
+      });
+    }
+
+    if (currentVersion < 17) {
+      // ADR-011 Phase 2: mid-execution escalation — columns on the tasks table.
+      ensureColumn(db, "tasks", "blocker_source", `ALTER TABLE tasks ADD COLUMN blocker_source TEXT NOT NULL DEFAULT ''`);
+      ensureColumn(db, "tasks", "escalation_pending", `ALTER TABLE tasks ADD COLUMN escalation_pending INTEGER NOT NULL DEFAULT 0`);
+      ensureColumn(db, "tasks", "escalation_awaiting_review", `ALTER TABLE tasks ADD COLUMN escalation_awaiting_review INTEGER NOT NULL DEFAULT 0`);
+      ensureColumn(db, "tasks", "escalation_artifact_path", `ALTER TABLE tasks ADD COLUMN escalation_artifact_path TEXT DEFAULT NULL`);
+      ensureColumn(db, "tasks", "escalation_override_applied_at", `ALTER TABLE tasks ADD COLUMN escalation_override_applied_at TEXT DEFAULT NULL`);
+      db.exec("CREATE INDEX IF NOT EXISTS idx_tasks_escalation_pending ON tasks(milestone_id, slice_id, escalation_pending)");
+      db.prepare("INSERT INTO schema_version (version, applied_at) VALUES (:version, :applied_at)").run({
+        ":version": 17,
+        ":applied_at": new Date().toISOString(),
+      });
+    }
+
     db.exec("COMMIT");
   } catch (err) {
     db.exec("ROLLBACK");
@@ -892,6 +1089,14 @@ export function vacuumDatabase(): void {
   } catch (e) { logWarning("db", `VACUUM failed: ${(e as Error).message}`); }
 }
 
+/** Flush WAL into gsd.db so `git add .gsd/gsd.db` stages current state — safe while DB is open. */
+export function checkpointDatabase(): void {
+  if (!currentDb) return;
+  try {
+    currentDb.exec('PRAGMA wal_checkpoint(TRUNCATE)');
+  } catch (e) { logWarning("db", `WAL checkpoint failed: ${(e as Error).message}`); }
+}
+
 let _txDepth = 0;
 
 export function transaction<T>(fn: () => T): T {
@@ -922,11 +1127,53 @@ export function transaction<T>(fn: () => T): T {
   }
 }
 
+/**
+ * Wrap a block of reads in a DEFERRED transaction so that all SELECTs observe
+ * a consistent snapshot of the DB even if a concurrent writer commits between
+ * them. Use this for multi-query read flows (e.g. tool executors that query
+ * milestone + slices + counts and want one snapshot). Re-entrant — if already
+ * inside a transaction, runs fn() without starting a nested one.
+ */
+export function readTransaction<T>(fn: () => T): T {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+
+  if (_txDepth > 0) {
+    _txDepth++;
+    try {
+      return fn();
+    } finally {
+      _txDepth--;
+    }
+  }
+
+  _txDepth++;
+  currentDb.exec("BEGIN DEFERRED");
+  try {
+    const result = fn();
+    currentDb.exec("COMMIT");
+    return result;
+  } catch (err) {
+    try {
+      currentDb.exec("ROLLBACK");
+    } catch (rollbackErr) {
+      // A failed ROLLBACK after a failed read is a split-brain signal —
+      // the transaction is in an indeterminate state. Surface it via the
+      // logger instead of swallowing it.
+      logError("db", "snapshotState ROLLBACK failed", {
+        error: (rollbackErr as Error).message,
+      });
+    }
+    throw err;
+  } finally {
+    _txDepth--;
+  }
+}
+
 export function insertDecision(d: Omit<Decision, "seq">): void {
   if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
   currentDb.prepare(
-    `INSERT INTO decisions (id, when_context, scope, decision, choice, rationale, revisable, made_by, superseded_by)
-     VALUES (:id, :when_context, :scope, :decision, :choice, :rationale, :revisable, :made_by, :superseded_by)`,
+    `INSERT INTO decisions (id, when_context, scope, decision, choice, rationale, revisable, made_by, source, superseded_by)
+     VALUES (:id, :when_context, :scope, :decision, :choice, :rationale, :revisable, :made_by, :source, :superseded_by)`,
   ).run({
     ":id": d.id,
     ":when_context": d.when_context,
@@ -936,6 +1183,7 @@ export function insertDecision(d: Omit<Decision, "seq">): void {
     ":rationale": d.rationale,
     ":revisable": d.revisable,
     ":made_by": d.made_by ?? "agent",
+    ":source": d.source ?? "discussion",
     ":superseded_by": d.superseded_by,
   });
 }
@@ -954,6 +1202,7 @@ export function getDecisionById(id: string): Decision | null {
     rationale: row["rationale"] as string,
     revisable: row["revisable"] as string,
     made_by: (row["made_by"] as string as import("./types.js").DecisionMadeBy) ?? "agent",
+    source: (row["source"] as string) ?? "discussion",
     superseded_by: (row["superseded_by"] as string) ?? null,
   };
 }
@@ -971,6 +1220,7 @@ export function getActiveDecisions(): Decision[] {
     rationale: row["rationale"] as string,
     revisable: row["revisable"] as string,
     made_by: (row["made_by"] as string as import("./types.js").DecisionMadeBy) ?? "agent",
+    source: (row["source"] as string) ?? "discussion",
     superseded_by: null,
   }));
 }
@@ -1059,8 +1309,8 @@ export function upsertDecision(d: Omit<Decision, "seq">): void {
   // seq column. INSERT OR REPLACE deletes then reinserts, resetting seq and
   // corrupting decision ordering in DECISIONS.md after reconcile replay.
   currentDb.prepare(
-    `INSERT INTO decisions (id, when_context, scope, decision, choice, rationale, revisable, made_by, superseded_by)
-     VALUES (:id, :when_context, :scope, :decision, :choice, :rationale, :revisable, :made_by, :superseded_by)
+    `INSERT INTO decisions (id, when_context, scope, decision, choice, rationale, revisable, made_by, source, superseded_by)
+     VALUES (:id, :when_context, :scope, :decision, :choice, :rationale, :revisable, :made_by, :source, :superseded_by)
      ON CONFLICT(id) DO UPDATE SET
        when_context = excluded.when_context,
        scope = excluded.scope,
@@ -1069,6 +1319,7 @@ export function upsertDecision(d: Omit<Decision, "seq">): void {
        rationale = excluded.rationale,
        revisable = excluded.revisable,
        made_by = excluded.made_by,
+       source = excluded.source,
        superseded_by = excluded.superseded_by`,
   ).run({
     ":id": d.id,
@@ -1079,6 +1330,7 @@ export function upsertDecision(d: Omit<Decision, "seq">): void {
     ":rationale": d.rationale,
     ":revisable": d.revisable,
     ":made_by": d.made_by ?? "agent",
+    ":source": d.source ?? "discussion",
     ":superseded_by": d.superseded_by ?? null,
   });
 }
@@ -1253,16 +1505,20 @@ export function insertSlice(s: {
   depends?: string[];
   demo?: string;
   sequence?: number;
+  isSketch?: boolean;
+  sketchScope?: string;
   planning?: Partial<SlicePlanningRecord>;
 }): void {
   if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
   currentDb.prepare(
     `INSERT INTO slices (
       milestone_id, id, title, status, risk, depends, demo, created_at,
-      goal, success_criteria, proof_level, integration_closure, observability_impact, sequence
+      goal, success_criteria, proof_level, integration_closure, observability_impact, sequence,
+      is_sketch, sketch_scope
     ) VALUES (
       :milestone_id, :id, :title, :status, :risk, :depends, :demo, :created_at,
-      :goal, :success_criteria, :proof_level, :integration_closure, :observability_impact, :sequence
+      :goal, :success_criteria, :proof_level, :integration_closure, :observability_impact, :sequence,
+      :is_sketch, :sketch_scope
     )
     ON CONFLICT (milestone_id, id) DO UPDATE SET
       title = CASE WHEN :raw_title IS NOT NULL THEN excluded.title ELSE slices.title END,
@@ -1275,7 +1531,9 @@ export function insertSlice(s: {
       proof_level = CASE WHEN :raw_proof_level IS NOT NULL THEN excluded.proof_level ELSE slices.proof_level END,
       integration_closure = CASE WHEN :raw_integration_closure IS NOT NULL THEN excluded.integration_closure ELSE slices.integration_closure END,
       observability_impact = CASE WHEN :raw_observability_impact IS NOT NULL THEN excluded.observability_impact ELSE slices.observability_impact END,
-      sequence = CASE WHEN :raw_sequence IS NOT NULL THEN excluded.sequence ELSE slices.sequence END`,
+      sequence = CASE WHEN :raw_sequence IS NOT NULL THEN excluded.sequence ELSE slices.sequence END,
+      is_sketch = CASE WHEN :raw_is_sketch IS NOT NULL THEN excluded.is_sketch ELSE slices.is_sketch END,
+      sketch_scope = CASE WHEN :raw_sketch_scope IS NOT NULL THEN excluded.sketch_scope ELSE slices.sketch_scope END`,
   ).run({
     ":milestone_id": s.milestoneId,
     ":id": s.id,
@@ -1291,6 +1549,8 @@ export function insertSlice(s: {
     ":integration_closure": s.planning?.integrationClosure ?? "",
     ":observability_impact": s.planning?.observabilityImpact ?? "",
     ":sequence": s.sequence ?? 0,
+    ":is_sketch": s.isSketch ? 1 : 0,
+    ":sketch_scope": s.sketchScope ?? "",
     // Raw sentinel params: NULL when caller omitted the field, used in ON CONFLICT guards
     ":raw_title": s.title ?? null,
     ":raw_risk": s.risk ?? null,
@@ -1301,9 +1561,52 @@ export function insertSlice(s: {
     ":raw_integration_closure": s.planning?.integrationClosure ?? null,
     ":raw_observability_impact": s.planning?.observabilityImpact ?? null,
     ":raw_sequence": s.sequence ?? null,
+    ":raw_is_sketch": s.isSketch === undefined ? null : (s.isSketch ? 1 : 0),
+    // NOTE: use !== undefined (not ??) so an explicit empty string "" is treated
+    // as a present value and correctly clears the existing sketch_scope on
+    // CONFLICT. ?? would incorrectly preserve the stale value.
+    ":raw_sketch_scope": s.sketchScope !== undefined ? s.sketchScope : null,
   });
 }
 
+// ADR-011: sketch-then-refine helpers
+export function setSliceSketchFlag(milestoneId: string, sliceId: string, isSketch: boolean): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE slices SET is_sketch = :is_sketch WHERE milestone_id = :mid AND id = :sid`,
+  ).run({ ":is_sketch": isSketch ? 1 : 0, ":mid": milestoneId, ":sid": sliceId });
+}
+
+/**
+ * ADR-011 auto-heal: reconcile stale is_sketch=1 rows whose PLAN already exists.
+ *
+ * Callers pass a predicate that resolves whether a plan file exists for a slice.
+ * The predicate MUST use the canonical path resolver (`resolveSliceFile`, etc.)
+ * to keep path logic in one place — do not hand-roll the path inside the callback.
+ *
+ * Recovers from two scenarios:
+ *   1. Crash between `gsd_plan_slice` write and the sketch flag flip.
+ *   2. Flag-OFF downgrade path: when `progressive_planning` is off, the dispatch
+ *      rule routes sketch slices to plan-slice, which writes PLAN.md but leaves
+ *      `is_sketch=1` — the next state derivation auto-heals it to 0 here.
+ *
+ * Not aggressive in practice: PLAN.md is only written via the DB-backed
+ * `gsd_plan_slice` tool (which also inserts tasks), so a "stale PLAN.md with
+ * is_sketch=1" is extremely unlikely to indicate anything other than the two
+ * recovery scenarios above.
+ */
+export function autoHealSketchFlags(milestoneId: string, hasPlanFile: (sliceId: string) => boolean): void {
+  if (!currentDb) return;
+  const rows = currentDb.prepare(
+    `SELECT id FROM slices WHERE milestone_id = :mid AND is_sketch = 1`,
+  ).all({ ":mid": milestoneId }) as Array<{ id: string }>;
+  for (const row of rows) {
+    if (hasPlanFile(row.id)) {
+      setSliceSketchFlag(milestoneId, row.id, false);
+    }
+  }
+}
+
 export function upsertSlicePlanning(milestoneId: string, sliceId: string, planning: Partial<SlicePlanningRecord>): void {
   if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
   currentDb.prepare(
@@ -1477,6 +1780,8 @@ export interface SliceRow {
   observability_impact: string;
   sequence: number;
   replan_triggered_at: string | null;
+  is_sketch: number;
+  sketch_scope: string;
 }
 
 function rowToSlice(row: Record<string, unknown>): SliceRow {
@@ -1499,6 +1804,8 @@ function rowToSlice(row: Record<string, unknown>): SliceRow {
     observability_impact: (row["observability_impact"] as string) ?? "",
     sequence: (row["sequence"] as number) ?? 0,
     replan_triggered_at: (row["replan_triggered_at"] as string) ?? null,
+    is_sketch: (row["is_sketch"] as number) ?? 0,
+    sketch_scope: (row["sketch_scope"] as string) ?? "",
   };
 }
 
@@ -1562,6 +1869,12 @@ export interface TaskRow {
   observability_impact: string;
   full_plan_md: string;
   sequence: number;
+  // ADR-011 Phase 2 escalation fields
+  blocker_source: string;
+  escalation_pending: number;
+  escalation_awaiting_review: number;
+  escalation_artifact_path: string | null;
+  escalation_override_applied_at: string | null;
 }
 
 function parseTaskArrayColumn(raw: unknown): string[] {
@@ -1632,6 +1945,11 @@ function rowToTask(row: Record<string, unknown>): TaskRow {
     observability_impact: (row["observability_impact"] as string) ?? "",
     full_plan_md: (row["full_plan_md"] as string) ?? "",
     sequence: (row["sequence"] as number) ?? 0,
+    blocker_source: (row["blocker_source"] as string) ?? "",
+    escalation_pending: (row["escalation_pending"] as number) ?? 0,
+    escalation_awaiting_review: (row["escalation_awaiting_review"] as number) ?? 0,
+    escalation_artifact_path: (row["escalation_artifact_path"] as string) ?? null,
+    escalation_override_applied_at: (row["escalation_override_applied_at"] as string) ?? null,
   };
 }
 
@@ -1652,6 +1970,125 @@ export function getSliceTasks(milestoneId: string, sliceId: string): TaskRow[] {
   return rows.map(rowToTask);
 }
 
+// ─── ADR-011 Phase 2 escalation helpers ──────────────────────────────────
+
+/** Set pause-on-escalation state on a completed task. Mutually exclusive with awaiting_review. */
+export function setTaskEscalationPending(
+  milestoneId: string, sliceId: string, taskId: string,
+  artifactPath: string,
+): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE tasks
+       SET escalation_pending = 1,
+           escalation_awaiting_review = 0,
+           escalation_artifact_path = :path
+     WHERE milestone_id = :mid AND slice_id = :sid AND id = :tid`,
+  ).run({ ":path": artifactPath, ":mid": milestoneId, ":sid": sliceId, ":tid": taskId });
+}
+
+/** Set awaiting-review state (artifact exists but continueWithDefault=true, no pause). Mutually exclusive with pending. */
+export function setTaskEscalationAwaitingReview(
+  milestoneId: string, sliceId: string, taskId: string,
+  artifactPath: string,
+): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE tasks
+       SET escalation_awaiting_review = 1,
+           escalation_pending = 0,
+           escalation_artifact_path = :path
+     WHERE milestone_id = :mid AND slice_id = :sid AND id = :tid`,
+  ).run({ ":path": artifactPath, ":mid": milestoneId, ":sid": sliceId, ":tid": taskId });
+}
+
+/** Clear escalation-pending and awaiting-review flags once the user has resolved it. */
+export function clearTaskEscalationFlags(
+  milestoneId: string, sliceId: string, taskId: string,
+): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE tasks
+       SET escalation_pending = 0,
+           escalation_awaiting_review = 0
+     WHERE milestone_id = :mid AND slice_id = :sid AND id = :tid`,
+  ).run({ ":mid": milestoneId, ":sid": sliceId, ":tid": taskId });
+}
+
+/**
+ * Atomically claim a resolved escalation override for injection into a downstream
+ * task's prompt. Returns true if this caller claimed it (must inject), false if
+ * another caller already claimed it (must skip).
+ */
+export function claimEscalationOverride(
+  milestoneId: string, sliceId: string, sourceTaskId: string,
+): boolean {
+  if (!currentDb) return false;
+  const now = new Date().toISOString();
+  const result = currentDb.prepare(
+    `UPDATE tasks
+       SET escalation_override_applied_at = :now
+     WHERE milestone_id = :mid AND slice_id = :sid AND id = :tid
+       AND escalation_override_applied_at IS NULL
+       AND escalation_artifact_path IS NOT NULL`,
+  ).run({ ":now": now, ":mid": milestoneId, ":sid": sliceId, ":tid": sourceTaskId });
+  // node:sqlite + better-sqlite3 both surface `changes` on the run result.
+  const changes = (result as { changes?: number }).changes ?? 0;
+  return changes > 0;
+}
+
+/** Find the most recent resolved-but-unapplied escalation override in a slice. */
+export function findUnappliedEscalationOverride(
+  milestoneId: string, sliceId: string,
+): { taskId: string; artifactPath: string } | null {
+  if (!currentDb) return null;
+  // Filter BOTH flags: escalation_pending=0 AND escalation_awaiting_review=0
+  // ensures we only claim overrides the user has explicitly resolved.
+  // Without the awaiting_review filter, continueWithDefault=true artifacts
+  // (not yet responded to) would be prematurely claimed, causing the override
+  // to be lost when the user later resolves (#ADR-011 Phase 2 peer-review Bug 2).
+  const row = currentDb.prepare(
+    `SELECT id, escalation_artifact_path AS path
+       FROM tasks
+      WHERE milestone_id = :mid AND slice_id = :sid
+        AND escalation_artifact_path IS NOT NULL
+        AND escalation_override_applied_at IS NULL
+        AND escalation_pending = 0
+        AND escalation_awaiting_review = 0
+      ORDER BY sequence DESC, id DESC
+      LIMIT 1`,
+  ).get({ ":mid": milestoneId, ":sid": sliceId }) as
+    | { id: string; path: string | null }
+    | undefined;
+  if (!row || !row.path) return null;
+  return { taskId: row.id, artifactPath: row.path };
+}
+
+/** Set the blocker_source provenance field (used when rejecting an escalation). */
+export function setTaskBlockerSource(
+  milestoneId: string, sliceId: string, taskId: string, source: string,
+): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE tasks
+       SET blocker_discovered = 1,
+           blocker_source = :src
+     WHERE milestone_id = :mid AND slice_id = :sid AND id = :tid`,
+  ).run({ ":src": source, ":mid": milestoneId, ":sid": sliceId, ":tid": taskId });
+}
+
+/** List tasks with active escalation artifacts across a milestone (for /gsd escalate list). */
+export function listEscalationArtifacts(milestoneId: string, includeResolved: boolean = false): TaskRow[] {
+  if (!currentDb) return [];
+  const filter = includeResolved
+    ? "escalation_artifact_path IS NOT NULL"
+    : "(escalation_pending = 1 OR escalation_awaiting_review = 1) AND escalation_artifact_path IS NOT NULL";
+  const rows = currentDb.prepare(
+    `SELECT * FROM tasks WHERE milestone_id = :mid AND ${filter} ORDER BY slice_id, sequence, id`,
+  ).all({ ":mid": milestoneId });
+  return rows.map(rowToTask);
+}
+
 export function insertVerificationEvidence(e: {
   taskId: string;
   sliceId: string;
@@ -1965,6 +2402,19 @@ export function reconcileWorktreeDb(
     try {
       const wtInfo = adapter.prepare("PRAGMA wt.table_info('decisions')").all();
       const hasMadeBy = wtInfo.some((col) => col["name"] === "made_by");
+      // ADR-011: worktree may predate schema v16/v17. For missing columns we
+      // fall through to the main DB's existing value (not a literal default)
+      // so reconcile never silently clears state the main tree has recorded.
+      const hasDecisionSource = wtInfo.some((col) => col["name"] === "source");
+      const wtSliceInfo = adapter.prepare("PRAGMA wt.table_info('slices')").all();
+      const hasIsSketch = wtSliceInfo.some((col) => col["name"] === "is_sketch");
+      const hasSketchScope = wtSliceInfo.some((col) => col["name"] === "sketch_scope");
+      const wtTaskInfo = adapter.prepare("PRAGMA wt.table_info('tasks')").all();
+      const hasBlockerSource = wtTaskInfo.some((col) => col["name"] === "blocker_source");
+      const hasEscalationPending = wtTaskInfo.some((col) => col["name"] === "escalation_pending");
+      const hasEscalationAwaiting = wtTaskInfo.some((col) => col["name"] === "escalation_awaiting_review");
+      const hasEscalationArtifact = wtTaskInfo.some((col) => col["name"] === "escalation_artifact_path");
+      const hasEscalationOverride = wtTaskInfo.some((col) => col["name"] === "escalation_override_applied_at");
 
       const decConf = adapter.prepare(
         `SELECT m.id FROM decisions m INNER JOIN wt.decisions w ON m.id = w.id WHERE m.decision != w.decision OR m.choice != w.choice OR m.rationale != w.rationale OR ${
@@ -1986,13 +2436,20 @@ export function reconcileWorktreeDb(
 
       adapter.exec("BEGIN");
       try {
+        // Join the target decisions so we can prefer an existing main.source
+        // when the worktree predates v16 — otherwise a write-through reconcile
+        // would clobber 'escalation'-sourced decisions with the literal default.
         merged.decisions = countChanges(adapter.prepare(`
           INSERT OR REPLACE INTO decisions (
-            id, when_context, scope, decision, choice, rationale, revisable, made_by, superseded_by
+            id, when_context, scope, decision, choice, rationale, revisable, made_by, source, superseded_by
           )
-          SELECT id, when_context, scope, decision, choice, rationale, revisable, ${
-            hasMadeBy ? "made_by" : "'agent'"
-          }, superseded_by FROM wt.decisions
+          SELECT w.id, w.when_context, w.scope, w.decision, w.choice, w.rationale, w.revisable, ${
+            hasMadeBy ? "w.made_by" : "COALESCE(m.made_by, 'agent')"
+          }, ${
+            hasDecisionSource ? "w.source" : "COALESCE(m.source, 'discussion')"
+          }, w.superseded_by
+          FROM wt.decisions w
+          LEFT JOIN decisions m ON m.id = w.id
         `).run());
 
         merged.requirements = countChanges(adapter.prepare(`
@@ -2029,13 +2486,15 @@ export function reconcileWorktreeDb(
         `).run());
 
         // Merge slices — preserve worktree progress but never downgrade completed status (#2558).
-        // Uses INSERT OR REPLACE with a subquery that picks the best status — if the main DB
-        // already has a completed slice, keep that status even if the worktree copy is stale.
+        // ADR-011 Phase 1: carry is_sketch + sketch_scope so reconcile doesn't
+        // silently clear sketch metadata. When the worktree predates v16,
+        // fall back to the main DB's existing value rather than a literal 0/''.
         merged.slices = countChanges(adapter.prepare(`
           INSERT OR REPLACE INTO slices (
             milestone_id, id, title, status, risk, depends, demo, created_at, completed_at,
             full_summary_md, full_uat_md, goal, success_criteria, proof_level,
-            integration_closure, observability_impact, sequence, replan_triggered_at
+            integration_closure, observability_impact, sequence, replan_triggered_at,
+            is_sketch, sketch_scope
           )
           SELECT w.milestone_id, w.id, w.title,
                  CASE
@@ -2048,19 +2507,25 @@ export function reconcileWorktreeDb(
                    THEN m.completed_at ELSE w.completed_at
                  END,
                  w.full_summary_md, w.full_uat_md, w.goal, w.success_criteria, w.proof_level,
-                 w.integration_closure, w.observability_impact, w.sequence, w.replan_triggered_at
+                 w.integration_closure, w.observability_impact, w.sequence, w.replan_triggered_at,
+                 ${hasIsSketch ? "w.is_sketch" : "COALESCE(m.is_sketch, 0)"},
+                 ${hasSketchScope ? "w.sketch_scope" : "COALESCE(m.sketch_scope, '')"}
           FROM wt.slices w
           LEFT JOIN slices m ON m.milestone_id = w.milestone_id AND m.id = w.id
         `).run());
 
-        // Merge tasks — preserve execution results, never downgrade completed status (#2558)
+        // Merge tasks — preserve execution results, never downgrade completed status (#2558).
+        // ADR-011 P2: carry blocker_source + escalation_* columns so worktree reconcile
+        // doesn't silently clear escalation state back to defaults.
         merged.tasks = countChanges(adapter.prepare(`
           INSERT OR REPLACE INTO tasks (
             milestone_id, slice_id, id, title, status, one_liner, narrative,
             verification_result, duration, completed_at, blocker_discovered,
             deviations, known_issues, key_files, key_decisions, full_summary_md,
             description, estimate, files, verify, inputs, expected_output,
-            observability_impact, full_plan_md, sequence
+            observability_impact, full_plan_md, sequence,
+            blocker_source, escalation_pending, escalation_awaiting_review,
+            escalation_artifact_path, escalation_override_applied_at
           )
           SELECT w.milestone_id, w.slice_id, w.id, w.title,
                  CASE
@@ -2076,7 +2541,12 @@ export function reconcileWorktreeDb(
                  w.blocker_discovered,
                  w.deviations, w.known_issues, w.key_files, w.key_decisions, w.full_summary_md,
                  w.description, w.estimate, w.files, w.verify, w.inputs, w.expected_output,
-                 w.observability_impact, w.full_plan_md, w.sequence
+                 w.observability_impact, w.full_plan_md, w.sequence,
+                 ${hasBlockerSource ? "w.blocker_source" : "COALESCE(m.blocker_source, '')"},
+                 ${hasEscalationPending ? "w.escalation_pending" : "COALESCE(m.escalation_pending, 0)"},
+                 ${hasEscalationAwaiting ? "w.escalation_awaiting_review" : "COALESCE(m.escalation_awaiting_review, 0)"},
+                 ${hasEscalationArtifact ? "w.escalation_artifact_path" : "m.escalation_artifact_path"},
+                 ${hasEscalationOverride ? "w.escalation_override_applied_at" : "m.escalation_override_applied_at"}
           FROM wt.tasks w
           LEFT JOIN tasks m ON m.milestone_id = w.milestone_id AND m.slice_id = w.slice_id AND m.id = w.id
         `).run());
@@ -2226,6 +2696,9 @@ export function deleteMilestone(milestoneId: string): void {
     currentDb!.prepare(
       `DELETE FROM quality_gates WHERE milestone_id = :mid`,
     ).run({ ":mid": milestoneId });
+    currentDb!.prepare(
+      `DELETE FROM gate_runs WHERE milestone_id = :mid`,
+    ).run({ ":mid": milestoneId });
     currentDb!.prepare(
       `DELETE FROM tasks WHERE milestone_id = :mid`,
     ).run({ ":mid": milestoneId });
@@ -2359,6 +2832,30 @@ export function saveGateResult(g: {
     ":findings": g.findings,
     ":evaluated_at": new Date().toISOString(),
   });
+
+  const outcome =
+    g.verdict === "pass"
+      ? "pass"
+      : g.verdict === "omitted"
+        ? "manual-attention"
+        : "fail";
+  insertGateRun({
+    traceId: `quality-gate:${g.milestoneId}:${g.sliceId}`,
+    turnId: `gate:${g.gateId}:${g.taskId ?? "slice"}`,
+    gateId: g.gateId,
+    gateType: "quality-gate",
+    milestoneId: g.milestoneId,
+    sliceId: g.sliceId,
+    taskId: g.taskId ?? undefined,
+    outcome,
+    failureClass: outcome === "fail" ? "verification" : outcome === "manual-attention" ? "manual-attention" : "none",
+    rationale: g.rationale,
+    findings: g.findings,
+    attempt: 1,
+    maxAttempts: 1,
+    retryable: false,
+    evaluatedAt: new Date().toISOString(),
+  });
 }
 
 export function getPendingGates(milestoneId: string, sliceId: string, scope?: GateScope): GateRow[] {
@@ -2451,3 +2948,568 @@ export function getPendingGateCountForTurn(
 ): number {
   return getPendingGatesForTurn(milestoneId, sliceId, turn).length;
 }
+
+export function insertGateRun(entry: {
+  traceId: string;
+  turnId: string;
+  gateId: string;
+  gateType: string;
+  unitType?: string;
+  unitId?: string;
+  milestoneId?: string;
+  sliceId?: string;
+  taskId?: string;
+  outcome: "pass" | "fail" | "retry" | "manual-attention";
+  failureClass: "none" | "policy" | "input" | "execution" | "artifact" | "verification" | "closeout" | "git" | "timeout" | "manual-attention" | "unknown";
+  rationale?: string;
+  findings?: string;
+  attempt: number;
+  maxAttempts: number;
+  retryable: boolean;
+  evaluatedAt: string;
+}): void {
+  if (!currentDb) return;
+  currentDb.prepare(
+    `INSERT INTO gate_runs (
+      trace_id, turn_id, gate_id, gate_type, unit_type, unit_id, milestone_id, slice_id, task_id,
+      outcome, failure_class, rationale, findings, attempt, max_attempts, retryable, evaluated_at
+    ) VALUES (
+      :trace_id, :turn_id, :gate_id, :gate_type, :unit_type, :unit_id, :milestone_id, :slice_id, :task_id,
+      :outcome, :failure_class, :rationale, :findings, :attempt, :max_attempts, :retryable, :evaluated_at
+    )`,
+  ).run({
+    ":trace_id": entry.traceId,
+    ":turn_id": entry.turnId,
+    ":gate_id": entry.gateId,
+    ":gate_type": entry.gateType,
+    ":unit_type": entry.unitType ?? null,
+    ":unit_id": entry.unitId ?? null,
+    ":milestone_id": entry.milestoneId ?? null,
+    ":slice_id": entry.sliceId ?? null,
+    ":task_id": entry.taskId ?? null,
+    ":outcome": entry.outcome,
+    ":failure_class": entry.failureClass,
+    ":rationale": entry.rationale ?? "",
+    ":findings": entry.findings ?? "",
+    ":attempt": entry.attempt,
+    ":max_attempts": entry.maxAttempts,
+    ":retryable": entry.retryable ? 1 : 0,
+    ":evaluated_at": entry.evaluatedAt,
+  });
+}
+
+export function upsertTurnGitTransaction(entry: {
+  traceId: string;
+  turnId: string;
+  unitType?: string;
+  unitId?: string;
+  stage: string;
+  action: "commit" | "snapshot" | "status-only";
+  push: boolean;
+  status: "ok" | "failed";
+  error?: string;
+  metadata?: Record<string, unknown>;
+  updatedAt: string;
+}): void {
+  if (!currentDb) return;
+  currentDb.prepare(
+    `INSERT OR REPLACE INTO turn_git_transactions (
+      trace_id, turn_id, unit_type, unit_id, stage, action, push, status, error, metadata_json, updated_at
+    ) VALUES (
+      :trace_id, :turn_id, :unit_type, :unit_id, :stage, :action, :push, :status, :error, :metadata_json, :updated_at
+    )`,
+  ).run({
+    ":trace_id": entry.traceId,
+    ":turn_id": entry.turnId,
+    ":unit_type": entry.unitType ?? null,
+    ":unit_id": entry.unitId ?? null,
+    ":stage": entry.stage,
+    ":action": entry.action,
+    ":push": entry.push ? 1 : 0,
+    ":status": entry.status,
+    ":error": entry.error ?? null,
+    ":metadata_json": JSON.stringify(entry.metadata ?? {}),
+    ":updated_at": entry.updatedAt,
+  });
+}
+
+export function insertAuditEvent(entry: {
+  eventId: string;
+  traceId: string;
+  turnId?: string;
+  causedBy?: string;
+  category: string;
+  type: string;
+  ts: string;
+  payload: Record<string, unknown>;
+}): void {
+  if (!currentDb) return;
+  transaction(() => {
+    currentDb!.prepare(
+      `INSERT OR IGNORE INTO audit_events (
+        event_id, trace_id, turn_id, caused_by, category, type, ts, payload_json
+      ) VALUES (
+        :event_id, :trace_id, :turn_id, :caused_by, :category, :type, :ts, :payload_json
+      )`,
+    ).run({
+      ":event_id": entry.eventId,
+      ":trace_id": entry.traceId,
+      ":turn_id": entry.turnId ?? null,
+      ":caused_by": entry.causedBy ?? null,
+      ":category": entry.category,
+      ":type": entry.type,
+      ":ts": entry.ts,
+      ":payload_json": JSON.stringify(entry.payload ?? {}),
+    });
+
+    if (entry.turnId) {
+      const row = currentDb!.prepare(
+        `SELECT event_count, first_ts, last_ts
+         FROM audit_turn_index
+         WHERE trace_id = :trace_id AND turn_id = :turn_id`,
+      ).get({
+        ":trace_id": entry.traceId,
+        ":turn_id": entry.turnId,
+      });
+      if (row) {
+        currentDb!.prepare(
+          `UPDATE audit_turn_index
+           SET first_ts = CASE WHEN :ts < first_ts THEN :ts ELSE first_ts END,
+               last_ts = CASE WHEN :ts > last_ts THEN :ts ELSE last_ts END,
+               event_count = event_count + 1
+           WHERE trace_id = :trace_id AND turn_id = :turn_id`,
+        ).run({
+          ":trace_id": entry.traceId,
+          ":turn_id": entry.turnId,
+          ":ts": entry.ts,
+        });
+      } else {
+        currentDb!.prepare(
+          `INSERT INTO audit_turn_index (trace_id, turn_id, first_ts, last_ts, event_count)
+           VALUES (:trace_id, :turn_id, :first_ts, :last_ts, :event_count)`,
+        ).run({
+          ":trace_id": entry.traceId,
+          ":turn_id": entry.turnId,
+          ":first_ts": entry.ts,
+          ":last_ts": entry.ts,
+          ":event_count": 1,
+        });
+      }
+    }
+  });
+}
+
+// ─── Single-writer bypass wrappers ───────────────────────────────────────
+// These wrappers exist so modules outside this file never need to call
+// `_getAdapter()` for writes. Each one is a byte-equivalent replacement for
+// a raw prepare/run previously issued from another module. Keep them
+// minimal and direct — they exist to hold SQL text in one place, not to
+// add new behavior.
+
+/** Delete a decision row by id. Used by db-writer.ts rollback on disk-write failure. */
+export function deleteDecisionById(id: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare("DELETE FROM decisions WHERE id = :id").run({ ":id": id });
+}
+
+/** Delete a requirement row by id. Used by db-writer.ts rollback on disk-write failure. */
+export function deleteRequirementById(id: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare("DELETE FROM requirements WHERE id = :id").run({ ":id": id });
+}
+
+/** Delete an artifact row by path. Used by db-writer.ts rollback on disk-write failure. */
+export function deleteArtifactByPath(path: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare("DELETE FROM artifacts WHERE path = :path").run({ ":path": path });
+}
+
+/**
+ * Drop all rows from tasks/slices/milestones in dependency order inside a
+ * transaction. Used by `gsd recover` to rebuild engine state from markdown.
+ */
+export function clearEngineHierarchy(): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  transaction(() => {
+    currentDb!.exec("DELETE FROM tasks");
+    currentDb!.exec("DELETE FROM slices");
+    currentDb!.exec("DELETE FROM milestones");
+  });
+}
+
+/**
+ * INSERT OR IGNORE a slice during event replay (workflow-reconcile.ts).
+ * Strict insert-or-ignore semantics are required here to avoid the
+ * `insertSlice` ON CONFLICT path that could downgrade an already-completed
+ * slice back to 'pending'.
+ */
+export function insertOrIgnoreSlice(args: {
+  milestoneId: string;
+  sliceId: string;
+  title: string;
+  createdAt: string;
+}): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `INSERT OR IGNORE INTO slices (milestone_id, id, title, status, created_at)
+     VALUES (:mid, :sid, :title, 'pending', :ts)`,
+  ).run({
+    ":mid": args.milestoneId,
+    ":sid": args.sliceId,
+    ":title": args.title,
+    ":ts": args.createdAt,
+  });
+}
+
+/**
+ * INSERT OR IGNORE a task during event replay (workflow-reconcile.ts).
+ * Same rationale as `insertOrIgnoreSlice`.
+ */
+export function insertOrIgnoreTask(args: {
+  milestoneId: string;
+  sliceId: string;
+  taskId: string;
+  title: string;
+  createdAt: string;
+}): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `INSERT OR IGNORE INTO tasks (milestone_id, slice_id, id, title, status, created_at)
+     VALUES (:mid, :sid, :tid, :title, 'pending', :ts)`,
+  ).run({
+    ":mid": args.milestoneId,
+    ":sid": args.sliceId,
+    ":tid": args.taskId,
+    ":title": args.title,
+    ":ts": args.createdAt,
+  });
+}
+
+/**
+ * Stamp the `replan_triggered_at` column on a slice. Used by triage-resolution
+ * when a user capture requests a replan so the dispatcher can detect the
+ * trigger via DB in addition to the on-disk REPLAN-TRIGGER.md marker.
+ */
+export function setSliceReplanTriggeredAt(milestoneId: string, sliceId: string, ts: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    "UPDATE slices SET replan_triggered_at = :ts WHERE milestone_id = :mid AND id = :sid",
+  ).run({ ":ts": ts, ":mid": milestoneId, ":sid": sliceId });
+}
+
+/**
+ * INSERT OR REPLACE a quality_gates row. Used by milestone-validation-gates.ts
+ * to persist milestone-level (MV*) gate outcomes after validate-milestone runs.
+ */
+export function upsertQualityGate(g: {
+  milestoneId: string;
+  sliceId: string;
+  gateId: string;
+  scope: string;
+  taskId: string;
+  status: string;
+  verdict: string;
+  rationale: string;
+  findings: string;
+  evaluatedAt: string;
+}): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `INSERT OR REPLACE INTO quality_gates
+     (milestone_id, slice_id, gate_id, scope, task_id, status, verdict, rationale, findings, evaluated_at)
+     VALUES (:mid, :sid, :gid, :scope, :tid, :status, :verdict, :rationale, :findings, :evaluated_at)`,
+  ).run({
+    ":mid": g.milestoneId,
+    ":sid": g.sliceId,
+    ":gid": g.gateId,
+    ":scope": g.scope,
+    ":tid": g.taskId,
+    ":status": g.status,
+    ":verdict": g.verdict,
+    ":rationale": g.rationale,
+    ":findings": g.findings,
+    ":evaluated_at": g.evaluatedAt,
+  });
+}
+
+/**
+ * Atomically replace all workflow state from a manifest. Lifted verbatim from
+ * workflow-manifest.ts so the single-writer invariant holds. Only touches
+ * engine tables + decisions. Does NOT modify artifacts or memories.
+ */
+export function restoreManifest(manifest: StateManifest): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  const db = currentDb;
+
+  transaction(() => {
+    // Clear engine tables (order matters for foreign-key-like consistency)
+    db.exec("DELETE FROM verification_evidence");
+    db.exec("DELETE FROM tasks");
+    db.exec("DELETE FROM slices");
+    db.exec("DELETE FROM milestones");
+    db.exec("DELETE FROM decisions WHERE 1=1");
+
+    // Restore milestones
+    const msStmt = db.prepare(
+      `INSERT INTO milestones (id, title, status, depends_on, created_at, completed_at,
+        vision, success_criteria, key_risks, proof_strategy,
+        verification_contract, verification_integration, verification_operational, verification_uat,
+        definition_of_done, requirement_coverage, boundary_map_markdown)
+       VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+    );
+    for (const m of manifest.milestones) {
+      msStmt.run(
+        m.id, m.title, m.status,
+        JSON.stringify(m.depends_on), m.created_at, m.completed_at,
+        m.vision, JSON.stringify(m.success_criteria), JSON.stringify(m.key_risks),
+        JSON.stringify(m.proof_strategy),
+        m.verification_contract, m.verification_integration, m.verification_operational, m.verification_uat,
+        JSON.stringify(m.definition_of_done), m.requirement_coverage, m.boundary_map_markdown,
+      );
+    }
+
+    // Restore slices (ADR-011 Phase 1: includes is_sketch + sketch_scope)
+    const slStmt = db.prepare(
+      `INSERT INTO slices (milestone_id, id, title, status, risk, depends, demo,
+        created_at, completed_at, full_summary_md, full_uat_md,
+        goal, success_criteria, proof_level, integration_closure, observability_impact,
+        sequence, replan_triggered_at, is_sketch, sketch_scope)
+       VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+    );
+    for (const s of manifest.slices) {
+      slStmt.run(
+        s.milestone_id, s.id, s.title, s.status, s.risk,
+        JSON.stringify(s.depends), s.demo,
+        s.created_at, s.completed_at, s.full_summary_md, s.full_uat_md,
+        s.goal, s.success_criteria, s.proof_level, s.integration_closure, s.observability_impact,
+        s.sequence, s.replan_triggered_at,
+        s.is_sketch ?? 0,
+        s.sketch_scope ?? "",
+      );
+    }
+
+    // Restore tasks (ADR-011 P2: includes blocker_source + escalation_* columns)
+    const tkStmt = db.prepare(
+      `INSERT INTO tasks (milestone_id, slice_id, id, title, status,
+        one_liner, narrative, verification_result, duration, completed_at,
+        blocker_discovered, deviations, known_issues, key_files, key_decisions,
+        full_summary_md, description, estimate, files, verify,
+        inputs, expected_output, observability_impact, sequence,
+        blocker_source, escalation_pending, escalation_awaiting_review,
+        escalation_artifact_path, escalation_override_applied_at)
+       VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+    );
+    for (const t of manifest.tasks) {
+      tkStmt.run(
+        t.milestone_id, t.slice_id, t.id, t.title, t.status,
+        t.one_liner, t.narrative, t.verification_result, t.duration, t.completed_at,
+        t.blocker_discovered ? 1 : 0, t.deviations, t.known_issues,
+        JSON.stringify(t.key_files), JSON.stringify(t.key_decisions),
+        t.full_summary_md, t.description, t.estimate, JSON.stringify(t.files), t.verify,
+        JSON.stringify(t.inputs), JSON.stringify(t.expected_output),
+        t.observability_impact, t.sequence,
+        t.blocker_source ?? "",
+        t.escalation_pending ?? 0,
+        t.escalation_awaiting_review ?? 0,
+        t.escalation_artifact_path ?? null,
+        t.escalation_override_applied_at ?? null,
+      );
+    }
+
+    // Restore decisions (ADR-011 P2: include source so escalation decisions survive)
+    const dcStmt = db.prepare(
+      `INSERT INTO decisions (seq, id, when_context, scope, decision, choice, rationale, revisable, made_by, source, superseded_by)
+       VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+    );
+    for (const d of manifest.decisions) {
+      dcStmt.run(d.seq, d.id, d.when_context, d.scope, d.decision, d.choice, d.rationale, d.revisable, d.made_by, d.source ?? "discussion", d.superseded_by);
+    }
+
+    // Restore verification evidence
+    const evStmt = db.prepare(
+      `INSERT INTO verification_evidence (task_id, slice_id, milestone_id, command, exit_code, verdict, duration_ms, created_at)
+       VALUES (?, ?, ?, ?, ?, ?, ?, ?)`,
+    );
+    for (const e of manifest.verification_evidence) {
+      evStmt.run(e.task_id, e.slice_id, e.milestone_id, e.command, e.exit_code, e.verdict, e.duration_ms, e.created_at);
+    }
+  });
+}
+
+// ─── Legacy markdown → DB bulk migration ─────────────────────────────────
+
+export interface LegacyMilestoneInsert {
+  id: string;
+  title: string;
+  status: string;
+}
+
+export interface LegacySliceInsert {
+  id: string;
+  milestoneId: string;
+  title: string;
+  status: string;
+  risk: string;
+  sequence: number;
+}
+
+export interface LegacyTaskInsert {
+  id: string;
+  sliceId: string;
+  milestoneId: string;
+  title: string;
+  status: string;
+  sequence: number;
+}
+
+/**
+ * Bulk delete + insert a legacy milestone hierarchy for markdown → DB migration.
+ * Used by workflow-migration.ts to populate engine tables from parsed ROADMAP/PLAN
+ * files. All operations run inside a single transaction.
+ */
+export function bulkInsertLegacyHierarchy(payload: {
+  milestones: LegacyMilestoneInsert[];
+  slices: LegacySliceInsert[];
+  tasks: LegacyTaskInsert[];
+  clearMilestoneIds: string[];
+  createdAt: string;
+}): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  const db = currentDb;
+  const { milestones, slices, tasks, clearMilestoneIds, createdAt } = payload;
+
+  if (clearMilestoneIds.length === 0) return;
+  const placeholders = clearMilestoneIds.map(() => "?").join(",");
+
+  transaction(() => {
+    db.prepare(`DELETE FROM tasks WHERE milestone_id IN (${placeholders})`).run(...clearMilestoneIds);
+    db.prepare(`DELETE FROM slices WHERE milestone_id IN (${placeholders})`).run(...clearMilestoneIds);
+    db.prepare(`DELETE FROM milestones WHERE id IN (${placeholders})`).run(...clearMilestoneIds);
+
+    const insertMilestone = db.prepare(
+      "INSERT INTO milestones (id, title, status, created_at) VALUES (?, ?, ?, ?)",
+    );
+    for (const m of milestones) {
+      insertMilestone.run(m.id, m.title, m.status, createdAt);
+    }
+
+    const insertSliceStmt = db.prepare(
+      "INSERT INTO slices (id, milestone_id, title, status, risk, depends, sequence, created_at) VALUES (?, ?, ?, ?, ?, ?, ?, ?)",
+    );
+    for (const s of slices) {
+      insertSliceStmt.run(s.id, s.milestoneId, s.title, s.status, s.risk, "[]", s.sequence, createdAt);
+    }
+
+    const insertTaskStmt = db.prepare(
+      "INSERT INTO tasks (id, slice_id, milestone_id, title, description, status, estimate, files, sequence) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)",
+    );
+    for (const t of tasks) {
+      insertTaskStmt.run(t.id, t.sliceId, t.milestoneId, t.title, "", t.status, "", "[]", t.sequence);
+    }
+  });
+}
+
+// ─── Memory store writers ────────────────────────────────────────────────
+// All memory writes go through gsd-db.ts so the single-writer invariant
+// holds. These are direct pass-throughs to the SQL previously in
+// memory-store.ts — same bindings, same behavior.
+
+export function insertMemoryRow(args: {
+  id: string;
+  category: string;
+  content: string;
+  confidence: number;
+  sourceUnitType: string | null;
+  sourceUnitId: string | null;
+  createdAt: string;
+  updatedAt: string;
+}): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `INSERT INTO memories (id, category, content, confidence, source_unit_type, source_unit_id, created_at, updated_at)
+     VALUES (:id, :category, :content, :confidence, :source_unit_type, :source_unit_id, :created_at, :updated_at)`,
+  ).run({
+    ":id": args.id,
+    ":category": args.category,
+    ":content": args.content,
+    ":confidence": args.confidence,
+    ":source_unit_type": args.sourceUnitType,
+    ":source_unit_id": args.sourceUnitId,
+    ":created_at": args.createdAt,
+    ":updated_at": args.updatedAt,
+  });
+}
+
+export function rewriteMemoryId(placeholderId: string, realId: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare("UPDATE memories SET id = :real_id WHERE id = :placeholder").run({
+    ":real_id": realId,
+    ":placeholder": placeholderId,
+  });
+}
+
+export function updateMemoryContentRow(
+  id: string,
+  content: string,
+  confidence: number | undefined,
+  updatedAt: string,
+): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  if (confidence != null) {
+    currentDb.prepare(
+      "UPDATE memories SET content = :content, confidence = :confidence, updated_at = :updated_at WHERE id = :id",
+    ).run({ ":content": content, ":confidence": confidence, ":updated_at": updatedAt, ":id": id });
+  } else {
+    currentDb.prepare(
+      "UPDATE memories SET content = :content, updated_at = :updated_at WHERE id = :id",
+    ).run({ ":content": content, ":updated_at": updatedAt, ":id": id });
+  }
+}
+
+export function incrementMemoryHitCount(id: string, updatedAt: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    "UPDATE memories SET hit_count = hit_count + 1, updated_at = :updated_at WHERE id = :id",
+  ).run({ ":updated_at": updatedAt, ":id": id });
+}
+
+export function supersedeMemoryRow(oldId: string, newId: string, updatedAt: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    "UPDATE memories SET superseded_by = :new_id, updated_at = :updated_at WHERE id = :old_id",
+  ).run({ ":new_id": newId, ":updated_at": updatedAt, ":old_id": oldId });
+}
+
+export function markMemoryUnitProcessed(
+  unitKey: string,
+  activityFile: string,
+  processedAt: string,
+): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `INSERT OR IGNORE INTO memory_processed_units (unit_key, activity_file, processed_at)
+     VALUES (:key, :file, :at)`,
+  ).run({ ":key": unitKey, ":file": activityFile, ":at": processedAt });
+}
+
+export function decayMemoriesBefore(cutoffTs: string, now: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE memories
+     SET confidence = MAX(0.1, confidence - 0.1), updated_at = :now
+     WHERE superseded_by IS NULL AND updated_at < :cutoff AND confidence > 0.1`,
+  ).run({ ":now": now, ":cutoff": cutoffTs });
+}
+
+export function supersedeLowestRankedMemories(limit: number, now: string): void {
+  if (!currentDb) throw new GSDError(GSD_STALE_STATE, "gsd-db: No database open");
+  currentDb.prepare(
+    `UPDATE memories SET superseded_by = 'CAP_EXCEEDED', updated_at = :now
+     WHERE id IN (
+       SELECT id FROM memories
+       WHERE superseded_by IS NULL
+       ORDER BY (confidence * (1.0 + hit_count * 0.1)) ASC
+       LIMIT :limit
+     )`,
+  ).run({ ":now": now, ":limit": limit });
+}