akm-cli 0.9.16 → 0.9.17-alpha.10

package/dist/commands/improve/preparation.js

@@ -1,271 +1,140 @@
 // This Source Code Form is subject to the terms of the Mozilla Public
 // License, v. 2.0. If a copy of the MPL was not distributed with this
 // file, You can obtain one at https://mozilla.org/MPL/2.0/.
+/**
+ * The improve preparation stage: consolidation and session extraction (which
+ * run before the loop), memory cleanup, structural validation, and candidate
+ * selection for the reflect/distill loop.
+ *
+ * Candidate selection reads the improve ledger plus one set of signals: a ref
+ * is eligible for a source when feedback newer than its last attempt landed and
+ * no ledger window holds it. Refs without recent feedback can still be picked
+ * by the fallback lanes (proactive maintenance, high salience); the survivors
+ * are ranked by salience, checked on disk and capped. A plan-only run
+ * evaluates the same selectors against read snapshots and writes nothing.
+ */
 import fs from "node:fs";
 import path from "node:path";
-import { makeBundleRef } from "../../core/asset/asset-ref.js";
 import { parseFrontmatter } from "../../core/asset/frontmatter.js";
-import { typeNameFromConceptId } from "../../core/asset/resolve-ref.js";
 import { daysToMs } from "../../core/common.js";
 import { loadConfig } from "../../core/config/config.js";
 import { ConfigError, rethrowIfTestIsolationError } from "../../core/errors.js";
 import { appendEvent, readEvents } from "../../core/events.js";
-import { openStateDatabase, withStateDb } from "../../core/state-db.js";
+import { withStateDb } from "../../core/state-db.js";
 import { info, warn } from "../../core/warn.js";
-import { countUsageEventsByType } from "../../indexer/usage/usage-events.js";
+import { countUsageEventsByType, USAGE_EVENT_RETENTION_DAYS } from "../../indexer/usage/usage-events.js";
 import { getAvailableHarnesses } from "../../integrations/session-logs/index.js";
-import { withLlmStage } from "../../llm/usage-telemetry.js";
-import { closeDatabase, openExistingDatabase, openReadonlyExistingDatabase, } from "../../storage/repositories/index-connection.js";
 import { getZeroResultSearches } from "../../storage/repositories/index-entries-repository.js";
 import { getRetrievalCounts } from "../../storage/repositories/index-utility-repository.js";
 import { listStateProposals } from "../../storage/repositories/proposals-repository.js";
 import { akmLint } from "../lint/index.js";
 import { runSchemaRepairPass } from "../sources/schema-repair.js";
 import { isAutonomyLaneAllowed } from "./autonomy-gate.js";
-import { akmConsolidate, inspectConsolidationPool } from "./consolidate.js";
+import { akmConsolidate, inspectConsolidationPool, loadExistingKnowledgeBodyHashes, makeConsolidateResult, } from "./consolidate.js";
 import { computeSafeChunkSize, DEFAULT_CONTEXT_LENGTH_TOKENS } from "./consolidate/chunking.js";
-// Eligibility / candidate-selection predicates live in ./eligibility.
-import { buildLatestFeedbackTsMap, buildLatestProposalTsMap, buildUtilityMap, dedupeRefs, findAssetFilePath, isDistillCandidateRef, isLessonCandidate, isSignalDeltaEligible, resolveImproveScope, } from "./eligibility.js";
+import { assetTypeOf, buildUtilityMap, dedupeRefs, findAssetFilePath, isDistillCandidateRef, isLessonCandidate, resolveImproveScope, withIndexDb, } from "./eligibility.js";
 import { akmExtract, countNewExtractCandidates } from "./extract.js";
-import { computeValenceScore, FEEDBACK_WEIGHT, UTILITY_WEIGHT } from "./feedback-valence.js";
+import { computeValenceScore } from "./feedback-valence.js";
+import { isLedgerBlocked, lastAttemptByRef, ledgerRowFor, loadLedgerSnapshot, stateKey, } from "./ledger.js";
 import { applyMemoryCleanup } from "./memory/memory-improve.js";
-import { computeProxyAdequacy, getAllAssetOutcomes, getAssetOutcome, getOutcomeScoresByRef, OUTCOME_SCORE_MAX, outcomeScoreToSalience, projectAssetOutcome, updateAssetOutcome, } from "./outcome-loop.js";
+import { getAllAssetOutcomes, getAssetOutcome, getOutcomeScoresByRef, OUTCOME_SCORE_MAX, outcomeScoreToSalience, projectAssetOutcome, updateAssetOutcome, } from "./outcome-loop.js";
 import { projectMemoryCleanup, selectEffectiveImproveRefs } from "./planner.js";
 import { DEFAULT_DUE_DAYS, DEFAULT_MAX_PER_RUN, selectProactiveMaintenanceRefs } from "./proactive-maintenance.js";
-import { buildRankChangeReport, computeSalience, getAllRankScores, getAssetSalience, getLastUseMsByRef, isContentEncodingRow, SALIENCE_NO_OP_DAMPEN_FACTOR, SALIENCE_NO_OP_DAMPEN_THRESHOLD, upsertAssetSalience, } from "./salience.js";
-import { bareImproveRef, improveStateReadRefs } from "./source-identity.js";
-function readAssetSalienceForImproveRef(db, ref, itemRef) {
-    for (const key of improveStateReadRefs(ref, itemRef)) {
-        const row = getAssetSalience(db, key);
-        if (row)
-            return row;
+import { isInRetrievalScope, loadRetrievalScope } from "./retrieval-scope.js";
+import { computeSalience, getAssetSalience, getLastUseMsByRef, isContentEncodingRow, SALIENCE_NO_OP_DAMPEN_FACTOR, SALIENCE_NO_OP_DAMPEN_THRESHOLD, upsertAssetSalience, } from "./salience.js";
+import { attributeStage, errMessage } from "./stage.js";
+/** The candidate's durable state key (salience, outcome, ledger). */
+const keyOf = (r) => stateKey(r.ref, r.itemRef);
+/**
+ * Run `fn` against the run's state.db (its long-lived handle when there is
+ * one). A plan-only run without a handle reads nothing. Best-effort.
+ */
+function withRunState(eventsCtx, persist, fn) {
+    if (!persist && !eventsCtx?.db)
+        return undefined;
+    try {
+        return withStateDb(fn, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
+    }
+    catch (err) {
+        rethrowIfTestIsolationError(err);
+        return undefined;
     }
-    return undefined;
-}
-function readConsecutiveNoOpsForImproveRef(db, ref, itemRef) {
-    return readAssetSalienceForImproveRef(db, ref, itemRef)?.consecutive_no_ops ?? 0;
 }
-// ── Durable-state write keys ──────────────────────────────────────────────────
-//
-// The durable improve-state writers (salience + outcome) key by the resolved
-// index entry's `item_ref` (`<bundle>//<conceptId>`) when the planner resolved
-// one (`ImproveEligibleRef.itemRef`). Both writers share one key expression,
-// `itemRefByRef.get(ref) ?? ref`.
-//
-// A direct `--scope <ref>` candidate does not flow through
-// `collectEligibleRefsFromIndex`, so its conceptId `ref` is the write key.
-//
-// itemRefByRef is `ref → item_ref | undefined`, built once per pass from the
-// candidate set.
-/** `ref → item_ref | undefined` for a run's candidate set. */
-function buildItemRefByRef(refs) {
-    const m = new Map();
-    for (const r of refs)
-        m.set(r.ref, r.itemRef);
-    return m;
+function fileSize(filePath) {
+    if (!filePath)
+        return undefined;
+    try {
+        return fs.statSync(filePath).size;
+    }
+    catch {
+        return undefined;
+    }
 }
-/** Durable `asset_salience` write key: the entry's item_ref, else its conceptId `ref` (scope-ref fallback). */
-function salienceWriteKey(ref, itemRefByRef) {
-    return itemRefByRef.get(ref) ?? ref;
+/** `{ key: value }` for each key `source` defines. */
+export function pickDefined(source, keys) {
+    const out = {};
+    for (const key of keys)
+        if (source?.[key] !== undefined)
+            out[key] = source[key];
+    return out;
 }
-/** Durable `asset_outcome` write key: the entry's item_ref, else its conceptId `ref` (scope-ref fallback). */
-function outcomeWriteKey(ref, itemRefByRef) {
-    return itemRefByRef.get(ref) ?? ref;
+export const CONSOLIDATION_CONFIG_KEYS = ["enabled", "minPoolSize", "limit", "maxChunkSize"];
+/** Emit an aggregate `improve_skipped` row (never one per ref). */
+export function recordImproveSkip(eventsCtx, ref, metadata) {
+    appendEvent({ eventType: "improve_skipped", ref, metadata }, eventsCtx);
 }
-/** Resolve an AKM asset type from a short or bundle-qualified conceptId. */
-function assetTypeOf(ref) {
-    const tail = ref.includes("//") ? ref.slice(ref.indexOf("//") + 2) : ref;
-    return typeNameFromConceptId(tail)?.type ?? "";
+/** Per-originator rolling error windows (3 each) shown to later prompts as patterns to avoid. */
+export function pushRecentError(recentErrors, originator, msg) {
+    const window = recentErrors[originator] ?? [];
+    window.push(msg);
+    if (window.length > 3)
+        window.shift();
+    recentErrors[originator] = window;
 }
+// ── Consolidation ────────────────────────────────────────────────────────────
 /**
- * Evaluate the consolidation gate flags (volume trigger, #551 pool-delta
- * cooldown, profile disable, #553 min-pool-size) up front, before any LLM call.
- * Extracted verbatim from `runConsolidationPass` — logic is byte-identical.
+ * The consolidation gates and pool, with no model call: the profile toggle,
+ * `minPoolSize` (not for a named strategy or ref scope, nor once the pool is
+ * over the 100-memory volume trigger), and the ledger delta (every memory
+ * judged recently and unchanged since means nothing to do).
  */
-function evaluateConsolidationEligibility(args) {
-    const { options, primaryStashDir, memorySummary, improveProfile, resolvedPlan, eventsCtx } = args;
-    const MEMORY_VOLUME_THRESHOLD = options.memoryVolumeConsolidationThreshold ?? 100;
-    const hasLlm = resolvedPlan.processes.consolidate.runner !== null;
-    const volumeTriggered = typeof memorySummary.eligible === "number" && memorySummary.eligible > MEMORY_VOLUME_THRESHOLD && hasLlm;
-    // 0.8.0 pool-delta gate for consolidate: re-eligible iff at least one
-    // memory file has been updated since the most recent successful
-    // consolidate_completed event. Time-based cooldowns produced the same
-    // synchronised-wave failure mode the reflect/distill cooldowns did; the
-    // pool-delta gate ties consolidation to actual work-to-do.
-    const sourceName = options.sourceName ?? options.writeTarget?.source.name ?? options.config?.defaultBundle ?? "stash";
-    const recentConsolidations = readEvents({ type: "consolidate_completed" }, eventsCtx);
-    const lastConsolidation = recentConsolidations.events
-        .filter((e) => e.metadata?.source === sourceName && Number(e.metadata?.processed) > 0)
-        .sort((a, b) => new Date(b.ts ?? 0).getTime() - new Date(a.ts ?? 0).getTime())[0];
-    const lastConsolidateTs = typeof lastConsolidation?.metadata?.completedThrough === "string"
-        ? lastConsolidation.metadata.completedThrough
-        : lastConsolidation?.ts;
-    // #551 smarter gate: build the set of memory asset paths whose only delta
-    // since the last consolidate is their OWN promotion. Those files
-    // have not had a full improve cycle to settle, so they offer no merge /
-    // contradiction candidates yet — excluding them stops the gate firing on
-    // freshly-promoted single-source memories. We read `promoted` events emitted
-    // after the last consolidate; each carries the written `assetPath`.
-    const promotedSinceConsolidate = (() => {
-        const paths = new Set();
-        try {
-            const promoted = readEvents({
-                type: "promoted",
-                ...(lastConsolidateTs ? { since: lastConsolidateTs } : {}),
-            }, eventsCtx).events;
-            for (const e of promoted) {
-                const ap = e.metadata?.assetPath;
-                if (typeof ap === "string" && ap.length > 0)
-                    paths.add(path.resolve(ap));
-            }
-        }
-        catch {
-            // best-effort: if the events query fails, fall back to no exclusions
-            // (preserves pre-#551 behaviour rather than over-skipping).
-        }
-        return paths;
-    })();
-    // Pool-delta: any memory file with mtime > lastConsolidateTs flags work to do,
-    // EXCEPT files whose only post-consolidate change was their own promotion.
-    // Using file mtime keeps this query DB-free and matches what the indexer
-    // already uses as the canonical `memory.updated_at` proxy.
-    //
-    // Bootstrap: when no successful consolidate_completed event has ever been
-    // recorded, we cannot evaluate the pool-delta — treat as eligible so a
-    // fresh stash runs consolidate once before the steady-state gate kicks in.
-    const memoryUpdatedAfterLastConsolidate = (() => {
-        if (volumeTriggered)
-            return true; // volume override forces the run regardless.
-        if (!lastConsolidateTs)
-            return true; // bootstrap path: never consolidated.
-        if (!primaryStashDir)
-            return false;
-        const memoriesDir = path.join(primaryStashDir, "memories");
-        if (!fs.existsSync(memoriesDir))
-            return false;
-        try {
-            const pending = [memoriesDir];
-            while (pending.length > 0) {
-                const current = pending.pop();
-                for (const entry of fs.readdirSync(current, { withFileTypes: true })) {
-                    const filePath = path.join(current, entry.name);
-                    if (entry.isDirectory()) {
-                        pending.push(filePath);
-                        continue;
-                    }
-                    if (!entry.isFile() || !entry.name.endsWith(".md"))
-                        continue;
-                    if (promotedSinceConsolidate.has(path.resolve(filePath)))
-                        continue;
-                    try {
-                        if (fs.statSync(filePath).mtime.toISOString() > lastConsolidateTs)
-                            return true;
-                    }
-                    catch {
-                        // Ignore files that disappear during the scan.
-                    }
-                }
-            }
-            return false;
-        }
-        catch {
-            return false;
-        }
-    })();
-    const consolidationOnCooldown = !volumeTriggered && !memoryUpdatedAfterLastConsolidate;
-    // Profile gate: if profile explicitly disables consolidate, skip the entire pass.
-    const consolidateDisabledByProfile = improveProfile?.processes?.consolidate?.enabled === false;
-    // #553 minPoolSize guard: skip consolidation when the eligible memory pool is
-    // below a minimum size, rather than spending an LLM pass on a handful of
-    // memories. This is an INDEPENDENT skip condition from #551's mtime pool-delta
-    // gate — either can skip. Default 0 (disabled) — every built-in strategy
-    // used to ship 500, which meant `akm improve --strategy consolidate`, typed
-    // by a human, silently did nothing on almost every real install. Evaluated
-    // against the eligible-pool count BEFORE entering the LLM loop so a skip
-    // costs ZERO LLM calls when an operator opts back into a floor.
-    const CONSOLIDATE_DEFAULT_MIN_POOL_SIZE = 0;
-    const configuredMinPoolSize = improveProfile?.processes?.consolidate?.minPoolSize;
-    const minPoolSize = typeof configuredMinPoolSize === "number" ? configuredMinPoolSize : CONSOLIDATE_DEFAULT_MIN_POOL_SIZE;
-    const eligiblePoolSize = typeof memorySummary.eligible === "number" ? memorySummary.eligible : 0;
-    const userNamedStrategyOrScope = options.strategy !== undefined || resolveImproveScope(options.scope).mode === "ref";
-    // volumeTriggered means the pool already exceeds the volume threshold (100),
-    // so a force-triggered run never trips the pool-size guard. The guard only
-    // engages when minPoolSize > 0 and the eligible pool is strictly below it.
-    const poolBelowMinSize = !volumeTriggered && !userNamedStrategyOrScope && minPoolSize > 0 && eligiblePoolSize < minPoolSize;
-    return {
-        volumeTriggered,
-        consolidationOnCooldown,
-        consolidateDisabledByProfile,
-        poolBelowMinSize,
-        eligiblePoolSize,
-        minPoolSize,
-        ...(lastConsolidateTs ? { lastConsolidationTs: lastConsolidateTs } : {}),
-    };
-}
-/** Build the no-dispatch consolidation projection consumed by dry and live. */
 function planConsolidationPass(args) {
-    const { options, primaryStashDir, memorySummary, improveProfile, resolvedPlan, eventsCtx } = args;
-    const processConfig = improveProfile?.processes?.consolidate;
-    const eligibility = evaluateConsolidationEligibility({
-        options,
-        primaryStashDir,
-        memorySummary,
-        improveProfile,
-        resolvedPlan,
-        eventsCtx,
-    });
-    const effectiveOptions = {
-        ...options.consolidateOptions,
-        config: options.config,
-        stashDir: options.stashDir,
-        writeTarget: options.writeTarget,
-        limit: processConfig?.limit,
-        incrementalSince: processConfig?.incrementalSince,
-        neighborsPerChanged: processConfig?.neighborsPerChanged,
-        maxChunkSize: processConfig?.maxChunkSize,
-    };
-    const poolWarnings = [];
+    const { options, primaryStashDir, memorySummary, resolvedPlan } = args;
+    const processConfig = args.improveProfile?.processes?.consolidate;
+    const volumeTriggered = memorySummary.eligible > 100 && resolvedPlan.processes.consolidate.runner !== null;
+    const minPoolSize = typeof processConfig?.minPoolSize === "number" ? processConfig.minPoolSize : 0;
+    const eligiblePoolSize = typeof memorySummary.eligible === "number" ? memorySummary.eligible : 0;
+    const userNamed = options.strategy !== undefined || resolveImproveScope(options.scope).mode === "ref";
+    const poolBelowMinSize = !volumeTriggered && !userNamed && minPoolSize > 0 && eligiblePoolSize < minPoolSize;
     const pool = primaryStashDir
-        ? inspectConsolidationPool(effectiveOptions, primaryStashDir, poolWarnings, {
-            readOnly: eventsCtx?.readOnly === true,
-        })
-        : { poolSize: 0, candidatePoolSize: 0, dedupPoolSize: 0, memories: [] };
-    // #800/#957 round 3 — a credential-unavailable consolidate engine still
-    // resolved a context length structurally; read it off the `engineUnavailable`
-    // entry instead of falling back to the generic default, so a dry-run
-    // preview reflects the real engine even when its credential isn't
-    // materialized here.
-    const consolidateUnavailable = resolvedPlan.engineUnavailable.find((item) => item.process === "consolidate");
+        ? inspectConsolidationPool({
+            config: options.config,
+            stashDir: options.stashDir,
+            writeTarget: options.writeTarget,
+            target: options.target,
+            limit: processConfig?.limit,
+            maxChunkSize: processConfig?.maxChunkSize,
+        }, primaryStashDir, [], args.existingKnowledgeBodyHashes ?? loadExistingKnowledgeBodyHashes(primaryStashDir), { readOnly: args.eventsCtx?.readOnly === true })
+        : { poolSize: 0, candidatePoolSize: 0, judgedUnchanged: 0 };
+    // A credential-unavailable engine still resolved its context length.
+    const unavailable = resolvedPlan.engineUnavailable.find((item) => item.process === "consolidate");
     const chunkSize = computeSafeChunkSize(resolvedPlan.processes.consolidate.runner?.connection.contextLength ??
-        consolidateUnavailable?.contextLength ??
+        unavailable?.contextLength ??
         DEFAULT_CONTEXT_LENGTH_TOKENS, 500, processConfig?.maxChunkSize);
-    const profilePassed = !eligibility.consolidateDisabledByProfile;
-    const minimumPoolPassed = !eligibility.poolBelowMinSize;
-    const deltaPassed = !eligibility.consolidationOnCooldown;
-    const nonEmptyPool = pool.candidatePoolSize > 0;
-    const wouldRun = profilePassed && minimumPoolPassed && deltaPassed && nonEmptyPool;
-    const reason = !profilePassed
-        ? "disabled by improve profile"
-        : !minimumPoolPassed
-            ? `pool ${eligibility.eligiblePoolSize} is below minPoolSize ${eligibility.minPoolSize}`
-            : !deltaPassed
-                ? "no memory updates since the last completed consolidation"
-                : !nonEmptyPool
-                    ? "candidate pool is empty after narrowing"
-                    : "all consolidation gates pass";
+    const profilePassed = processConfig?.enabled !== false;
+    const deltaPassed = pool.candidatePoolSize > 0 || pool.judgedUnchanged === 0;
+    const wouldRun = profilePassed && !poolBelowMinSize && deltaPassed && pool.candidatePoolSize > 0;
+    const belowMin = `pool ${eligiblePoolSize} is below minPoolSize ${minPoolSize}`;
+    const unchanged = "every memory was judged recently and is unchanged since";
     return {
-        eligibility,
+        poolBelowMinSize,
+        eligiblePoolSize,
+        minPoolSize,
         plan: {
-            configured: {
-                ...(processConfig?.enabled !== undefined ? { enabled: processConfig.enabled } : {}),
-                ...(processConfig?.minPoolSize !== undefined ? { minPoolSize: processConfig.minPoolSize } : {}),
-                ...(processConfig?.limit !== undefined ? { limit: processConfig.limit } : {}),
-                ...(processConfig?.maxChunkSize !== undefined ? { maxChunkSize: processConfig.maxChunkSize } : {}),
-                ...(processConfig?.incrementalSince !== undefined ? { incrementalSince: processConfig.incrementalSince } : {}),
-            },
+            configured: pickDefined(processConfig, CONSOLIDATION_CONFIG_KEYS),
             effective: {
                 enabled: profilePassed,
-                minPoolSize: eligibility.minPoolSize,
+                minPoolSize,
                 ...(processConfig?.limit !== undefined ? { limit: processConfig.limit } : {}),
                 chunkSize,
             },
@@ -277,314 +146,180 @@ function planConsolidationPass(args) {
                     reason: profilePassed ? "consolidation enabled" : "disabled by improve profile",
                 },
                 minimumPool: {
-                    passed: minimumPoolPassed,
-                    reason: minimumPoolPassed
-                        ? `pool satisfies minPoolSize ${eligibility.minPoolSize}`
-                        : `pool ${eligibility.eligiblePoolSize} is below minPoolSize ${eligibility.minPoolSize}`,
+                    passed: !poolBelowMinSize,
+                    reason: poolBelowMinSize ? belowMin : `pool satisfies minPoolSize ${minPoolSize}`,
                 },
                 delta: {
                     passed: deltaPassed,
-                    reason: deltaPassed ? "memory pool has work" : "no updates since the last completed consolidation",
+                    reason: !deltaPassed
+                        ? unchanged
+                        : pool.judgedUnchanged > 0
+                            ? `${pool.judgedUnchanged} recently judged, unchanged memories skipped`
+                            : "no memory was judged recently",
                 },
             },
             wouldRun,
-            reason,
+            reason: !profilePassed
+                ? "disabled by improve profile"
+                : poolBelowMinSize
+                    ? belowMin
+                    : !deltaPassed
+                        ? unchanged
+                        : pool.candidatePoolSize === 0
+                            ? "candidate pool is empty after narrowing"
+                            : "all consolidation gates pass",
             estimatedChunks: wouldRun ? Math.ceil(pool.candidatePoolSize / chunkSize) : 0,
         },
     };
 }
-export async function runConsolidationPass(args) {
-    const { options, primaryStashDir, memorySummary, improveProfile, resolvedPlan, eventsCtx, budgetSignal, runBudgetMs, } = args;
-    const baseConfig = options.config ?? loadConfig();
-    const consolidationConfig = baseConfig;
-    const planned = planConsolidationPass({
-        options,
-        primaryStashDir,
-        memorySummary,
-        improveProfile,
-        resolvedPlan,
-        eventsCtx,
-    });
-    const { volumeTriggered, consolidationOnCooldown, consolidateDisabledByProfile, poolBelowMinSize, eligiblePoolSize, minPoolSize, lastConsolidationTs, } = planned.eligibility;
-    let consolidation = {
-        schemaVersion: 1,
-        ok: true,
-        shape: "consolidate-result",
-        dryRun: false,
-        previewOnly: false,
-        target: "",
-        processed: 0,
-        merged: 0,
-        deleted: 0,
-        promoted: [],
-        contradicted: 0,
-        warnings: [],
-        durationMs: 0,
-    };
-    if (consolidateDisabledByProfile) {
+async function runConsolidationPass(args) {
+    const { options, primaryStashDir, improveProfile, resolvedPlan, eventsCtx } = args;
+    // Walked once and shared with the live pass.
+    const existingKnowledgeBodyHashes = primaryStashDir ? loadExistingKnowledgeBodyHashes(primaryStashDir) : undefined;
+    const planned = planConsolidationPass({ ...args, existingKnowledgeBodyHashes });
+    const processConfig = improveProfile?.processes?.consolidate;
+    let consolidation = makeConsolidateResult({ target: "", durationMs: 0 });
+    if (!planned.plan.gates.profile.passed) {
         info("[improve] consolidation skipped (disabled by improve profile)");
     }
-    else if (poolBelowMinSize) {
-        // #553: eligible pool below the configured minimum — skip with zero LLM
-        // calls. Reuse the #551 `improve_skipped` emission path so health surfaces
-        // it via the dynamic skipReasons aggregation under `pool_below_min_size`.
-        appendEvent({
-            eventType: "improve_skipped",
-            ref: "memories/_consolidation",
-            metadata: {
-                reason: "pool_below_min_size",
-                poolSize: eligiblePoolSize,
-                minPoolSize,
-            },
-        }, eventsCtx);
-        info(`[improve] consolidation skipped (pool ${eligiblePoolSize} < minPoolSize ${minPoolSize})`);
+    else if (planned.poolBelowMinSize) {
+        recordImproveSkip(eventsCtx, "memories/_consolidation", {
+            reason: "pool_below_min_size",
+            poolSize: planned.eligiblePoolSize,
+            minPoolSize: planned.minPoolSize,
+        });
+        info(`[improve] consolidation skipped (pool ${planned.eligiblePoolSize} < minPoolSize ${planned.minPoolSize})`);
     }
-    else if (!consolidationOnCooldown) {
-        const consolidationStartedAt = new Date().toISOString();
-        consolidation = await withLlmStage("consolidate", () => akmConsolidate({
-            ...options.consolidateOptions,
-            config: consolidationConfig,
+    else if (!planned.plan.gates.delta.passed) {
+        recordImproveSkip(eventsCtx, "memories/_consolidation", { reason: "consolidation_no_memory_updates" });
+        info("[improve] consolidation skipped (every memory was judged recently and is unchanged)");
+    }
+    else {
+        consolidation = await attributeStage(resolvedPlan, "consolidate", () => akmConsolidate({
+            target: options.target,
+            ...(options.writeTarget ? { writeTarget: options.writeTarget } : {}),
+            config: options.config ?? loadConfig(),
             dryRun: options.dryRun ?? false,
             stashDir: options.stashDir,
-            // Active profile for this improve run — lets consolidate's secondary
-            // process-config reads honor `--profile <name>` instead of `default`.
             improveProfile,
             llmRunner: resolvedPlan.processes.consolidate.runner,
-            autoTriggered: volumeTriggered,
-            // Tie consolidate proposals back to this improve invocation so
-            // accept-rate-per-run aggregation works. Mirrors reflect/propose/extract.
+            existingKnowledgeBodyHashes,
             sourceRun: `consolidate-${Date.now()}`,
-            // Pass profile-configured options. incrementalSince narrows the pool to
-            // recently-changed memories + graph neighbours — use this for frequent
-            // passes (quick-shredder). Leave absent in the nightly default profile for
-            // a full-pool sweep that catches stale-but-unmerged duplicates.
-            limit: improveProfile?.processes?.consolidate?.limit,
-            incrementalSince: improveProfile?.processes?.consolidate?.incrementalSince,
-            neighborsPerChanged: improveProfile?.processes?.consolidate?.neighborsPerChanged,
-            maxChunkSize: improveProfile?.processes?.consolidate?.maxChunkSize,
-            // WS-3a: forward budget signal for graceful abort on timeout, and pass
-            // the profile's p90 estimate for cold-start budget reduction.
-            signal: budgetSignal,
-            p90ChunkSecondsDefault: improveProfile?.processes?.consolidate?.p90ChunkSecondsDefault,
-            // WS-5: pass total run budget so perfTelemetry.estimatedBudgetFractionUsed
-            // can flag when consolidation alone exceeded the budget.
-            runBudgetMs,
-        }), { engine: resolvedPlan.processes.consolidate.runner?.engine, process: "consolidate" });
-        const sourceName = options.sourceName ?? options.writeTarget?.source.name ?? baseConfig.defaultBundle ?? "stash";
-        const complete = (consolidation.failedChunks ?? 0) === 0 &&
-            (consolidation.failedChunkMemories ?? 0) === 0 &&
-            (consolidation.failedPromotions ?? 0) === 0 &&
-            (consolidation.deferredMemories ?? 0) === 0;
-        const hasUnappliedAdvisoryOperations = consolidation.planned?.some((op) => op.op !== "promote") ?? false;
-        if (consolidation.ok &&
-            !consolidation.dryRun &&
-            complete &&
-            !hasUnappliedAdvisoryOperations &&
-            consolidation.processed > 0) {
-            appendEvent({
-                eventType: "consolidate_completed",
-                ref: makeBundleRef(sourceName, "memories/_consolidation"),
-                metadata: {
-                    processed: consolidation.processed,
-                    source: sourceName,
-                    completedThrough: consolidationStartedAt,
-                    merged: consolidation.merged,
-                    deleted: consolidation.deleted,
-                    contradicted: consolidation.contradicted,
-                    failedChunks: consolidation.failedChunks ?? 0,
-                    durationMs: consolidation.durationMs,
-                },
-            }, eventsCtx);
-        }
+            limit: processConfig?.limit,
+            maxChunkSize: processConfig?.maxChunkSize,
+            signal: args.budgetSignal,
+            p90ChunkSecondsDefault: processConfig?.p90ChunkSecondsDefault,
+        }));
     }
-    else {
-        appendEvent({
-            eventType: "improve_skipped",
-            ref: "memories/_consolidation",
-            metadata: {
-                reason: "consolidation_no_memory_updates",
-                lastEventTs: lastConsolidationTs ?? null,
-            },
-        }, eventsCtx);
-        info("[improve] consolidation skipped (no memory updates since last run)");
-    }
-    // D9: track whether consolidation wrote any data so graph extraction can reindex if needed
-    const consolidationRan = !consolidateDisabledByProfile &&
-        !poolBelowMinSize &&
-        !consolidationOnCooldown &&
-        !consolidation.previewOnly &&
-        consolidation.processed > 0;
-    return { consolidation, consolidationRan, plan: planned.plan };
+    return { consolidation, plan: planned.plan };
 }
-/**
- * Evaluate the exact pre-dispatch extract gates. Live execution consumes this
- * snapshot and dry-run only reports it, so neither path reconstructs the
- * selector independently.
- */
-function inspectExtractPass(args) {
-    const { options, improveProfile, resolvedPlan, eventsCtx, readOnly } = args;
+/** The extract gates, evaluated once for both the live pass and the dry-run report. */
+function inspectExtractPass(args, readOnly) {
+    const { options, improveProfile, resolvedPlan, eventsCtx } = args;
     const enabled = resolvedPlan.processes.extract.enabled;
     const hasRunner = resolvedPlan.processes.extract.runner?.engine !== undefined;
-    const availableHarnesses = (options.extractHarnesses ?? getAvailableHarnesses()).filter((harness) => harness.isAvailable());
-    const configuredMinNewSessions = improveProfile.processes?.extract?.minNewSessions;
-    const minNewSessions = typeof configuredMinNewSessions === "number" ? configuredMinNewSessions : 0;
+    const availableHarnesses = (options.extractHarnesses ?? getAvailableHarnesses()).filter((h) => h.isAvailable());
+    const configured = improveProfile.processes?.extract?.minNewSessions;
+    const minNewSessions = typeof configured === "number" ? configured : 0;
     let newCandidateCount;
     if (enabled && hasRunner && availableHarnesses.length > 0 && minNewSessions > 0) {
-        const countFn = options.extractCandidateCountFn ?? countNewExtractCandidates;
-        newCandidateCount = countFn(options.config ?? loadConfig(), {
+        const defaultSince = improveProfile.processes?.extract?.defaultSince;
+        newCandidateCount = (options.extractCandidateCountFn ?? countNewExtractCandidates)(options.config ?? loadConfig(), {
             harnesses: availableHarnesses,
             improveProfile,
-            ...(improveProfile.processes?.extract?.defaultSince
-                ? { since: improveProfile.processes.extract.defaultSince }
-                : {}),
+            ...(defaultSince ? { since: defaultSince } : {}),
             ...(eventsCtx?.db ? { stateDb: eventsCtx.db } : {}),
             ...(!readOnly && eventsCtx?.dbPath ? { stateDbPath: eventsCtx.dbPath } : {}),
             ...(readOnly ? { readOnly: true } : {}),
         });
     }
     const belowMinNewSessions = minNewSessions > 0 && newCandidateCount !== undefined && newCandidateCount < minNewSessions;
-    const wouldRun = enabled && hasRunner && availableHarnesses.length > 0 && !belowMinNewSessions;
-    const reason = !enabled
-        ? "disabled"
-        : !hasRunner
-            ? "enabled but no runner is resolved"
-            : availableHarnesses.length === 0
-                ? "enabled but no session-log harness is available"
-                : belowMinNewSessions
-                    ? `${newCandidateCount ?? 0} new sessions is below minNewSessions ${minNewSessions}`
-                    : minNewSessions > 0
-                        ? `${newCandidateCount ?? 0} new sessions satisfies minNewSessions ${minNewSessions}`
-                        : `enabled with ${availableHarnesses.length} available session-log harness(es); minNewSessions is disabled`;
+    const count = `${newCandidateCount ?? 0} new sessions`;
     return {
         availableHarnesses,
         minNewSessions,
         ...(newCandidateCount !== undefined ? { newCandidateCount } : {}),
         belowMinNewSessions,
-        wouldRun,
-        reason,
+        wouldRun: enabled && hasRunner && availableHarnesses.length > 0 && !belowMinNewSessions,
+        reason: !enabled
+            ? "disabled"
+            : !hasRunner
+                ? "enabled but no runner is resolved"
+                : availableHarnesses.length === 0
+                    ? "enabled but no session-log harness is available"
+                    : belowMinNewSessions
+                        ? `${count} is below minNewSessions ${minNewSessions}`
+                        : minNewSessions > 0
+                            ? `${count} satisfies minNewSessions ${minNewSessions}`
+                            : `enabled with ${availableHarnesses.length} available session-log harness(es); minNewSessions is disabled`,
     };
 }
 /**
- * Phase 0.4 — session-extract pass. Reads native session files through the
- * SessionLogHarness registry, and asks a bounded LLM for candidate proposals.
- * Failures are non-fatal (collected into `warnings`). Returns the extract
- * results + any warnings collected along the way.
+ * One `akmExtract` per available harness under the strategy's frozen plan. A
+ * harness that throws is a warning; the `minNewSessions` gate skips the whole
+ * pass with no model call.
  */
-async function runSessionExtractPass(args) {
-    const { options, primaryStashDir, improveProfile, resolvedPlan, eventsCtx, budgetSignal } = args;
+async function runSessionExtractPass(args, plan) {
+    const { options, primaryStashDir, resolvedPlan, eventsCtx } = args;
     const warnings = [];
-    // Phase 0.4 — session-extract pass.
-    //
-    // Reads native session files (claude JSONL, opencode storage tree)
-    // through the SessionLogHarness registry, pre-filters noise, and asks a
-    // bounded in-tree LLM to produce candidate memory/lesson/knowledge
-    // proposals for content the agent did NOT preserve via inline `akm remember`
-    // / `akm feedback` invocations. Replaces the akm-plugin session-checkpoint
-    // hook with an on-demand pull pipeline.
-    //
-    // Runs only when the ACTIVE strategy resolves
-    // `processes.extract.enabled: true` (#593: the gate respects the resolved
-    // improve strategy, not just the hardcoded `default` path the legacy feature
-    // flag read). Shipped `default` and `frequent` strategies leave this off.
-    // Each available harness gets one call with the default --since window;
-    // already-seen sessions (tracked in state.db.extract_sessions_seen) are
-    // skipped automatically so re-runs don't burn LLM calls on unchanged data.
-    //
-    // Failures are non-fatal — one harness throwing doesn't abort improve.
-    // The extract envelope's own `warnings` field surfaces what went wrong.
-    let extractResults;
-    const extractConfig = options.config ?? loadConfig();
-    // #554 minNewSessions gate: skip the entire extract pass (ensureIndex was
-    // already done upstream; here we elide every akmExtract/processSession call)
-    // when the NEW (unseen, in-window) candidate-session pool is below a minimum.
-    // 22% of improve runs produce zero memory-inference writes because extract
-    // finds no new sessions, yet still burns the full extract pipeline. Default 0
-    // (disabled) preserves existing always-run behaviour; only opted-in profiles
-    // (e.g. a user-enabled `frequent` strategy) set it. Evaluated BEFORE any LLM
-    // call so a skip costs zero LLM work AND writes nothing. A skipped extract
-    // never flags work for the NEXT run's consolidation mtime-gate (the
-    // downstream trigger #554 asks us to suppress).
-    const plan = args.plan ?? inspectExtractPass({ options, improveProfile, resolvedPlan, eventsCtx, readOnly: false });
-    // #593/#594: the ACTIVE resolved improve profile is the single source of
-    // truth for whether extract runs. (Previously this also ANDed in the legacy
-    // `session_extraction` feature flag, which only reads
-    // a retired global feature path; the selected strategy is authoritative.)
-    // `akmExtract` re-checks the same active profile internally via `improveProfile`.
-    if (resolvedPlan.processes.extract.enabled) {
-        const extractRunner = resolvedPlan.processes.extract.runner;
-        if (!extractRunner?.engine) {
-            throw new ConfigError("Resolved improve plan has no runner for enabled extract process.", "LLM_NOT_CONFIGURED");
-        }
-        const extractPlan = Object.freeze({
-            strategy: resolvedPlan.strategy.name,
-            engine: extractRunner.engine,
-            enabled: true,
-            process: resolvedPlan.processes.extract.config,
-            runner: extractRunner,
-            timeoutMs: extractRunner.timeoutMs === undefined ? 600_000 : extractRunner.timeoutMs,
-            embeddingConfig: Object.freeze(structuredClone(extractConfig.embedding)),
-            ...(resolvedPlan.processes.extract.notices?.length ? { notices: resolvedPlan.processes.extract.notices } : {}),
+    if (!resolvedPlan.processes.extract.enabled)
+        return { warnings };
+    const runner = resolvedPlan.processes.extract.runner;
+    if (!runner?.engine) {
+        throw new ConfigError("Resolved improve plan has no runner for enabled extract process.", "LLM_NOT_CONFIGURED");
+    }
+    const config = options.config ?? loadConfig();
+    const extractPlan = Object.freeze({
+        strategy: resolvedPlan.strategy.name,
+        engine: runner.engine,
+        enabled: true,
+        process: resolvedPlan.processes.extract.config,
+        runner,
+        timeoutMs: runner.timeoutMs === undefined ? 600_000 : runner.timeoutMs,
+        embeddingConfig: Object.freeze(structuredClone(config.embedding)),
+        ...(resolvedPlan.processes.extract.notices?.length ? { notices: resolvedPlan.processes.extract.notices } : {}),
+    });
+    if (plan.belowMinNewSessions) {
+        recordImproveSkip(eventsCtx, "memories/_extract", {
+            reason: "below_min_new_sessions",
+            newSessions: plan.newCandidateCount ?? 0,
+            minNewSessions: plan.minNewSessions,
         });
-        const availableHarnesses = plan.availableHarnesses;
-        if (plan.belowMinNewSessions) {
-            // Reuse the #551/#553 `improve_skipped` emission path so health's dynamic
-            // skipReasons aggregation surfaces this under `below_min_new_sessions`.
-            appendEvent({
-                eventType: "improve_skipped",
-                ref: "memories/_extract",
-                metadata: {
-                    reason: "below_min_new_sessions",
-                    newSessions: plan.newCandidateCount ?? 0,
-                    minNewSessions: plan.minNewSessions,
-                },
-            }, eventsCtx);
-            info(`[improve] extract skipped (new sessions ${plan.newCandidateCount ?? 0} < minNewSessions ${plan.minNewSessions})`);
+        info(`[improve] extract skipped (new sessions ${plan.newCandidateCount ?? 0} < minNewSessions ${plan.minNewSessions})`);
+    }
+    if (!plan.wouldRun)
+        return { warnings };
+    const extractResults = [];
+    for (const harness of plan.availableHarnesses) {
+        try {
+            extractResults.push(await attributeStage(resolvedPlan, "extract", () => akmExtract({
+                type: harness.name,
+                ...(primaryStashDir !== undefined ? { stashDir: primaryStashDir } : {}),
+                config,
+                resolvedPlan: extractPlan,
+                dryRun: options.dryRun ?? false,
+                signal: args.budgetSignal,
+                ...(options.extractHarnesses ? { harnesses: options.extractHarnesses } : {}),
+                ...(eventsCtx?.dbPath ? { stateDbPath: eventsCtx.dbPath } : {}),
+                eventsCtx,
+            })));
         }
-        if (plan.wouldRun) {
-            extractResults = [];
-            for (const h of availableHarnesses) {
-                try {
-                    const result = await withLlmStage("session-extraction", () => akmExtract({
-                        type: h.name,
-                        ...(primaryStashDir !== undefined ? { stashDir: primaryStashDir } : {}),
-                        config: extractConfig,
-                        resolvedPlan: extractPlan,
-                        dryRun: options.dryRun ?? false,
-                        signal: budgetSignal,
-                        ...(options.extractHarnesses ? { harnesses: options.extractHarnesses } : {}),
-                        // C2: pin extract's skip-tracking state.db open to the boundary path.
-                        ...(eventsCtx?.dbPath ? { stateDbPath: eventsCtx.dbPath } : {}),
-                        // R25: extract's event emits reuse the run's events context
-                        // (fast path when it carries the long-lived handle).
-                        eventsCtx,
-                    }), { engine: resolvedPlan.processes.extract.runner?.engine, process: "extract" });
-                    extractResults.push(result);
-                }
-                catch (err) {
-                    const msg = err instanceof Error ? err.message : String(err);
-                    warnings.push(`extract(${h.name}) failed: ${msg}`);
-                }
-            }
-            if (extractResults.length === 0) {
-                // All harnesses threw — clear so the envelope's `extract` field is
-                // absent rather than misleadingly empty.
-                extractResults = undefined;
-            }
+        catch (err) {
+            warnings.push(`extract(${harness.name}) failed: ${errMessage(err)}`);
         }
     }
-    return {
-        extractResults,
-        warnings,
-    };
+    // Every harness threw: no `extract` field rather than a misleadingly empty one.
+    return { ...(extractResults.length > 0 ? { extractResults } : {}), warnings };
 }
+// ── Validation ───────────────────────────────────────────────────────────────
 /**
- * Phase 1 — validation + schema-repair pass. Scans postCleanupRefs for assets
- * with structural problems (missing file, missing lesson description), attempts
- * LLM schema repair, and returns the still-failing ref set + the repair records.
+ * Structural validation (file on disk, lesson description) with optional LLM
+ * schema repair. A repair is advisory: a ref leaves the failure set only when a
+ * fresh read of the live asset passes.
  */
 export async function runValidationAndRepairPass(args) {
-    const { postCleanupRefs, options, startMs, budgetMs, primaryStashDir, resolvedPlan, repairValidationFailures, schemaRepairFn = runSchemaRepairPass, } = args;
-    const validateCandidate = async (candidate) => {
+    const { postCleanupRefs, options, resolvedPlan, repairValidationFailures } = args;
+    const validate = async (candidate) => {
         try {
             const filePath = candidate.filePath && fs.existsSync(candidate.filePath)
                 ? candidate.filePath
@@ -593,10 +328,8 @@ export async function runValidationAndRepairPass(args) {
                 return "file not found on disk";
             if (path.extname(filePath).toLowerCase() !== ".md")
                 return undefined;
-            if (isLessonCandidate(candidate.ref)) {
-                const fm = parseFrontmatter(fs.readFileSync(filePath, "utf8")).data;
-                if (!fm.description)
-                    return "missing description";
+            if (isLessonCandidate(candidate.ref) && !parseFrontmatter(fs.readFileSync(filePath, "utf8")).data.description) {
+                return "missing description";
             }
             return undefined;
         }
@@ -606,7 +339,7 @@ export async function runValidationAndRepairPass(args) {
     };
     const validationFailures = [];
     for (const candidate of postCleanupRefs) {
-        const reason = await validateCandidate(candidate);
+        const reason = await validate(candidate);
         if (reason)
             validationFailures.push({ ref: candidate.ref, reason });
     }
@@ -616,166 +349,122 @@ export async function runValidationAndRepairPass(args) {
             info(`  ${f.ref}: ${f.reason}`);
     }
     let schemaRepairs = [];
-    const repairedRefs = new Set();
-    // Schema repair pass: attempt to fix validation failures via LLM before skipping.
-    if (repairValidationFailures && validationFailures.length > 0) {
-        const validationRunner = resolvedPlan.processes.validation.runner;
-        if (validationRunner) {
-            const result = await withLlmStage("validation", () => schemaRepairFn(validationFailures, {
-                startMs,
-                budgetMs,
-                llmRunner: validationRunner,
-                // #591/#379 regression: options.stashDir is the raw, unresolved CLI
-                // flag (only set when --stash-dir is passed explicitly — never true
-                // for the scheduled tasks). primaryStashDir is the already-resolved
-                // source path and is what runSchemaRepairPass's `stashDir` param
-                // documents itself as needing ("proposal-queue writes"). Passing
-                // options.stashDir here made every schema-repair attempt throw
-                // `runSchemaRepairPass requires stashDir` on every cron invocation.
-                stashDir: primaryStashDir,
-                findFilePath: findAssetFilePath,
-                isLessonCandidateFn: isLessonCandidate,
-            }), { engine: resolvedPlan.processes.validation.runner?.engine, process: "validation" });
-            schemaRepairs = result.repairs;
-            // A repair result is advisory. Only a fresh structural read of the live
-            // asset can remove it from the failure set; queued content is not live.
-            const failedRefs = new Set(validationFailures.map((failure) => failure.ref));
-            const candidatesByRef = new Map(postCleanupRefs.map((candidate) => [candidate.ref, candidate]));
-            for (const ref of failedRefs) {
-                const candidate = candidatesByRef.get(ref);
-                if (candidate && !(await validateCandidate(candidate)))
-                    repairedRefs.add(ref);
-            }
+    const repaired = new Set();
+    const runner = resolvedPlan.processes.validation.runner;
+    if (repairValidationFailures && validationFailures.length > 0 && runner) {
+        const result = await attributeStage(resolvedPlan, "validation", () => (args.schemaRepairFn ?? runSchemaRepairPass)(validationFailures, {
+            startMs: args.startMs,
+            budgetMs: args.budgetMs,
+            llmRunner: runner,
+            // The resolved source path, not the raw `--stash-dir` flag.
+            stashDir: args.primaryStashDir,
+            findFilePath: findAssetFilePath,
+            isLessonCandidateFn: isLessonCandidate,
+        }));
+        schemaRepairs = result.repairs;
+        const byRef = new Map(postCleanupRefs.map((candidate) => [candidate.ref, candidate]));
+        for (const { ref } of validationFailures) {
+            const candidate = byRef.get(ref);
+            if (candidate && !(await validate(candidate)))
+                repaired.add(ref);
         }
     }
-    const validationFailureRefs = new Set(validationFailures.filter((f) => !repairedRefs.has(f.ref)).map((f) => f.ref));
-    if (repairedRefs.size > 0) {
-        info(`[improve] schema repair fixed ${repairedRefs.size}/${validationFailures.length} validation failures; ${validationFailureRefs.size} remain`);
+    const validationFailureRefs = new Set(validationFailures.filter((f) => !repaired.has(f.ref)).map((f) => f.ref));
+    if (repaired.size > 0) {
+        info(`[improve] schema repair fixed ${repaired.size}/${validationFailures.length} validation failures; ${validationFailureRefs.size} remain`);
     }
     return { validationFailures, validationFailureRefs, schemaRepairs };
 }
-/**
- * Resolve the preparation stages that precede candidate ranking. Keeping these
- * lifecycle decisions in one named pass preserves the 220-line orchestrator
- * ratchet while giving dry and live execution one implementation.
- */
-async function runPreparationPrelude(args) {
-    const { scope, options, plannedRefs, memoryCleanupPlan, primaryStashDir, memorySummary, reindexFn, startMs, budgetMs, eventsCtx, improveProfile, resolvedPlan, strategyName, budgetSignal, planOnly, actions, cleanupWarnings, } = args;
-    const memoryBudget = assessMemoryIndexBudget(primaryStashDir);
-    if (memoryBudget.warning)
-        cleanupWarnings.push(memoryBudget.warning);
-    // Consolidation intentionally precedes extract so current-run promotions
-    // cannot force the pool-delta gate open (#551).
+export async function runImprovePreparationStage(args) {
+    const { scope, options, plannedRefs, memoryCleanupPlan, primaryStashDir, eventsCtx, resolvedPlan } = args;
+    const planOnly = args.planOnly ?? options.dryRun === true;
+    const persist = !planOnly;
+    const actions = [];
+    const cleanupWarnings = [...(args.initialCleanupWarnings ?? [])];
+    const memoryIndexHealth = assessMemoryIndex(primaryStashDir, cleanupWarnings);
+    // Consolidation precedes extract, so it only judges memories from earlier runs.
     const consolidationPass = planOnly
-        ? (() => {
-            const planned = planConsolidationPass({
-                options,
-                primaryStashDir,
-                memorySummary,
-                improveProfile,
-                resolvedPlan,
-                eventsCtx,
-            });
-            return {
-                consolidation: {
-                    schemaVersion: 1,
-                    ok: true,
-                    shape: "consolidate-result",
-                    dryRun: true,
-                    previewOnly: true,
-                    target: options.target ?? options.stashDir ?? "",
-                    processed: 0,
-                    merged: 0,
-                    deleted: 0,
-                    promoted: [],
-                    contradicted: 0,
-                    warnings: [],
-                    durationMs: 0,
-                },
-                consolidationRan: false,
-                plan: planned.plan,
-            };
-        })()
-        : await runConsolidationPass({
-            options,
-            primaryStashDir,
-            memorySummary,
-            improveProfile,
-            resolvedPlan,
-            eventsCtx,
-            budgetSignal,
-            runBudgetMs: budgetMs,
-        });
-    const extractPlan = inspectExtractPass({ options, improveProfile, resolvedPlan, eventsCtx, readOnly: planOnly });
-    const extractPass = planOnly
-        ? { extractResults: undefined, warnings: [] }
-        : await runSessionExtractPass({
-            options,
-            primaryStashDir,
-            improveProfile,
-            resolvedPlan,
-            eventsCtx,
-            budgetSignal,
-            plan: extractPlan,
-        });
-    if (extractPass.warnings.length > 0)
-        cleanupWarnings.push(...extractPass.warnings);
-    if (!planOnly) {
+        ? {
+            consolidation: makeConsolidateResult({
+                dryRun: true,
+                previewOnly: true,
+                target: options.target ?? options.stashDir ?? "",
+                durationMs: 0,
+            }),
+            plan: planConsolidationPass(args).plan,
+        }
+        : await runConsolidationPass(args);
+    const extractPlan = inspectExtractPass(args, planOnly);
+    const extractPass = planOnly ? { warnings: [] } : await runSessionExtractPass(args, extractPlan);
+    cleanupWarnings.push(...extractPass.warnings);
+    if (persist) {
         appendEvent({
             eventType: "improve_invoked",
             ref: scope.mode === "ref" ? scope.value : `improve:${scope.mode}:${scope.value ?? "all"}`,
-            metadata: { strategy: strategyName, scope, dryRun: options.dryRun ?? false, eligibleCount: plannedRefs.length },
+            metadata: {
+                strategy: args.strategyName,
+                scope,
+                dryRun: options.dryRun ?? false,
+                eligibleCount: plannedRefs.length,
+            },
         }, eventsCtx);
     }
-    const allowCleanupApply = isAutonomyLaneAllowed("memoryCleanup", options.config ?? loadConfig());
-    const cleanup = planOnly
-        ? {
-            ...projectMemoryCleanup({
-                mode: "estimate",
-                plannedRefs,
-                candidateRefs: memoryCleanupPlan?.pruneCandidates.map((candidate) => candidate.ref) ?? [],
-                allowApply: allowCleanupApply,
-            }),
-            pruneActions: [],
-            warnings: [],
-            appliedCleanup: undefined,
+    // Memory cleanup: archive redundant derived memories (autonomy-gated).
+    const allowCleanup = isAutonomyLaneAllowed("memoryCleanup", options.config ?? loadConfig());
+    let appliedCleanup;
+    if (persist) {
+        try {
+            appliedCleanup =
+                primaryStashDir && memoryCleanupPlan && allowCleanup
+                    ? applyMemoryCleanup(primaryStashDir, memoryCleanupPlan)
+                    : undefined;
+        }
+        catch (err) {
+            cleanupWarnings.push(`applyMemoryCleanup failed: ${errMessage(err)}`);
         }
-        : await applyCleanupPass({
-            primaryStashDir,
-            memoryCleanupPlan,
+    }
+    const cleanup = planOnly
+        ? projectMemoryCleanup({
+            mode: "estimate",
+            plannedRefs,
+            candidateRefs: memoryCleanupPlan?.pruneCandidates.map((candidate) => candidate.ref) ?? [],
+            allowApply: allowCleanup,
+        })
+        : projectMemoryCleanup({
+            mode: "execution",
             plannedRefs,
-            reindexFn,
-            budgetSignal,
-            allowApply: allowCleanupApply,
+            archivedRefs: appliedCleanup?.archived.map((record) => record.ref) ?? [],
+            allowApply: allowCleanup,
         });
-    actions.push(...cleanup.pruneActions);
-    cleanupWarnings.push(...cleanup.warnings);
-    const validation = await runValidationAndRepairPass({
-        postCleanupRefs: cleanup.postCleanupRefs,
+    if (appliedCleanup) {
+        for (const candidate of memoryCleanupPlan?.pruneCandidates ?? []) {
+            if (!appliedCleanup.archived.some((record) => record.ref === candidate.ref))
+                continue;
+            actions.push({
+                ref: candidate.ref,
+                mode: "memory-prune",
+                result: { ok: true, pruned: true, reason: candidate.reason },
+            });
+        }
+        if ((appliedCleanup.archived.length > 0 || appliedCleanup.beliefStateTransitions.length > 0) && primaryStashDir) {
+            try {
+                await args.reindexFn({ stashDir: primaryStashDir, signal: args.budgetSignal });
+            }
+            catch (err) {
+                cleanupWarnings.push(`reindex after cleanup failed: ${errMessage(err)}`);
+            }
+        }
+    }
+    const { postCleanupRefs } = cleanup;
+    const { validationFailures, validationFailureRefs, schemaRepairs } = await runValidationAndRepairPass({
+        postCleanupRefs,
         options,
-        startMs,
-        budgetMs,
+        startMs: args.startMs,
+        budgetMs: args.budgetMs,
         primaryStashDir,
         resolvedPlan,
-        repairValidationFailures: !planOnly && resolvedPlan.processes.validation.enabled && options.repairValidationFailures !== false,
+        repairValidationFailures: persist && resolvedPlan.processes.validation.enabled && options.repairValidationFailures !== false,
     });
-    return {
-        memoryIndexHealth: memoryBudget.memoryIndexHealth,
-        consolidationPass,
-        extractPlan,
-        extractResults: extractPass.extractResults,
-        appliedCleanup: cleanup.appliedCleanup,
-        postCleanupRefs: cleanup.postCleanupRefs,
-        cleanupGate: cleanup.gate,
-        ...validation,
-    };
-}
-export async function runImprovePreparationStage(args) {
-    const { scope, options, primaryStashDir, eventsCtx, initialCleanupWarnings, improveProfile, resolvedPlan, planOnly = options.dryRun === true, } = args;
-    const actions = [];
-    const cleanupWarnings = initialCleanupWarnings ? [...initialCleanupWarnings] : [];
-    const { memoryIndexHealth, consolidationPass, extractPlan, extractResults, appliedCleanup, postCleanupRefs, cleanupGate, validationFailures, validationFailureRefs, schemaRepairs, } = await runPreparationPrelude({ ...args, planOnly, actions, cleanupWarnings });
-    // Phase 0.5 — structural hygiene pass
     let lintSummary;
     if (primaryStashDir) {
         try {
@@ -783,1796 +472,567 @@ export async function runImprovePreparationStage(args) {
             lintSummary = { fixed: lintResult.summary.fixed, flagged: lintResult.summary.flagged };
         }
         catch {
-            // lint is best-effort; never block improve
+            // lint never blocks improve
         }
     }
-    const recentErrors = seedRecentErrorWindows(schemaRepairs);
-    const snapshot = buildSnapshotManifest({ postCleanupRefs, validationFailureRefs, eventsCtx });
-    const gathered = gatherCandidates({
-        scope,
-        options,
-        primaryStashDir,
-        eventsCtx,
-        improveProfile,
-        resolvedPlan,
-        postCleanupRefs,
-        validationFailureRefs,
-        snapshot,
-        persist: !planOnly,
-    });
-    const eligibilitySourceByRef = stampEligibilitySource({
-        scope,
-        processableRefs: gathered.processableRefs,
-        mergedRefs: gathered.mergedRefs,
-        signalFiltered: gathered.signalFiltered,
-        proactiveRefs: gathered.proactiveRefs,
-        highSalienceRefs: gathered.highSalienceRefs,
-    });
-    // Shared admission boundary for every synthetic fallback lane. Cleanup and
-    // structural validation are exclusive selectors: no later rank/replay state
-    // may re-create a candidate they removed. Keep the exact surviving objects
-    // so any admitted fallback preserves its index-resolved file/item provenance.
-    const fallbackEligibleRefs = postCleanupRefs.filter((candidate) => !validationFailureRefs.has(candidate.ref));
-    const scored = scoreSalience({
-        scope,
-        options,
-        primaryStashDir,
-        eventsCtx,
-        mergedRefs: gathered.mergedRefs,
-        eligibilitySourceByRef,
-        feedbackSummary: gathered.feedbackSummary,
-        retrievalCounts: gathered.retrievalCounts,
-        signalFiltered: gathered.signalFiltered,
-        proactiveRefs: gathered.proactiveRefs,
-        highSalienceRefs: gathered.highSalienceRefs,
-        forgettingEligibleRefs: fallbackEligibleRefs,
-        persist: !planOnly,
-    });
-    // Replay is additive to the signal lanes, but it must not bypass selectors
-    // that have already removed a ref. Use the exact surviving objects so a
-    // replay admission preserves the index-resolved file/item provenance while
-    // excluding cleanup-pruned and structurally-invalid candidates.
-    const filtered = await filterEligibility({
-        scope,
-        options,
-        replayEligibleRefs: fallbackEligibleRefs,
-        eventsCtx,
-        mergedRefs: scored.mergedRefs,
-        salienceMap: scored.salienceMap,
-        eligibilitySourceByRef,
-        distillOnlyRefs: gathered.distillOnlyRefs,
-        validationFailureRefs,
-        summary: {
-            signalAndRetrievalRefs: gathered.signalAndRetrievalRefs,
-            signalFiltered: gathered.signalFiltered,
-        },
-        persist: !planOnly,
-    });
-    const preDiskRefSet = new Set(filtered.preDiskRefs.map((candidate) => candidate.ref));
-    const terminalSignalSkippedRefs = fallbackEligibleRefs.filter((candidate) => !preDiskRefSet.has(candidate.ref));
-    recordSignalSkipObservability({
-        actions,
-        terminalSignalSkippedRefs,
-        distillCooledRefs: gathered.distillCooledRefs,
-        eventsCtx,
-        persist: !planOnly,
-    });
-    // Gate counts are an exclusive, sequential accounting of the raw pool.
-    // Replay, proactive maintenance, high-salience, and forgetting-safety are
-    // legitimate signal-gate fallback lanes, so a ref admitted by any of them
-    // was not removed by the signal gate. Derive this count from the actual
-    // pre-disk survivor set instead of the earlier lane-rescue snapshot; the
-    // latter is intentionally assembled before replay and is also broader than
-    // the effective pool when --require-feedback-signal suppresses fallbacks.
-    const signalRemoved = terminalSignalSkippedRefs.length;
-    const totalReflectBlocked = terminalSignalSkippedRefs.length + gathered.distillOnlyRefs.length;
-    if (totalReflectBlocked > 0) {
-        info(`[improve] ${totalReflectBlocked} of ${gathered.preCooldownCount} indexed refs blocked by reflect signal-delta ` +
-            `(${terminalSignalSkippedRefs.length} fully skipped, ${gathered.distillOnlyRefs.length} routed to distill-only)`);
+    // Schema-repair errors get their own window; they are never shown to reflect.
+    const recentErrors = {};
+    for (const repair of schemaRepairs) {
+        if (repair.outcome !== "error")
+            continue;
+        pushRecentError(recentErrors, "schema-repair", repair.error ?? `schema repair error: ${repair.reason}`);
     }
-    const planningGates = [
-        cleanupGate,
-        {
-            name: "validation",
-            removed: validationFailureRefs.size,
-            reason: "structural validation failures",
-        },
-        {
-            name: "signal",
-            removed: signalRemoved,
-            reason: "no fresh signal and no fallback lane selected the ref",
-        },
-        {
-            name: "disk",
-            removed: filtered.missingDiskCount,
-            reason: "backing asset is absent on disk",
-        },
-        {
-            name: "limit",
-            removed: filtered.limitRemoved,
-            reason: "deferred by the effective run limit",
-        },
-    ];
+    const selection = await selectLoopCandidates(args, postCleanupRefs, validationFailureRefs, actions, persist);
     return {
         actions,
         cleanupWarnings,
         appliedCleanup,
         memoryIndexHealth,
-        extract: extractResults,
-        actionableRefs: filtered.actionableRefs,
-        signalBearingSet: gathered.signalBearingSet,
+        extract: extractPass.extractResults,
+        actionableRefs: selection.actionableRefs,
+        signalBearingSet: selection.signalBearingSet,
         validationFailures,
         schemaRepairs,
         lintSummary,
-        loopRefs: filtered.loopRefs,
-        distillCooledRefs: gathered.distillCooledRefs,
-        distillOnlyRefs: filtered.distillOnlyRefs,
-        coverageGaps: filtered.coverageGaps,
+        loopRefs: selection.loopRefs,
+        distillCooledRefs: selection.distillCooledRefs,
+        distillOnlyRefs: selection.distillOnlyRefs,
+        coverageGaps: selection.coverageGaps,
         recentErrors,
-        utilityMap: scored.utilityMap,
         consolidation: consolidationPass.consolidation,
-        consolidationRan: consolidationPass.consolidationRan,
-        ...(gathered.proactiveMaintenanceSummary ? { proactiveMaintenance: gathered.proactiveMaintenanceSummary } : {}),
+        ...(selection.proactive.proactiveMaintenanceSummary
+            ? { proactiveMaintenance: selection.proactive.proactiveMaintenanceSummary }
+            : {}),
         planning: {
-            gates: planningGates,
-            replayBudget: filtered.replayBudget,
-            ...(gathered.proactivePlan ? { proactive: gathered.proactivePlan } : {}),
+            gates: [
+                cleanup.gate,
+                { name: "validation", removed: validationFailureRefs.size, reason: "structural validation failures" },
+                ...selection.gates,
+            ],
+            ...(selection.proactive.proactivePlan ? { proactive: selection.proactive.proactivePlan } : {}),
             consolidation: consolidationPass.plan,
             extract: { wouldRun: extractPlan.wouldRun, reason: extractPlan.reason },
         },
     };
 }
-// ── preparation-stage passes (WI-7.6 decomposition, R31) ────────────────────
-// The six-pass split prescribed by the chunk-7 brief §WI-7.6, adapted to the
-// code as it exists at HEAD (anchors re-measured; see the chunk-7 ledger):
-//   snapshot-manifest  → buildSnapshotManifest
-//   candidate-gather   → gatherCandidates (+ its five lane/sub-passes)
-//   salience-score     → scoreSalience (+ outcome/vector/persist sub-passes)
-//   valence-score      → the two computeValenceScore call sites move VERBATIM
-//                        inside the salience passes (pure fn; no separate pass)
-//   standards-context  → does not exist in preparation.ts (assembly lives in
-//                        extract.ts — recorded in the ledger, no empty pass)
-//   eligibility-filter → filterEligibility (+ replay/disk-check sub-passes)
-// Every pass takes an args object and returns its results; the orchestrator
-// folds them. Shared-by-reference structures (the ImproveEligibleRef objects,
-// eligibilitySourceByRef, salienceMap, actions, recentErrors) keep their
-// identity — attribution stamps must travel with the ref objects into the
-// loop stage exactly as before.
-/** Phase 0 — MEMORY.md budget check (200-line cap; warn at 180). */
-function assessMemoryIndexBudget(primaryStashDir) {
-    let warning;
-    // Phase 0 — MEMORY.md budget check (200-line cap; warn at 180)
-    let memoryIndexHealth;
-    if (primaryStashDir) {
-        const memoryMdPath = path.join(primaryStashDir, "memories", "MEMORY.md");
-        if (fs.existsSync(memoryMdPath)) {
-            try {
-                const lines = fs.readFileSync(memoryMdPath, "utf8").split("\n").length;
-                const overBudget = lines >= 180;
-                memoryIndexHealth = { lineCount: lines, overBudget };
-                if (overBudget) {
-                    warning = `MEMORY.md has ${lines} lines (budget: 200). Consolidation strongly recommended.`;
-                }
-            }
-            catch {
-                // best-effort
-            }
-        }
-    }
-    return { memoryIndexHealth, warning };
-}
-/**
- * Memory-cleanup apply + prune-action recording + the post-cleanup reindex.
- * Returns the surviving ref set and the prune actions/warnings for the
- * orchestrator to fold (same order as the old inline pushes).
- */
-async function applyCleanupPass(args) {
-    const { primaryStashDir, memoryCleanupPlan, plannedRefs, reindexFn, budgetSignal, allowApply } = args;
-    const pruneActions = [];
-    const warnings = [];
-    let appliedCleanup;
+/** MEMORY.md line budget: warn at 180 of 200 lines. */
+function assessMemoryIndex(primaryStashDir, warnings) {
+    if (!primaryStashDir)
+        return undefined;
+    const memoryMdPath = path.join(primaryStashDir, "memories", "MEMORY.md");
+    if (!fs.existsSync(memoryMdPath))
+        return undefined;
     try {
-        appliedCleanup =
-            primaryStashDir && memoryCleanupPlan && allowApply
-                ? applyMemoryCleanup(primaryStashDir, memoryCleanupPlan)
-                : undefined;
-    }
-    catch (err) {
-        warnings.push(`applyMemoryCleanup failed: ${err instanceof Error ? err.message : String(err)}`);
-    }
-    const projection = projectMemoryCleanup({
-        mode: "execution",
-        plannedRefs,
-        archivedRefs: appliedCleanup?.archived.map((record) => record.ref) ?? [],
-        allowApply,
-    });
-    // ── Phase 1: validation pass + schema repair (run on full postCleanupRefs) ──
-    // Identifies refs whose on-disk asset has structural problems. Validation
-    // failures are excluded from every downstream bucket. Run early so the
-    // cooldown partition operates on a clean set.
-    if (appliedCleanup) {
-        for (const candidate of memoryCleanupPlan?.pruneCandidates ?? []) {
-            const archived = appliedCleanup.archived.find((record) => record.ref === candidate.ref);
-            if (!archived)
-                continue;
-            pruneActions.push({
-                ref: candidate.ref,
-                mode: "memory-prune",
-                result: { ok: true, pruned: true, reason: candidate.reason },
-            });
-        }
-        if ((appliedCleanup.archived.length > 0 || appliedCleanup.beliefStateTransitions.length > 0) && primaryStashDir) {
-            try {
-                await reindexFn({ stashDir: primaryStashDir, signal: budgetSignal });
-            }
-            catch (err) {
-                warnings.push(`reindex after cleanup failed: ${err instanceof Error ? err.message : String(err)}`);
-            }
+        const lineCount = fs.readFileSync(memoryMdPath, "utf8").split("\n").length;
+        if (lineCount >= 180) {
+            warnings.push(`MEMORY.md has ${lineCount} lines (budget: 200). Consolidation strongly recommended.`);
         }
+        return { lineCount, overBudget: lineCount >= 180 };
     }
-    return { appliedCleanup, ...projection, pruneActions, warnings };
-}
-/** Seed the per-originator rolling error windows from schema-repair errors. */
-function seedRecentErrorWindows(schemaRepairs) {
-    // O-5 / #378: Per-originator rolling error windows.
-    // Reflexion (arXiv:2303.11366) warns that cross-task verbal critique
-    // contamination degrades below single-shot baseline. Each originator key
-    // ("schema-repair", "reflect") maintains its own rolling window so that
-    // schema-repair failures are not injected as avoidPatterns into reflect calls.
-    const recentErrors = {};
-    const RECENT_ERRORS_CAP = 3;
-    // Helper: push an error onto an originator's rolling window.
-    function pushRecentError(originator, msg) {
-        if (!recentErrors[originator])
-            recentErrors[originator] = [];
-        recentErrors[originator].push(msg);
-        if (recentErrors[originator].length > RECENT_ERRORS_CAP)
-            recentErrors[originator].shift();
-    }
-    // Seed schema-repair originator window from any schema-repair errors.
-    for (const repair of schemaRepairs) {
-        if (repair.outcome === "error") {
-            const errMsg = repair.error ?? `schema repair error: ${repair.reason}`;
-            pushRecentError("schema-repair", errMsg);
-        }
+    catch {
+        return undefined;
     }
-    return recentErrors;
 }
-/** Pass: snapshot-manifest — the three timestamp maps + the 30-day signal window. */
+// ── Candidate selection ──────────────────────────────────────────────────────
+const FEEDBACK_SIGNAL_WINDOW_DAYS = 30;
+/** Feedback that counts as a signal carries a signal or a note (a bare `akm feedback` does not). */
+function isSignalEvent(metadata) {
+    const meta = metadata;
+    return meta !== undefined && (typeof meta.signal === "string" || typeof meta.note === "string");
+}
+/** One read of the feedback events and the ledger's reflect/distill rows. */
 export function buildSnapshotManifest(args) {
-    const { postCleanupRefs, validationFailureRefs, eventsCtx } = args;
-    // ── Phase 2: signal-delta eligibility sets built EARLY ────────────────────
-    // 0.8.0 replaces the flat time-based cooldowns (which produced synchronised
-    // waves whenever many refs cooled at the same instant — see the 2026-05-26
-    // 54-ref simultaneous-reflect incident) with a *signal-delta* gate:
-    //
-    //   reflectEligible(ref) ≡ latestFeedbackTs(ref) > lastReflectProposalTs(ref)
-    //   distillEligible(ref) ≡ latestFeedbackTs(ref) > lastDistillProposalTs(ref)
-    //
-    // i.e. a ref is re-eligible iff new feedback has landed since the last
-    // proposal was generated for it. Stable content with no new signal stays
-    // out of the queue regardless of clock time; a sudden burst of feedback
-    // surfaces only the refs that the burst actually touches.
-    //
-    // The 30-day FEEDBACK_SIGNAL_WINDOW_DAYS bound still applies — only feedback
-    // events newer than that count as "current signal". Ancient one-off
-    // negatives don't permanently lock a ref into every run.
-    const FEEDBACK_SIGNAL_WINDOW_DAYS = 30;
+    const { eventsCtx, stashDir } = args;
     const feedbackSinceCutoff = new Date(Date.now() - daysToMs(FEEDBACK_SIGNAL_WINDOW_DAYS)).toISOString();
-    // Build the three timestamp maps once across the entire postCleanupRefs set.
-    // Per-ref queries would be N+1 and the planner is already the hottest path
-    // in `akm improve`.
-    const candidateRefs = postCleanupRefs.filter((r) => !validationFailureRefs.has(r.ref)).map((r) => r.ref);
-    // Carry each candidate's item_ref into the feedback/proposal timestamp reads.
-    const itemRefByRef = buildItemRefByRef(postCleanupRefs);
-    const latestFeedbackTs = buildLatestFeedbackTsMap(candidateRefs, feedbackSinceCutoff, itemRefByRef, eventsCtx);
-    const lastReflectProposalTs = buildLatestProposalTsMap(candidateRefs, "reflect", itemRefByRef, eventsCtx);
-    const lastDistillProposalTs = buildLatestProposalTsMap(candidateRefs, "distill", itemRefByRef, eventsCtx);
-    return { feedbackSinceCutoff, latestFeedbackTs, lastReflectProposalTs, lastDistillProposalTs };
-}
-/**
- * Pass: candidate-gather — the signal-delta partition, the bulk feedback
- * summary, retrieval signals, the Layer-2 proactive and Layer-3 high-salience
- * rescue lanes, and the merged candidate set. Lane attribution stamping is a
- * separate pass — see `stampEligibilitySource` — run by the caller once
- * `mergedRefs` is known.
- */
-function gatherCandidates(args) {
-    const { scope, options, primaryStashDir, eventsCtx, improveProfile, resolvedPlan, postCleanupRefs, persist } = args;
-    const { feedbackSinceCutoff, lastReflectProposalTs, lastDistillProposalTs } = args.snapshot;
-    const partition = partitionBySignalDelta({
-        scope,
-        options,
-        postCleanupRefs,
-        validationFailureRefs: args.validationFailureRefs,
-        snapshot: args.snapshot,
-    });
-    const { distillCooledRefs, preCooldownCount, eligibleRefs, distillOnlyRefs, noFeedbackPool } = partition;
-    // ── Phase 4: signal/feedback/utility/sort on the reduced set ──────────────
-    // Everything from here works on (eligibleRefs ∪ distillOnlyRefs) plus the
-    // deferred noFeedbackPool that may be rescued by the proactive-maintenance
-    // (Layer 2) or high-salience (Layer 3) fallbacks below. The fully-skipped
-    // bucket is retained as partition metadata only; terminal skip observability
-    // is delayed until every fallback lane has finalized. We deliberately avoid
-    // spending DB/CPU on refs that the signal-delta gate rejected with feedback
-    // already on record.
-    const processableRefs = [...eligibleRefs, ...distillOnlyRefs];
-    const feedbackSummary = buildFeedbackSummaryMap({
-        processableRefs,
-        noFeedbackPool,
-        eventsCtx,
-        feedbackSinceCutoff,
-    });
-    const signalFiltered = processableRefs.filter((candidate) => feedbackSummary.get(candidate.ref)?.hasSignal === true);
-    const signalBearingSet = new Set(signalFiltered.map((r) => r.ref));
-    // Zero-feedback candidates for the proactive/high-salience fallbacks:
-    // processableRefs without a recent signal, plus the deferred noFeedbackPool.
-    // Dedupe by ref (the two sources are disjoint by construction, but guard
-    // against overlap defensively).
-    const noFeedbackSeen = new Set();
-    const noFeedbackCandidates = [];
-    for (const r of [...processableRefs.filter((r) => !signalBearingSet.has(r.ref)), ...noFeedbackPool]) {
-        if (noFeedbackSeen.has(r.ref))
-            continue;
-        noFeedbackSeen.add(r.ref);
-        noFeedbackCandidates.push(r);
+    const candidates = args.postCleanupRefs.filter((r) => !args.validationFailureRefs.has(r.ref));
+    const refByKey = new Map(candidates.map((r) => [keyOf(r), r.ref]));
+    const latestFeedbackTs = new Map();
+    const feedback = new Map(candidates.map((r) => [r.ref, { hasSignal: false, positive: 0, negative: 0 }]));
+    if (candidates.length > 0) {
+        for (const e of readEvents({ type: "feedback" }, eventsCtx).events) {
+            const ref = e.ref ? refByKey.get(e.ref) : undefined;
+            const entry = ref ? feedback.get(ref) : undefined;
+            if (!ref || !entry)
+                continue;
+            const ts = e.ts ?? "";
+            if (ts >= feedbackSinceCutoff && isSignalEvent(e.metadata)) {
+                entry.hasSignal = true;
+                if (ts > (latestFeedbackTs.get(ref) ?? ""))
+                    latestFeedbackTs.set(ref, ts);
+            }
+            const signal = e.metadata?.signal;
+            if (signal === "positive")
+                entry.positive++;
+            else if (signal === "negative")
+                entry.negative++;
+        }
     }
-    const { retrievalCounts, lastUseMsForProactive } = fetchRetrievalSignals({
-        options,
-        primaryStashDir,
-        signalFiltered,
-        noFeedbackCandidates,
-        eventsCtx,
-        persist,
-    });
-    // `--require-feedback-signal` is a hard policy boundary, not merely a final
-    // list filter. Do not run or report fallback selectors that the invocation
-    // explicitly disabled (and do not emit their live selection events).
-    const allowFallbacks = options.requireFeedbackSignal !== true;
-    const proactive = allowFallbacks
-        ? selectProactiveMaintenanceLane({
-            scope,
-            improveProfile,
-            resolvedPlan,
-            eventsCtx,
-            noFeedbackCandidates,
-            lastReflectProposalTs,
-            lastDistillProposalTs,
-            retrievalCounts,
-            lastUseMsForProactive,
-            persist,
-        })
-        : { proactiveRefs: [] };
-    const proactiveRefs = proactive.proactiveRefs;
-    const proactiveMaintenanceSummary = proactive.proactiveMaintenanceSummary;
-    const highSalienceRefs = allowFallbacks
-        ? selectHighSalienceLane({
-            options,
-            improveProfile,
-            eventsCtx,
-            noFeedbackCandidates,
-            proactiveRefs,
-            lastReflectProposalTs,
-            persist,
-        })
-        : [];
-    // If the user explicitly scoped to a single ref, always act on it —
-    // skip the signal/retrieval filter entirely. The filter exists to avoid
-    // noisy "improve everything" runs; it should not gate an intentional
-    // per-ref invocation where the user's explicit choice is the signal.
-    //
-    // For type/all scope: only process refs with usage signals (recent feedback
-    // or a proactive/high-salience rescue). A stash with no signals has 0
-    // eligible refs — usage is the gate. Run `akm feedback <ref> --positive` or
-    // retrieve assets to bring them into the eligible pool.
-    // Layer-2 proactive refs join the eligible set alongside feedback-signal
-    // refs. The three sources are disjoint by construction (proactive draws from
-    // noFeedbackCandidates, and high-salience draws from the remainder), but
-    // dedupe defensively so a ref can never enter the loop twice.
-    // `requireFeedbackSignal` still suppresses all fallback sources for callers
-    // that want feedback-only runs.
-    const signalAndRetrievalRefs = dedupeRefs([...signalFiltered, ...proactiveRefs, ...highSalienceRefs]);
-    const mergedRefs = scope.mode === "ref" ? processableRefs : options.requireFeedbackSignal ? signalFiltered : signalAndRetrievalRefs;
+    const ledger = stashDir
+        ? loadLedgerSnapshot({ eventsCtx, ...(args.readOnly ? { readOnly: true } : {}) }, stashDir, ["reflect", "distill"])
+        : new Map();
     return {
-        distillCooledRefs,
-        preCooldownCount,
-        distillOnlyRefs,
-        feedbackSummary,
-        signalFiltered,
-        signalBearingSet,
-        retrievalCounts,
-        proactiveRefs,
-        proactiveMaintenanceSummary,
-        proactivePlan: proactive.proactivePlan,
-        highSalienceRefs,
-        signalAndRetrievalRefs,
-        mergedRefs,
-        processableRefs,
+        feedbackSinceCutoff,
+        nowIso: new Date().toISOString(),
+        latestFeedbackTs,
+        ledger,
+        lastReflectAttemptAt: lastAttemptByRef(ledger, "reflect", candidates),
+        lastDistillAttemptAt: lastAttemptByRef(ledger, "distill", candidates),
+        feedback,
     };
 }
 /**
- * Attribution tagging: stamp each ref with the eligibility lane that selected
- * it. Every reflect/distill proposal must record WHICH lane chose its source
- * asset so downstream accept/reject/revert/retrieval outcomes can be sliced by
- * lane (does the PROACTIVE lane produce value vs the reactive lanes?). We
- * build the lane map here — the one place all three lanes are known — and
- * stamp it onto each ImproveEligibleRef object. Because the ref objects are
- * shared by reference across buckets, the stamp travels with the ref through
- * the sort, disk-check, and loop stages down to the reflect/distill event
- * emit sites and createProposal calls. See EligibilitySource for the lane
- * vocabulary.
- *
- * Precedence (prefer the most specific reactive signal):
- *   scope > signal-delta > proactive > high-salience
- * A ref with real feedback is attributed to feedback even if it was also due
- * for proactive maintenance or had high encoding salience. We apply lanes
- * weakest-first so the strongest overwrites; the explicit --scope <ref> bypass
- * wins outright (user intent).
- */
-function stampEligibilitySource(args) {
-    const { scope, processableRefs, mergedRefs, signalFiltered, proactiveRefs, highSalienceRefs } = args;
-    const eligibilitySourceByRef = new Map();
-    for (const r of highSalienceRefs)
-        eligibilitySourceByRef.set(r.ref, "high-salience");
-    for (const r of proactiveRefs)
-        eligibilitySourceByRef.set(r.ref, "proactive");
-    for (const r of signalFiltered)
-        eligibilitySourceByRef.set(r.ref, "signal-delta");
-    if (scope.mode === "ref") {
-        // O-2 (#365): explicit --scope <ref> bypass — every ref in processableRefs
-        // arrived via the scopeRefBypass branch, so attribute the whole set to scope.
-        for (const r of processableRefs)
-            eligibilitySourceByRef.set(r.ref, "scope");
-    }
-    for (const r of mergedRefs) {
-        // "unknown" is a genuine fallback, never a silent alias for signal-delta:
-        // only refs we truly cannot attribute land here (none in practice, since
-        // mergedRefs is always a subset of the four lanes above).
-        r.eligibilitySource = eligibilitySourceByRef.get(r.ref) ?? "unknown";
-    }
-    return eligibilitySourceByRef;
-}
-/**
- * The signal-delta partition of postCleanupRefs into the four buckets (pass:
- * candidate-gather, phase 3). The 2026-05-26 54-ref incident semantics move
- * VERBATIM — see the phase-2/3 comments inside.
+ * Partition the post-cleanup refs against the ledger:
+ *  - eligibleRefs: reflect's signal delta passes (distill may still be cooled);
+ *  - distillOnlyRefs: only distill's passes, on a distill candidate;
+ *  - noFeedbackPool: no recent feedback and no reflect window, left to the
+ *    fallback lanes;
+ *  - fullySkippedCount: feedback on record but nothing new, or a live window.
+ * An explicit `--scope <ref>` bypasses every gate.
  */
 export function partitionBySignalDelta(args) {
-    const { scope, options, postCleanupRefs, validationFailureRefs } = args;
-    const { latestFeedbackTs, lastReflectProposalTs, lastDistillProposalTs } = args.snapshot;
-    // Refs the distill signal-delta gate rejected at planning time. The main
-    // loop reads this to skip distill for these refs without re-checking
-    // eligibility per iteration.
-    const distillCooledRefs = new Set();
-    const preCooldownCount = postCleanupRefs.length;
-    // ── Phase 3: partition postCleanupRefs by signal-delta eligibility ────────
-    // Three buckets (validation failures are excluded entirely):
-    //   eligibleRefs        — reflect signal-delta passes (full reflect+distill
-    //                         loop path; distill guard remains in the loop for
-    //                         refs that fail the distill signal-delta gate).
-    //   distillOnlyRefs     — reflect blocked but distill signal-delta passes
-    //                         AND ref is a distill candidate.
-    //   noFeedbackPool      — neither signal-delta gate passes *and* the ref has
-    //                         no recent feedback signal at all. These are NOT
-    //                         skipped here: they are handed to the proactive
-    //                         (Layer 2) and high-salience (Layer 3) fallbacks
-    //                         below so never-rated assets can still be improved.
-    //                         Only refs those lanes decline are fully skipped.
-    //   fullySkippedCount   — has stale feedback but no signal delta → genuine
-    //                         skip candidate, excluded from sort. Final skip
-    //                         observability is emitted only after fallbacks.
-    const eligibleRefs = [];
-    const distillOnlyRefs = [];
-    // Zero-(recent-)feedback refs deferred to the proactive/high-salience fallbacks.
-    const noFeedbackPool = [];
-    let fullySkippedCount = 0;
-    // O-2 (#365): explicit --scope <ref> bypasses every gate (user intent wins).
-    const scopeRefBypass = scope.mode === "ref";
+    const { postCleanupRefs, validationFailureRefs } = args;
+    const { latestFeedbackTs, ledger, nowIso } = args.snapshot;
+    // Newer feedback lifts a revisit window, never a rejection.
+    const deltaPasses = (candidate, source) => {
+        const feedbackAt = latestFeedbackTs.get(candidate.ref);
+        if (!feedbackAt)
+            return false;
+        const row = ledgerRowFor(ledger, source, candidate.ref, candidate.itemRef);
+        return feedbackAt > (row?.lastAttemptAt ?? "") && !isLedgerBlocked(row, nowIso, feedbackAt);
+    };
+    const out = {
+        distillCooledRefs: new Set(),
+        preCooldownCount: postCleanupRefs.length,
+        eligibleRefs: [],
+        distillOnlyRefs: [],
+        noFeedbackPool: [],
+        fullySkippedCount: 0,
+    };
     for (const r of postCleanupRefs) {
         if (validationFailureRefs.has(r.ref))
             continue;
-        if (scopeRefBypass) {
-            eligibleRefs.push(r);
+        if (args.scope.mode === "ref") {
+            out.eligibleRefs.push(r);
             continue;
         }
-        const reflectOk = isSignalDeltaEligible(r.ref, latestFeedbackTs, lastReflectProposalTs);
-        const distillOk = isSignalDeltaEligible(r.ref, latestFeedbackTs, lastDistillProposalTs);
-        const isDistillCandidate = isDistillCandidateRef(r.ref, options.stashDir);
+        const reflectOk = deltaPasses(r, "reflect");
+        const distillOk = deltaPasses(r, "distill");
         if (reflectOk) {
-            if (!distillOk && isDistillCandidate) {
-                // Reflect passes the gate, distill does not. Record only partition
-                // metadata here; observability is emitted after every fallback selector
-                // has finalized the invocation's terminal skipped set.
-                distillCooledRefs.add(r.ref);
-            }
-            else if (!distillOk) {
-                // Not a distill candidate AND distill gate doesn't pass — just mark
-                // distillCooled so the loop's distill section is a no-op.
-                distillCooledRefs.add(r.ref);
-            }
-            eligibleRefs.push(r);
+            if (!distillOk)
+                out.distillCooledRefs.add(r.ref);
+            out.eligibleRefs.push(r);
         }
-        else if (distillOk && isDistillCandidate) {
-            // Reflect blocked but distill passes → distill-only bucket.
-            distillOnlyRefs.push(r);
+        else if (distillOk && isDistillCandidateRef(r.ref, args.options.stashDir)) {
+            out.distillOnlyRefs.push(r);
         }
-        else if (!latestFeedbackTs.has(r.ref)) {
-            // Neither signal-delta gate passes AND there is no recent feedback signal
-            // at all. Rather than skip outright, defer to the proactive-maintenance
-            // and high-salience fallbacks below: a never-rated asset is exactly what
-            // those lanes are meant to rescue. Refs those lanes decline are skipped there.
-            noFeedbackPool.push(r);
+        else if (!latestFeedbackTs.has(r.ref) &&
+            !isLedgerBlocked(ledgerRowFor(ledger, "reflect", r.ref, r.itemRef), nowIso)) {
+            out.noFeedbackPool.push(r);
         }
         else {
-            // Has feedback on record but no signal delta since the last proposal —
-            // genuinely a fully-skipped candidate. Count it as partition metadata;
-            // final observability waits until replay and every other fallback lane
-            // has had a chance to rescue it.
-            fullySkippedCount++;
+            out.fullySkippedCount++;
         }
     }
-    return {
-        distillCooledRefs,
-        preCooldownCount,
-        eligibleRefs,
-        distillOnlyRefs,
-        noFeedbackPool,
-        fullySkippedCount,
-    };
+    return out;
 }
 /**
- * Emit signal-delta skip observability only after every fallback lane has
- * finalized the pre-disk survivor set. This prevents replay, proactive,
- * high-salience, or forgetting-safety winners from also being recorded as
- * terminally skipped work.
+ * Pick the loop's refs: signal delta, the fallback lanes (unless
+ * `--require-feedback-signal`), lane attribution, salience, the
+ * no-op-dampened ranking, the disk check and the limit.
  */
-function recordSignalSkipObservability(args) {
-    const { actions, terminalSignalSkippedRefs, distillCooledRefs, eventsCtx, persist } = args;
-    for (const ref of distillCooledRefs) {
+async function selectLoopCandidates(args, postCleanupRefs, validationFailureRefs, actions, persist) {
+    const { scope, options, primaryStashDir, eventsCtx, improveProfile } = args;
+    const snapshot = buildSnapshotManifest({
+        postCleanupRefs,
+        validationFailureRefs,
+        eventsCtx,
+        stashDir: primaryStashDir ?? options.stashDir,
+        readOnly: !persist,
+    });
+    const partition = partitionBySignalDelta({ scope, options, postCleanupRefs, validationFailureRefs, snapshot });
+    const processableRefs = [...partition.eligibleRefs, ...partition.distillOnlyRefs];
+    const signalFiltered = processableRefs.filter((c) => snapshot.feedback.get(c.ref)?.hasSignal === true);
+    const signalBearingSet = new Set(signalFiltered.map((r) => r.ref));
+    // The fallback lanes (proactive, high salience) have no usage evidence of
+    // their own: they pick only what retrieval returned or new material improve
+    // never processed (#986). Evaluated once per candidate.
+    const fallbackEligible = postCleanupRefs.filter((c) => !validationFailureRefs.has(c.ref));
+    const allowFallbacks = options.requireFeedbackSignal !== true;
+    const retrievalScope = scope.mode === "ref" || !allowFallbacks
+        ? undefined
+        : loadRetrievalScope({ eventsCtx, ...(persist ? {} : { readOnly: true }) }, primaryStashDir ?? options.stashDir);
+    const unscoped = new Set(fallbackEligible.filter((c) => !isInRetrievalScope(retrievalScope, c.ref, c.filePath)).map((c) => c.ref));
+    const noFeedbackPool = dedupeRefs([
+        ...processableRefs.filter((r) => !signalBearingSet.has(r.ref)),
+        ...partition.noFeedbackPool,
+    ]);
+    const noFeedbackCandidates = noFeedbackPool.filter((r) => !unscoped.has(r.ref));
+    // Only a ref no fallback lane may pick anymore is charged to the retrieval gate.
+    const outOfScope = new Set(noFeedbackPool.filter((r) => unscoped.has(r.ref)).map((r) => r.ref));
+    const retrieval = fetchRetrievalSignals(options, signalFiltered, noFeedbackCandidates, eventsCtx, persist);
+    const proactive = allowFallbacks
+        ? selectProactiveMaintenanceLane(args, noFeedbackCandidates, snapshot, retrieval, persist)
+        : { proactiveRefs: [] };
+    const highSalienceRefs = allowFallbacks
+        ? selectHighSalienceLane(options, improveProfile, eventsCtx, noFeedbackCandidates.filter((r) => !proactive.proactiveRefs.some((p) => p.ref === r.ref)), snapshot.lastReflectAttemptAt, persist)
+        : [];
+    // An explicit ref scope always acts on its ref; otherwise usage signals gate the pool.
+    const signalAndRetrievalRefs = dedupeRefs([...signalFiltered, ...proactive.proactiveRefs, ...highSalienceRefs]);
+    const mergedRefs = scope.mode === "ref" ? processableRefs : options.requireFeedbackSignal ? signalFiltered : signalAndRetrievalRefs;
+    // Lane attribution, weakest first so the strongest wins: high-salience <
+    // proactive < signal-delta, and an explicit ref scope over everything.
+    const sourceByRef = new Map();
+    for (const r of highSalienceRefs)
+        sourceByRef.set(r.ref, "high-salience");
+    for (const r of proactive.proactiveRefs)
+        sourceByRef.set(r.ref, "proactive");
+    for (const r of signalFiltered)
+        sourceByRef.set(r.ref, "signal-delta");
+    if (scope.mode === "ref")
+        for (const r of processableRefs)
+            sourceByRef.set(r.ref, "scope");
+    for (const r of mergedRefs)
+        r.eligibilitySource = sourceByRef.get(r.ref) ?? "unknown";
+    const salienceMap = scoreSalience(args, mergedRefs, snapshot.feedback, retrieval.retrievalCounts, persist);
+    // Rank by salience; a ref skipped as a no-op repeatedly sorts lower (its stored rank is untouched).
+    const noOps = new Map();
+    withRunState(eventsCtx, persist, (db) => {
+        for (const r of mergedRefs)
+            noOps.set(r.ref, getAssetSalience(db, keyOf(r))?.consecutive_no_ops ?? 0);
+    });
+    const effectiveScore = (ref) => {
+        const rank = salienceMap.get(ref)?.rankScore ?? 0;
+        return (noOps.get(ref) ?? 0) >= SALIENCE_NO_OP_DAMPEN_THRESHOLD ? rank * SALIENCE_NO_OP_DAMPEN_FACTOR : rank;
+    };
+    const sorted = [...mergedRefs].sort((a, b) => effectiveScore(b.ref) - effectiveScore(a.ref) || (a.ref < b.ref ? -1 : a.ref > b.ref ? 1 : 0));
+    const coverageGaps = withIndexDb(!persist, getZeroResultSearches) ?? [];
+    const { actionableRefs, missing } = await dropRefsMissingOnDisk(sorted, options, eventsCtx, persist);
+    const selection = selectEffectiveImproveRefs({
+        rankedRefs: actionableRefs,
+        distillOnlyRefs: partition.distillOnlyRefs,
+        limit: options.limit,
+    });
+    if (signalAndRetrievalRefs.length > 0) {
+        info(`[improve] ${signalAndRetrievalRefs.length} refs with usage signals (${signalFiltered.length} feedback)`);
+    }
+    if (validationFailureRefs.size > 0)
+        info(`[improve] ${validationFailureRefs.size} with validation failures excluded`);
+    if (persist && missing.length > 0)
+        info(`[improve] ${missing.length} candidates dropped — file not on disk`);
+    const deferred = actionableRefs.length - selection.loopRefs.length;
+    info(`[improve] ${actionableRefs.length} actionable; ${selection.loopRefs.length} will be processed` +
+        (options.limit && deferred > 0 ? ` (--limit ${options.limit} applied; ${deferred} deferred)` : ""));
+    // Skip observability waits until every fallback lane has finalized the
+    // survivors, so a rescued ref is never also reported skipped.
+    const survivors = new Set(sorted.map((c) => c.ref));
+    const skipped = fallbackEligible.filter((c) => !survivors.has(c.ref));
+    const retrievalSkipped = skipped.filter((c) => outOfScope.has(c.ref));
+    const signalSkipped = skipped.filter((c) => !outOfScope.has(c.ref));
+    for (const ref of partition.distillCooledRefs) {
         actions.push({ ref, mode: "distill-skipped", result: { ok: true, reason: "distill signal-delta" } });
-        if (persist) {
-            appendEvent({
-                eventType: "improve_skipped",
-                ref,
-                metadata: { reason: "distill_no_new_signal" },
-            }, eventsCtx);
-        }
+        if (persist)
+            recordImproveSkip(eventsCtx, ref, { reason: "distill_no_new_signal" });
     }
-    for (const candidate of terminalSignalSkippedRefs) {
+    for (const candidate of skipped) {
         actions.push({
             ref: candidate.ref,
             mode: "distill-skipped",
-            result: { ok: true, reason: "no new signal since last proposal" },
-        });
-    }
-    // One aggregate row preserves health accounting without restoring the old
-    // O(n) event-write path. The count now exactly matches the signal gate.
-    if (persist && terminalSignalSkippedRefs.length > 0) {
-        appendEvent({
-            eventType: "improve_skipped",
-            ref: undefined,
-            metadata: {
-                reason: "no_new_signal",
-                count: terminalSignalSkippedRefs.length,
+            result: {
+                ok: true,
+                reason: outOfScope.has(candidate.ref)
+                    ? "not retrieved inside the usage window"
+                    : "no new signal since last proposal",
             },
-        }, eventsCtx);
+        });
     }
-}
-/** Bulk per-ref feedback summary in a SINGLE readEvents pass (candidate-gather). */
-function buildFeedbackSummaryMap(args) {
-    const { processableRefs, noFeedbackPool, eventsCtx, feedbackSinceCutoff } = args;
-    // Gap 6: only surface feedback signals from the last 30 days so that
-    // ancient one-off feedback events don't permanently lock an asset into
-    // every improve run. Assets with only stale signals fall through to the
-    // proactive/high-salience fallbacks or are skipped until new signals arrive.
-    // (FEEDBACK_SIGNAL_WINDOW_DAYS / feedbackSinceCutoff are already defined in
-    // Phase 2 above for the signal-delta gate; we reuse them here.)
-    // Pre-compute feedback summary per ref in a SINGLE bulk read so we don't
-    // open state.db once per asset (which caused 5000+ accumulated FDs and a
-    // 2-hour runaway on a 13K-asset stash). Pattern mirrors buildLatestFeedbackTsMap
-    // above: one readEvents() call fetches ALL feedback events, then we aggregate
-    // in-memory by ref — O(1) DB opens regardless of candidate set size.
-    // Cover processableRefs *and* the deferred noFeedbackPool so utility/feedback
-    // ratios are available for any noFeedbackPool ref the fallback lanes rescue below.
-    //
-    // Behavioral note: positive/negative COUNTS are all-time (same as the old
-    // per-ref readEvents call which had no `since` filter); hasSignal is bounded
-    // to feedbackSinceCutoff (same as the old inline `(e.ts ?? "") >= cutoff` guard).
-    const feedbackSummary = new Map();
-    {
-        const feedbackCandidates = [...processableRefs, ...noFeedbackPool];
-        const feedbackCandidateSet = new Set(feedbackCandidates.map((r) => r.ref));
-        // Map each candidate's single durable event key back to its display ref.
-        const feedbackRefByDurableKey = new Map(feedbackCandidates.flatMap((r) => improveStateReadRefs(r.ref, r.itemRef).map((key) => [key, r.ref])));
-        if (feedbackCandidateSet.size > 0) {
-            // Fetch ALL feedback events in one query (no ref filter, no since filter =
-            // single full table scan). Filtering per-ref in memory avoids N sequential
-            // state.db opens — the dominant FD-leak path on large stashes.
-            const { events: allFeedbackEvents } = readEvents({ type: "feedback" }, eventsCtx);
-            for (const e of allFeedbackEvents) {
-                const ref = e.ref ? feedbackRefByDurableKey.get(e.ref) : undefined;
-                if (!ref)
-                    continue;
-                const entry = feedbackSummary.get(ref) ?? { hasSignal: false, positive: 0, negative: 0 };
-                const meta = e.metadata;
-                // hasSignal: only count feedback events within the 30-day window.
-                if (!entry.hasSignal &&
-                    (e.ts ?? "") >= feedbackSinceCutoff &&
-                    meta !== undefined &&
-                    (typeof meta.signal === "string" || typeof meta.note === "string")) {
-                    entry.hasSignal = true;
-                }
-                // positive/negative: all-time counts (no since filter, matching prior behaviour).
-                if (meta?.signal === "positive")
-                    entry.positive++;
-                else if (meta?.signal === "negative")
-                    entry.negative++;
-                feedbackSummary.set(ref, entry);
-            }
-            // Ensure every candidate has an entry (even refs with zero feedback events).
-            for (const ref of feedbackCandidateSet) {
-                if (!feedbackSummary.has(ref)) {
-                    feedbackSummary.set(ref, { hasSignal: false, positive: 0, negative: 0 });
-                }
-            }
-        }
+    if (persist && signalSkipped.length > 0) {
+        recordImproveSkip(eventsCtx, undefined, { reason: "no_new_signal", count: signalSkipped.length });
     }
-    return feedbackSummary;
-}
-/** Retrieval counts + last-use timestamps for the candidate pools (candidate-gather). */
-function fetchRetrievalSignals(args) {
-    const { options, signalFiltered, noFeedbackCandidates, eventsCtx, persist } = args;
-    // Retrieval counts for the zero-feedback pool, hoisted so the Layer-2
-    // proactive-maintenance selector below can reuse them without a second DB pass.
-    // Also fetch lastUseMs here for the proactive-maintenance recency term (plan §WS-1
-    // step 2: recency is MANDATORY — never pinned to floor).
-    let retrievalCounts = new Map();
-    let lastUseMsForProactive = new Map();
-    let dbForRetrieval;
-    try {
-        dbForRetrieval = persist
-            ? openExistingDatabase()
-            : openReadonlyExistingDatabase(undefined, { isolatedSnapshot: true });
-        if (!dbForRetrieval)
-            return { retrievalCounts, lastUseMsForProactive };
-        // usage_events lives in state.db (Chunk-8 WI-8.3); entries stay in index.db,
-        // so the retrieval-count reads take both handles.
-        const dbForRetrievalIndex = dbForRetrieval;
-        if (persist || eventsCtx?.db) {
-            withStateDb((stateDb) => {
-                const showEventCount = countUsageEventsByType(stateDb, "show");
-                if (showEventCount === 0) {
-                    warn("Warning: show events not yet in usage_events — zero-feedback fallback will match only search-retrieved assets.");
-                }
-                // Fetch retrieval counts for ALL candidates — not only the zero-feedback pool.
-                // Previously only noFeedbackCandidates were looked up, so feedback-bearing refs
-                // had retrievalFreq=0 in computeSalience(), collapsing their retrievalSalience
-                // to 0 regardless of actual use. Two assets of the same type — one
-                // heavily-retrieved, one never-touched — would receive identical rankScores.
-                // Fix (WS-1 blocker 3): union the feedback pool into the lookup.
-                const allCandidateRefs = [...new Set([...signalFiltered, ...noFeedbackCandidates].map((r) => r.ref))];
-                retrievalCounts = getRetrievalCounts(dbForRetrievalIndex, stateDb, allCandidateRefs, {
-                    sourceName: options.sourceName,
-                });
-            }, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
-        }
-        lastUseMsForProactive = getLastUseMsByRef(dbForRetrieval, noFeedbackCandidates);
+    if (persist && retrievalSkipped.length > 0) {
+        recordImproveSkip(eventsCtx, undefined, { reason: "not_retrieved", count: retrievalSkipped.length });
     }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort: if DB unavailable, retrievalCounts/lastUseMsForProactive stay empty
+    const blocked = signalSkipped.length + partition.distillOnlyRefs.length;
+    if (blocked > 0) {
+        info(`[improve] ${blocked} of ${partition.preCooldownCount} indexed refs blocked by reflect signal-delta ` +
+            `(${signalSkipped.length} fully skipped, ${partition.distillOnlyRefs.length} routed to distill-only)`);
     }
-    finally {
-        if (dbForRetrieval)
-            closeDatabase(dbForRetrieval);
+    if (retrievalSkipped.length > 0) {
+        info(`[improve] ${retrievalSkipped.length} refs left out: not retrieved in the last ${USAGE_EVENT_RETENTION_DAYS} days, and not new material`);
     }
-    return { retrievalCounts, lastUseMsForProactive };
+    const gates = [
+        {
+            name: "retrieval",
+            removed: retrievalSkipped.length,
+            reason: `no feedback, and neither returned by search, curate or show in the last ${USAGE_EVENT_RETENTION_DAYS} days nor new material improve never processed`,
+        },
+        {
+            name: "signal",
+            removed: signalSkipped.length,
+            reason: "no fresh signal since the last attempt (or an improve-ledger window) and no fallback lane selected the ref",
+        },
+        { name: "disk", removed: missing.length, reason: "backing asset is absent on disk" },
+        { name: "limit", removed: selection.limitRemoved, reason: "deferred by the effective run limit" },
+    ];
+    return {
+        actionableRefs,
+        loopRefs: selection.loopRefs,
+        distillOnlyRefs: selection.distillOnlyRefs,
+        distillCooledRefs: partition.distillCooledRefs,
+        signalBearingSet,
+        coverageGaps,
+        gates,
+        proactive,
+    };
 }
-/** Layer 2 — the proactive-maintenance selector lane (candidate-gather). */
-function selectProactiveMaintenanceLane(args) {
-    const { scope, improveProfile, resolvedPlan, eventsCtx, noFeedbackCandidates, lastReflectProposalTs, lastDistillProposalTs, retrievalCounts, lastUseMsForProactive, persist, } = args;
-    // ── Layer 2: PROACTIVE MAINTENANCE SELECTOR (second eligibility source) ────
-    // The signal-delta gate only surfaces assets with fresh feedback. It never
-    // revisits a stable, high-value asset on a schedule, so on a quiet stash
-    // useful assets drift stale and are never refreshed. When the
-    // `proactiveMaintenance` process is enabled (DEFAULT OFF)
-    // and the run is whole-stash / type scope, this selector ranks the eligible
-    // population by a composite maintenance priority, gates on staleness ("due"),
-    // bounds to top-N, and folds the winners into the SAME candidate set the other
-    // sources feed — so they flow through the existing #580 empty-diff /
-    // cosmetic suppression and additive-distill gates. It adds no new mutation
-    // logic of its own. The due gate doubles as the rotation cooldown: a freshly
-    // reflected asset is excluded until it ages back past `dueDays`, so successive
-    // runs rotate through the due pool rather than re-selecting the same heads.
-    let proactiveRefs = [];
-    let proactiveMaintenanceSummary;
-    let proactivePlan;
-    const proactiveEnabled = scope.mode !== "ref" && resolvedPlan.processes.proactiveMaintenance.enabled;
-    if (proactiveEnabled) {
-        const pmCfg = improveProfile.processes?.proactiveMaintenance;
-        const dueDays = pmCfg?.dueDays ?? DEFAULT_DUE_DAYS;
-        const maxPerRun = pmCfg?.maxPerRun ?? pmCfg?.limit ?? DEFAULT_MAX_PER_RUN;
-        // Candidate population: the zero-feedback / non-signal pool — exactly the
-        // assets the signal-delta gate would NOT pick this run.
-        const pmCandidates = noFeedbackCandidates;
-        const selection = selectProactiveMaintenanceRefs({
-            candidates: pmCandidates,
-            lastReflectTs: lastReflectProposalTs,
-            lastDistillTs: lastDistillProposalTs,
-            retrievalCounts,
-            // WS-1: wire lastUseMs so the recency decay term is genuine (plan §step 2).
-            lastUseMs: lastUseMsForProactive,
-            sizeBytesOf: (r) => {
-                const fp = r.filePath;
-                if (!fp)
-                    return undefined;
-                try {
-                    return fs.statSync(fp).size;
-                }
-                catch {
-                    return undefined;
-                }
-            },
-            dueDays,
-            maxPerRun,
+/** Retrieval counts for every candidate, and last-use times for the zero-feedback pool. */
+function fetchRetrievalSignals(options, signalFiltered, noFeedbackCandidates, eventsCtx, persist) {
+    const out = { retrievalCounts: new Map(), lastUseMs: new Map() };
+    withIndexDb(!persist, (indexDb) => {
+        // usage_events live in state.db, entries in index.db.
+        withRunState(eventsCtx, persist, (stateDb) => {
+            if (countUsageEventsByType(stateDb, "show") === 0) {
+                warn("Warning: show events not yet in usage_events — zero-feedback fallback will match only search-retrieved assets.");
+            }
+            const refs = [...new Set([...signalFiltered, ...noFeedbackCandidates].map((r) => r.ref))];
+            out.retrievalCounts = getRetrievalCounts(indexDb, stateDb, refs, { sourceName: options.sourceName });
         });
-        proactiveRefs = selection.selected;
-        proactiveMaintenanceSummary = {
-            selected: selection.selected.length,
-            dueTotal: selection.dueTotal,
-            neverReflected: selection.neverReflected,
-            selectedRefs: selection.selected.map((entry) => entry.ref),
-        };
-        proactivePlan = {
-            configured: {
-                ...(pmCfg?.dueDays !== undefined ? { dueDays: pmCfg.dueDays } : {}),
-                ...(pmCfg?.maxPerRun !== undefined ? { maxPerRun: pmCfg.maxPerRun } : {}),
-                ...(pmCfg?.limit !== undefined ? { limit: pmCfg.limit } : {}),
-            },
-            effective: { dueDays, maxPerRun },
-            candidatePool: pmCandidates.length,
-            dueTotal: selection.dueTotal,
-            neverReflected: selection.neverReflected,
-            selected: selection.selected.length,
-            selectedRefs: selection.selected.map((entry) => entry.ref),
-        };
-        // Aggregated observability event (never per-ref — avoids the event flood the
-        // Layer-1 work eliminated). Mirrors the `no_new_signal` aggregation pattern.
-        if (persist) {
-            appendEvent({
-                eventType: "proactive_selected",
-                ref: undefined,
-                metadata: {
-                    count: selection.selected.length,
-                    dueTotal: selection.dueTotal,
-                    neverReflected: selection.neverReflected,
-                },
-            }, eventsCtx);
-        }
-        if (selection.selected.length > 0) {
-            info(`[improve] proactive maintenance selected ${selection.selected.length}/${selection.dueTotal} due refs ` +
-                `(${selection.neverReflected} never reflected, dueDays=${dueDays}, maxPerRun=${maxPerRun})`);
-        }
-    }
-    return { proactiveRefs, proactiveMaintenanceSummary, proactivePlan };
+        out.lastUseMs = getLastUseMsByRef(indexDb, noFeedbackCandidates);
+    });
+    return out;
 }
-/** Layer 3 — the high-salience admission gate (#608/#644; candidate-gather). */
-function selectHighSalienceLane(args) {
-    const { options, improveProfile, eventsCtx, noFeedbackCandidates, proactiveRefs, lastReflectProposalTs, persist } = args;
-    // ── Layer 3: HIGH-SALIENCE ADMISSION GATE (#608) ──────────────────────────
-    // Zero-feedback refs whose encoding_salience (set at distill time by
-    // scoreEncodingSalience) exceeds the configured salienceThreshold are admitted
-    // into the improve run even without retrieval or feedback signal. This rescues
-    // newly distilled assets that the stash has not yet surfaced to users.
-    //
-    // Cap: at most 10% of the effective run limit so the lane cannot crowd out
-    // reactive feedback. Requires state.db to have an asset_salience row — refs
-    // without a row (pre-#608 assets still on the type-weight stub) are skipped.
-    //
-    // Cooldown: a ref qualifies at most once — when no prior reflect proposal
-    // exists for it (`!lastReflectProposalTs.has`). Without this guard the lane
-    // re-selects the same high-salience refs on EVERY run (promotion emits a
-    // `promoted` event, not `feedback`, so the ref never leaves
-    // noFeedbackCandidates), burning LLM calls and churning the asset. This
-    // mirrors the same `!lastReflectProposalTs.has(r.ref)` once-per-asset
-    // semantics the other "rescue" lanes share.
-    //
-    // Content-provenance gate (#644 follow-up): the row must ALSO carry a genuine
-    // content-derived encoding score (`isContentEncodingRow`). Otherwise the lane
-    // admits the per-type WEIGHT STUB (skill/agent 0.9, command/workflow 0.8,
-    // lesson 0.75 from DEFAULT_TYPE_ENCODING_WEIGHTS) for every distill-unscored
-    // asset — i.e. "high-salience" degenerates into "is a skill/agent/command/
-    // lesson", which selected the lore-writer type-stub agent on every run. Only
-    // content-scored assets earn the high-salience rescue; type-stub rows must
-    // earn retrieval/feedback signal via the other lanes. This PRESERVES #608's
-    // intent — distilled assets (the lane's real targets) keep their real content
-    // score and still qualify — while cutting the type-stub waste. See §5 F1 of
-    // #608/#644.
-    const highSalienceRefs = [];
-    const salienceCfg = (options.config ?? loadConfig()).improve?.salience;
-    const salienceThreshold = salienceCfg?.salienceThreshold ?? 0.75;
-    const proactiveSelectedSet = new Set(proactiveRefs.map((r) => r.ref));
-    try {
-        if (!persist && !eventsCtx?.db)
-            return highSalienceRefs;
-        withStateDb((dbForHighSalience) => {
-            // Derive the cap from the resolved reflect limit (mirrors improve.ts's
-            // options.limit resolution) so an unbounded whole-stash run does not
-            // collapse the lane to exactly 1 ref via the bare `?? 10` fallback.
-            const effectiveLimit = options.limit ?? improveProfile?.processes?.reflect?.limit ?? improveProfile.limit ?? 10;
-            const highSalienceCap = Math.max(1, Math.floor(effectiveLimit * 0.1));
-            const candidates = noFeedbackCandidates.filter((r) => !proactiveSelectedSet.has(r.ref));
-            // Collect ALL qualifying candidates, then take the top-N BY SCORE — the
-            // previous first-N-in-scan-order break meant a higher-salience candidate
-            // found later in the scan lost its slot to an earlier lower-scoring one.
-            const qualifying = [];
-            for (const r of candidates) {
-                const row = readAssetSalienceForImproveRef(dbForHighSalience, r.ref, r.itemRef);
-                if (row &&
-                    isContentEncodingRow(row) &&
-                    row.encoding_salience >= salienceThreshold &&
-                    !lastReflectProposalTs.has(r.ref)) {
-                    qualifying.push({ ref: r, score: row.encoding_salience });
-                }
-            }
-            qualifying.sort((a, b) => b.score - a.score);
-            for (const q of qualifying.slice(0, highSalienceCap)) {
-                highSalienceRefs.push(q.ref);
-            }
-        }, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
+/**
+ * Proactive maintenance (default off, whole-stash/type runs): revisit stable
+ * assets on a schedule. The due gate doubles as the rotation cooldown: a
+ * freshly reflected asset waits `dueDays` before it is picked again.
+ */
+function selectProactiveMaintenanceLane(args, candidates, snapshot, retrieval, persist) {
+    if (args.scope.mode === "ref" || !args.resolvedPlan.processes.proactiveMaintenance.enabled) {
+        return { proactiveRefs: [] };
     }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort: if DB unavailable, highSalienceRefs stays empty
+    const pmCfg = args.improveProfile.processes?.proactiveMaintenance;
+    const dueDays = pmCfg?.dueDays ?? DEFAULT_DUE_DAYS;
+    const maxPerRun = pmCfg?.maxPerRun ?? pmCfg?.limit ?? DEFAULT_MAX_PER_RUN;
+    const selection = selectProactiveMaintenanceRefs({
+        candidates,
+        lastReflectTs: snapshot.lastReflectAttemptAt,
+        lastDistillTs: snapshot.lastDistillAttemptAt,
+        retrievalCounts: retrieval.retrievalCounts,
+        lastUseMs: retrieval.lastUseMs,
+        sizeBytesOf: (r) => fileSize(r.filePath),
+        dueDays,
+        maxPerRun,
+    });
+    const summary = {
+        selected: selection.selected.length,
+        dueTotal: selection.dueTotal,
+        neverReflected: selection.neverReflected,
+    };
+    if (persist) {
+        appendEvent({
+            eventType: "proactive_selected",
+            ref: undefined,
+            metadata: { count: summary.selected, dueTotal: summary.dueTotal, neverReflected: summary.neverReflected },
+        }, args.eventsCtx);
     }
-    if (highSalienceRefs.length > 0) {
-        info(`[improve] high-salience lane admitted ${highSalienceRefs.length} content-scored ref(s) ` +
-            `(threshold=${salienceThreshold}, requires content-derived encoding_source)`);
+    if (summary.selected > 0) {
+        info(`[improve] proactive maintenance selected ${summary.selected}/${summary.dueTotal} due refs ` +
+            `(${summary.neverReflected} never reflected, dueDays=${dueDays}, maxPerRun=${maxPerRun})`);
     }
-    return highSalienceRefs;
+    const selectedRefs = selection.selected.map((entry) => entry.ref);
+    return {
+        proactiveRefs: selection.selected,
+        proactiveMaintenanceSummary: { ...summary, selectedRefs },
+        proactivePlan: {
+            configured: pickDefined(pmCfg, ["dueDays", "maxPerRun", "limit"]),
+            effective: { dueDays, maxPerRun },
+            candidatePool: candidates.length,
+            ...summary,
+            selectedRefs,
+        },
+    };
 }
 /**
- * Pass: salience-score — the WS-2 outcome loop, the WS-1 salience vector
- * computation (#644 provenance preserved), persistence + rank-change report,
- * and the forgetting-safety injection. The valence-score call sites
- * (computeValenceScore) live VERBATIM inside the outcome/persist sub-passes.
- * Mutates the shared eligibilitySourceByRef map and ref objects in place —
- * attribution identity is load-bearing (see the candidate-gather comments).
+ * High salience: zero-feedback refs whose content-derived encoding score (not
+ * a per-type stub) reaches `salienceThreshold` and that were never reflected,
+ * top-N by score, capped at 10% of the effective limit.
  */
-function scoreSalience(args) {
-    const { scope, options, primaryStashDir, eventsCtx, eligibilitySourceByRef, feedbackSummary, retrievalCounts, signalFiltered, proactiveRefs, highSalienceRefs, forgettingEligibleRefs, persist, } = args;
-    const mergedRefs = args.mergedRefs;
-    // Chunk-5 flip F5e — resolve each candidate's durable item_ref ONCE for this
-    // pass (the write/read key source for the outcome + salience state writers).
-    const itemRefByRef = buildItemRefByRef(mergedRefs);
-    // WS-1 — Unified salience vector (S1 seam).
-    //
-    // WS-1 converges utility, valence, and proactive-maintenance signals into one
-    // `computeSalience()` call per ref, with
-    // three independently-stored sub-scores and one documented rankScore projection.
-    //
-    // Fetch last-use timestamps from the index DB for the full merged set so the
-    // recency term in retrievalSalience is genuinely decayable (plan §WS-1 step 2).
-    // This reuses the index DB opened earlier for retrieval counts; a separate
-    // lightweight open is used here to avoid holding the connection longer than needed.
-    let lastUseMsByRef = new Map();
-    // Health and outcome reporting consume the utility projection.
-    const utilityMap = buildUtilityMap(mergedRefs, !persist);
-    let dbForSalience;
-    try {
-        dbForSalience = persist
-            ? openExistingDatabase()
-            : openReadonlyExistingDatabase(undefined, { isolatedSnapshot: true });
-        if (dbForSalience) {
-            lastUseMsByRef = getLastUseMsByRef(dbForSalience, mergedRefs);
-        }
+function selectHighSalienceLane(options, improveProfile, eventsCtx, candidates, lastReflectAttemptAt, persist) {
+    const threshold = (options.config ?? loadConfig()).improve?.salience?.salienceThreshold ?? 0.75;
+    const effectiveLimit = options.limit ?? improveProfile?.processes?.reflect?.limit ?? improveProfile.limit ?? 10;
+    const selected = withRunState(eventsCtx, persist, (db) => candidates
+        .flatMap((r) => {
+        const row = getAssetSalience(db, keyOf(r));
+        return row &&
+            isContentEncodingRow(row) &&
+            row.encoding_salience >= threshold &&
+            !lastReflectAttemptAt.has(r.ref)
+            ? [{ ref: r, score: row.encoding_salience }]
+            : [];
+    })
+        .sort((a, b) => b.score - a.score)
+        .slice(0, Math.max(1, Math.floor(effectiveLimit * 0.1)))
+        .map((q) => q.ref)) ?? [];
+    if (selected.length > 0) {
+        info(`[improve] high-salience lane admitted ${selected.length} content-scored ref(s) ` +
+            `(threshold=${threshold}, requires content-derived encoding_source)`);
     }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort: if DB unavailable, recency term stays at floor (lastUseMs=0)
-    }
-    finally {
-        if (dbForSalience)
-            closeDatabase(dbForSalience);
-    }
-    const outcomeSalienceByRef = updateOutcomeScores({
-        mergedRefs,
-        itemRefByRef,
-        feedbackSummary,
-        retrievalCounts,
-        lastUseMsByRef,
-        utilityMap,
-        primaryStashDir,
-        eventsCtx,
-        persist,
-    });
-    const { salienceMap, nowForSalience } = computeSalienceVectors({
+    return selected;
+}
+/**
+ * Score the merged refs: update `asset_outcome` (projected on a plan-only
+ * run), compute each salience vector (keeping a stored content-derived
+ * encoding score), then persist it.
+ */
+function scoreSalience(args, mergedRefs, feedback, retrievalCounts, persist) {
+    const { options, eventsCtx } = args;
+    const utilityMap = buildUtilityMap(mergedRefs, !persist);
+    const lastUseMsByRef = withIndexDb(!persist, (db) => getLastUseMsByRef(db, mergedRefs)) ?? new Map();
+    const outcomeSalience = updateOutcomeScores({
         mergedRefs,
-        itemRefByRef,
-        options,
-        eventsCtx,
+        feedback,
         retrievalCounts,
         lastUseMsByRef,
         utilityMap,
-        outcomeSalienceByRef,
-        persist,
-    });
-    const pendingForgettingRefs = persistSalienceAndReportRanks({
-        salienceMap,
-        itemRefByRef,
-        utilityMap,
-        feedbackSummary,
-        options,
+        primaryStashDir: args.primaryStashDir,
         eventsCtx,
-        nowForSalience,
         persist,
     });
-    const finalMergedRefs = applyForgettingSafety({
-        pendingForgettingRefs,
-        scope,
-        mergedRefs,
-        eligibleRefs: forgettingEligibleRefs,
-        allowFallbacks: options.requireFeedbackSignal !== true,
-        eligibilitySourceByRef,
-        highSalienceRefs,
-        proactiveRefs,
-        signalFiltered,
-    });
-    return { mergedRefs: finalMergedRefs, utilityMap, lastUseMsByRef, salienceMap, nowForSalience };
-}
-/** WS-2 — update asset_outcome for the merged set; returns outcomeSalience by ref. */
-function updateOutcomeScores(args) {
-    const { mergedRefs, itemRefByRef, feedbackSummary, retrievalCounts, lastUseMsByRef, utilityMap, primaryStashDir, eventsCtx, persist, } = args;
-    // ── WS-2 Outcome loop ─────────────────────────────────────────────────────
-    //
-    // Update asset_outcome for every ref in the merged set BEFORE computing the
-    // salience vector so the updated outcome_score feeds outcomeSalience this run.
-    //
-    // Inputs per ref:
-    //   - currentRetrievalCount: from retrievalCounts (index DB)
-    //   - lastRetrievedAt: from lastUseMsByRef (utility_scores.last_used_at)
-    //   - negativeFeedbackCount: cumulative negatives from feedbackSummary
-    //   - acceptedChangeCount: accepted proposals for this ref (state.db)
-    //   - valence: net valence from computeValenceScore(feedbackSummary.get(ref))
-    //   - utilityScore: from utilityMap (for warm-start seed on new rows)
-    //
-    // Best-effort: outcome failures never block the salience or ranking pass.
-    const outcomeSalienceByRef = new Map();
-    // Missing state.db is itself a complete snapshot: no prior outcome rows and
-    // no accepted proposals. Project the same warm-start values a live run would
-    // insert, without creating the database merely to represent empty tables.
-    if (!persist && !eventsCtx?.db) {
-        const projectedScores = new Map();
-        const nowForOutcome = Date.now();
-        for (const ref of mergedRefs) {
-            const feedback = feedbackSummary.get(ref.ref) ?? { positive: 0, negative: 0 };
-            const result = projectAssetOutcome(undefined, {
-                ref: outcomeWriteKey(ref.ref, itemRefByRef),
-                currentRetrievalCount: retrievalCounts.get(ref.ref) ?? 0,
-                lastRetrievedAt: lastUseMsByRef.get(ref.ref) ?? 0,
-                acceptedChangeCount: 0,
-                negativeFeedbackCount: feedback.negative,
-                valence: computeValenceScore(feedback).valence,
-                utilityScore: utilityMap.get(ref.ref),
-                now: nowForOutcome,
-            });
-            projectedScores.set(ref.ref, result.outcomeScore);
-        }
-        const maxOutcomeScore = Math.min(OUTCOME_SCORE_MAX, Math.max(0, ...projectedScores.values()));
-        for (const [ref, score] of projectedScores) {
-            outcomeSalienceByRef.set(ref, outcomeScoreToSalience(score, maxOutcomeScore));
+    const outcomeWeightEnabled = (options.config ?? loadConfig()).improve?.salience?.outcomeWeightEnabled !== false;
+    const storedEncoding = new Map();
+    withRunState(eventsCtx, persist, (db) => {
+        for (const r of mergedRefs) {
+            const row = getAssetSalience(db, keyOf(r));
+            if (row && isContentEncodingRow(row))
+                storedEncoding.set(r.ref, row.encoding_salience);
         }
-        return outcomeSalienceByRef;
-    }
-    try {
-        withStateDb((outcomeDb) => {
-            // Count accepted proposals per ref in one pass (avoid N separate queries).
-            // Scoped to primaryStashDir when available so multi-stash installs don't
-            // inflate counts with proposals from other stashes.
-            const acceptedCountByRef = new Map();
-            try {
-                // #858/#859: listStateProposals() now skips-and-warns on individual
-                // unparseable rows (including legacy pre-#578 rows with no
-                // persisted `changes`, which it tolerates directly) instead of
-                // throwing, so this no longer silently zeroes out every ref's
-                // count on a single bad row. The outer try/catch stays as a
-                // defense-in-depth fallback for unexpected failures (e.g. a query
-                // error), not the primary safeguard it used to be.
-                const acceptedProposals = listStateProposals(outcomeDb, {
-                    status: "accepted",
-                    ...(primaryStashDir ? { stashDir: primaryStashDir } : {}),
-                });
-                for (const p of acceptedProposals) {
-                    acceptedCountByRef.set(p.ref, (acceptedCountByRef.get(p.ref) ?? 0) + 1);
-                }
-            }
-            catch {
-                // best-effort: if the query itself fails, accepted counts stay at 0
-            }
-            // Update each ref's outcome row and collect the resulting outcome scores.
-            const rawOutcomeScores = new Map();
-            const projectedByWriteKey = new Map();
-            const nowForOutcome = Date.now();
-            for (const r of mergedRefs) {
-                const fb = feedbackSummary.get(r.ref) ?? { positive: 0, negative: 0 };
-                const valenceResult = computeValenceScore(fb);
-                try {
-                    const writeKey = outcomeWriteKey(r.ref, itemRefByRef);
-                    const inputs = {
-                        // Key by item_ref when resolved, else by the conceptId. Keep
-                        // rawOutcomeScores keyed by r.ref, its in-memory identity.
-                        ref: writeKey,
-                        currentRetrievalCount: retrievalCounts.get(r.ref) ?? 0,
-                        lastRetrievedAt: lastUseMsByRef.get(r.ref) ?? 0,
-                        acceptedChangeCount: acceptedCountByRef.get(r.ref) ?? 0,
-                        negativeFeedbackCount: fb.negative,
-                        valence: valenceResult.valence,
-                        utilityScore: utilityMap.get(r.ref),
-                        now: nowForOutcome,
-                    };
-                    const result = persist
-                        ? updateAssetOutcome(outcomeDb, inputs)
-                        : projectAssetOutcome(getAssetOutcome(outcomeDb, writeKey), inputs);
-                    rawOutcomeScores.set(r.ref, result.outcomeScore);
-                    projectedByWriteKey.set(writeKey, result.outcomeScore);
-                }
-                catch {
-                    // best-effort per-ref: skip this ref's outcome update on failure
-                }
-            }
-            // Compute stash-wide max outcome_score for normalisation (diversity floor).
-            // Read ALL rows (not just this run's batch) so the normalisation is
-            // stash-relative, not pool-relative.
-            let maxOutcomeScore = 0;
-            try {
-                const allOutcomes = getAllAssetOutcomes(outcomeDb);
-                const scoreByRef = new Map(allOutcomes.map((row) => [row.asset_ref, row.outcome_score]));
-                for (const [ref, score] of projectedByWriteKey)
-                    scoreByRef.set(ref, score);
-                for (const score of scoreByRef.values()) {
-                    if (score > maxOutcomeScore)
-                        maxOutcomeScore = score;
-                }
-                // Keep the normalization denominator within the writer's score bounds.
-                maxOutcomeScore = Math.min(maxOutcomeScore, OUTCOME_SCORE_MAX);
-                // Proxy-adequacy tripwire (two-tailed): inverted (corr < −0.3) and
-                // dead (|corr| < 0.1 at n ≥ 500) both emit health events.
-                const adequacy = persist ? computeProxyAdequacy(allOutcomes) : undefined;
-                if (adequacy?.isInverted) {
-                    appendEvent({
-                        eventType: "outcome_proxy_inverted",
-                        ref: undefined,
-                        metadata: {
-                            correlation: adequacy.correlation,
-                            n: adequacy.n,
-                            note: "corr(outcome_score, accepted_change_rate) < −0.3: high-outcome_score assets have LOW accepted-change rates — the proxy's 'doing well' signal is inverted, so the coarse retrieval-delta signal is no longer trustworthy and the 0.10+ rich in-session signal is no longer deferrable. See plan §WS-2 proxy-adequacy tripwire.",
-                        },
-                    }, eventsCtx);
-                }
-                if (adequacy?.isDead) {
-                    appendEvent({
-                        eventType: "outcome_proxy_dead",
-                        ref: undefined,
-                        metadata: {
-                            correlation: adequacy.correlation,
-                            n: adequacy.n,
-                            note: "|corr(outcome_score, accepted_change_rate)| < 0.1 at n ≥ 500: outcome_score is statistically unrelated to improvement outcomes — the proxy is noise, not signal. Rank contributions derived from it are not currently informative.",
-                        },
-                    }, eventsCtx);
-                }
-            }
-            catch {
-                // best-effort: tripwire failure never blocks ranking
-            }
-            // Convert raw outcome scores → normalised outcomeSalience values in [0,1].
-            for (const [ref, score] of rawOutcomeScores) {
-                const normalised = outcomeScoreToSalience(score, maxOutcomeScore);
-                outcomeSalienceByRef.set(ref, normalised);
-            }
-            // Also fetch outcome scores for refs NOT updated this run (stale or absent)
-            // so the outcomeSalience read path works for all refs in the batch.
-            // Chunk-5 flip F5e — query by each missing ref's WRITE key (item_ref,
-            // else bare) and map the stored-key result back to the bare `r.ref`
-            // identity that outcomeSalienceByRef is keyed on.
-            const missingRefs = mergedRefs.map((r) => r.ref).filter((ref) => !rawOutcomeScores.has(ref));
-            if (missingRefs.length > 0) {
-                const refByWriteKey = new Map();
-                for (const ref of missingRefs)
-                    refByWriteKey.set(outcomeWriteKey(ref, itemRefByRef), ref);
-                const storedScores = getOutcomeScoresByRef(outcomeDb, [...refByWriteKey.keys()]);
-                for (const [writeKey, score] of storedScores) {
-                    const bareRef = refByWriteKey.get(writeKey) ?? writeKey;
-                    outcomeSalienceByRef.set(bareRef, outcomeScoreToSalience(score, maxOutcomeScore));
-                }
-            }
-        }, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
-    }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort: outcome failures never block salience computation
-    }
-    return outcomeSalienceByRef;
-}
-/** WS-1 — compute the salience vector per ref (#644 provenance preserved). */
-function computeSalienceVectors(args) {
-    const { mergedRefs, itemRefByRef, options, eventsCtx, retrievalCounts, lastUseMsByRef, utilityMap, outcomeSalienceByRef, persist, } = args;
-    // Compute the salience vector for every ref in the merged set.
-    // retrievalCounts now covers the full candidate set (feedback-bearing + zero-feedback)
-    // so feedback refs get their genuine retrieval frequency, not a 0-floor fallback.
-    // outcomeSalienceByRef is populated by WS-2 above (or empty on first run).
-    //
-    // R1 loop closure: the outcome weight is ON by default (the G2 saturation
-    // cap makes it safe). Operators opt out with
-    // improve.salience.outcomeWeightEnabled: false in the config.
-    const salienceConfig = (options.config ?? loadConfig()).improve?.salience;
-    const outcomeWeightEnabled = salienceConfig?.outcomeWeightEnabled !== false;
+    });
+    const now = Date.now();
     const salienceMap = new Map();
-    const nowForSalience = Date.now();
-    // #644 — preserve content-derived encoding scores across runs.
-    //
-    // Before computing the salience vector, load each ref's stored encoding score
-    // and its provenance. When the stored row carries a genuine content-derived
-    // score (written by the distill path via `scoreEncodingSalience`), pass that
-    // value back in as `inputs.encodingSalience` so `computeSalience` does NOT fall
-    // back to the type-weight stub — keeping both the persisted `encoding_salience`
-    // AND the derived `rank_score` keyed on real novelty/magnitude/prediction-error.
-    // Refs that have never been content-scored keep the type-weight stub fallback.
-    const storedEncodingByRef = new Map();
-    try {
-        if (persist || eventsCtx?.db) {
-            withStateDb((dbForStoredEncoding) => {
-                for (const r of mergedRefs) {
-                    const row = readAssetSalienceForImproveRef(dbForStoredEncoding, r.ref, itemRefByRef.get(r.ref));
-                    if (row && isContentEncodingRow(row)) {
-                        storedEncodingByRef.set(r.ref, row.encoding_salience);
-                    }
-                }
-            }, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
-        }
-    }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort: if DB unavailable, fall back to type-weight stub (prior behaviour)
-    }
     for (const r of mergedRefs) {
-        const type = assetTypeOf(r.ref);
-        const sizeBytes = (() => {
-            const fp = r.filePath;
-            if (!fp)
-                return undefined;
-            try {
-                return fs.statSync(fp).size;
-            }
-            catch {
-                return undefined;
-            }
-        })();
-        const storedEncoding = storedEncodingByRef.get(r.ref);
-        const vector = computeSalience({
+        const encoding = storedEncoding.get(r.ref);
+        salienceMap.set(r.ref, computeSalience({
             ref: r.ref,
-            type,
-            // #644: pass the stored content-derived score (if any) so the type-weight
-            // stub is NOT re-asserted over a real distill-written encoding score.
-            ...(storedEncoding !== undefined ? { encodingSalience: storedEncoding } : {}),
+            type: assetTypeOf(r.ref),
+            ...(encoding !== undefined ? { encodingSalience: encoding } : {}),
             retrievalFreq: retrievalCounts.get(r.ref) ?? 0,
             lastUseMs: lastUseMsByRef.get(r.ref),
             utilityScore: utilityMap.get(r.ref),
-            outcomeSalience: outcomeSalienceByRef.get(r.ref),
-            sizeBytes,
-            now: nowForSalience,
+            outcomeSalience: outcomeSalience.get(r.ref),
+            sizeBytes: fileSize(r.filePath),
+            now,
             outcomeWeightEnabled,
+        }));
+    }
+    if (persist) {
+        withRunState(eventsCtx, persist, (db) => {
+            for (const r of mergedRefs)
+                upsertAssetSalience(db, keyOf(r), salienceMap.get(r.ref), now);
         });
-        salienceMap.set(r.ref, vector);
     }
-    return { salienceMap, nowForSalience };
-}
-/** 1-indexed rank positions sorted by score desc (deterministic ref-asc tie-break). */
-function toRankPositions(scores) {
-    const sorted = [...scores.entries()].sort(([refA, a], [refB, b]) => b !== a ? b - a : refA < refB ? -1 : refA > refB ? 1 : 0);
-    return new Map(sorted.map(([ref], i) => [ref, i + 1]));
+    return salienceMap;
 }
 /**
- * Chunk-5 flip F5e — the durable salience write-key maps for one improve pass.
- * `wk(ref)` is a pool asset's write key (item_ref, else conceptId);
- * `normalizeStoredKey` maps each in-pool asset's durable spelling to its write
- * key (no stashSize double-count); and
- * `refByWriteKey` reverses a write key back to its filesystem-facing bare ref.
+ * Update each ref's outcome row and return its outcome salience, normalized
+ * against the stash-wide maximum. Without state.db on a plan-only run, the
+ * values a live run would insert are projected instead.
  */
-function buildSalienceWriteKeyMaps(itemRefByRef) {
-    const wk = (ref) => salienceWriteKey(ref, itemRefByRef);
-    const normalizeStoredKey = new Map();
-    const refByWriteKey = new Map();
-    for (const [ref, itemRef] of itemRefByRef) {
-        const writeKey = wk(ref);
-        refByWriteKey.set(writeKey, ref);
-        for (const spelling of improveStateReadRefs(ref, itemRef)) {
-            normalizeStoredKey.set(spelling, writeKey);
-        }
-    }
-    return { wk, normalizeStoredKey, refByWriteKey };
-}
-/** Persist salience vectors + the WS-1 step-7 rank-change/forgetting report. */
-function persistSalienceAndReportRanks(args) {
-    const { salienceMap, itemRefByRef, utilityMap, feedbackSummary, options, eventsCtx, nowForSalience, persist } = args;
-    // Chunk-5 flip F5e — the WRITE-key space. salienceMap stays keyed by each
-    // candidate's own short `r.ref`; the state.db boundary keys by item_ref when
-    // available and otherwise by conceptId.
-    const { wk, normalizeStoredKey, refByWriteKey } = buildSalienceWriteKeyMaps(itemRefByRef);
-    // Persist salience vectors to state.db (best-effort, non-blocking).
-    // The canonical store enables WS-3 homeostatic demotion and WS-2 outcome reads.
-    //
-    // Forgetting-safety report (plan §WS-1 step 7) — stash-wide rank comparison:
-    //
-    // BEFORE persisting the new rankScores, read ALL existing rows from state.db
-    // (not just the per-run candidate pool). This gives stash-wide rank positions so
-    // the top-200/below-500 thresholds are meaningful.
-    //
-    // Two distinct scenarios:
-    //
-    // A. First WS-1 run (table empty): the old stash-wide combinedEligibilityScore
-    //    ordering was never persisted in state.db (asset_salience is a new WS-1 table).
-    //    However, the old formula's inputs are available in-scope for every candidate
-    //    in the current pool: utility comes from utilityMap and the attention term
-    //    from feedbackSummary (positive/negative counts). We reconstruct the old
-    //    combinedEligibilityScore = utility * UTILITY_WEIGHT + attention * FEEDBACK_WEIGHT
-    //    for every ref in salienceMap and rank them, giving a candidate-pool-scoped
-    //    old ordering. This is a partial reconstruction (only current-pool refs, not
-    //    stash-wide), but it is the most faithful comparison possible at cutover and
-    //    allows the top-200→below-500 forgetting guard to fire if the formula change
-    //    dramatically reorders the candidate pool.
-    //    WS-1 step 7 — the stash-wide
-    //    ordering was unreconstructable (no prior state.db snapshot), so this candidate-
-    //    pool partial reconstruction is the documented resolution for the first-run case.
-    //    Emit `improve_salience_first_run` to mark the cutover moment and include the
-    //    reconstructed comparison result in the metadata.
-    //
-    // B. Subsequent runs (table has rows): use ALL existing rows as old ranks, merge
-    //    them with the current run's salienceMap updates for new ranks, and call
-    //    buildRankChangeReport with stash-wide positions. This detects real rank drift
-    //    — e.g. a retrieval-pattern shift causing a previously top-200 asset to slip
-    //    below position 500.
-    //
-    // Measurement-protocol deferral (plan §269, Part-V):
-    // The Part-V T0 baseline (scripts/akm-eval + health report) and the throughput/
-    // quality gate are deferred pending owner sign-off. Full measurement requires a
-    // before/after `akm health` report. Owner-acknowledged deferral: WS-2 landing
-    // will re-introduce outcome salience and trigger the full re-tuning pass at that
-    // time. salience.ts already accepts outcomeSalience directly as an input
-    // (see SalienceInputs.outcomeSalience); no separate hook is needed.
-    //
-    // Forgetting-safety collection: populated inside scenario B below, consumed
-    // after the try/catch to union candidates into mergedRefs before the sort.
-    // Only refs from a real pre-existing ordering (scenario B) are collected;
-    // empty on scenario A or when no candidates dropped below the threshold.
-    let pendingForgettingRefs = [];
-    try {
-        if (!persist && !eventsCtx?.db)
-            return pendingForgettingRefs;
-        withStateDb((stateDb) => {
-            // Step 7: stash-wide rank-change report BEFORE overwriting the table.
-            //
-            // Load ALL existing rows so rank positions are stash-relative, not pool-relative.
-            // Source-scope by the `<bundle>//` prefix and fold each in-pool asset's
-            // stored spelling onto its single write key so
-            // the merge below never double-counts one asset across two spellings.
-            const allStoredScores = getAllRankScores(stateDb);
-            const existingAllScores = new Map();
-            for (const [ref, score] of allStoredScores) {
-                if (options.sourceName) {
-                    const boundary = ref.indexOf("//");
-                    const prefix = boundary >= 0 ? ref.slice(0, boundary) : undefined;
-                    const belongs = prefix === options.sourceName;
-                    if (!belongs)
-                        continue;
-                }
-                existingAllScores.set(normalizeStoredKey.get(ref) ?? ref, score);
-            }
-            if (existingAllScores.size === 0) {
-                // Scenario A: first WS-1 run — table empty.
-                //
-                // Reconstruct the old combinedEligibilityScore ordering for the current
-                // candidate pool using inputs that are already in-scope: utility from
-                // utilityMap and the attention term from feedbackSummary (positive/negative
-                // counts). Old formula: score = utility * UTILITY_WEIGHT + attention * FEEDBACK_WEIGHT.
-                //
-                // Limitation: this covers only the current-run candidate pool, not the full
-                // stash. The stash-wide ordering was never persisted (asset_salience is a new
-                // WS-1 table), so this is the most faithful comparison possible at cutover.
-                // WS-1 step 7.
-                const reconstructedOldScores = new Map();
-                for (const ref of salienceMap.keys()) {
-                    const utility = utilityMap.get(ref) ?? 0;
-                    const fb = feedbackSummary.get(ref) ?? { positive: 0, negative: 0 };
-                    const attention = computeValenceScore(fb).attention;
-                    reconstructedOldScores.set(ref, utility * UTILITY_WEIGHT + attention * FEEDBACK_WEIGHT);
-                }
-                // Assign 1-indexed rank positions sorted by score desc (tie-break: ref asc).
-                const oldRanks = toRankPositions(reconstructedOldScores);
-                const newRanks = toRankPositions(new Map([...salienceMap.entries()].map(([ref, v]) => [ref, v.rankScore])));
-                const firstRunReport = buildRankChangeReport(oldRanks, newRanks);
-                if (firstRunReport.forgettingCandidates.length > 0) {
-                    warn(`[improve/salience] WS-1 first-run rank-change report: ${firstRunReport.forgettingCandidates.length} asset(s) fell from top-200 to below position 500 (cutover formula change). ` +
-                        `Top drops: ${firstRunReport.forgettingCandidates
-                            .slice(0, 5)
-                            .map((e) => `${e.ref} (#${e.oldRank}→#${e.newRank})`)
-                            .join(", ")}`);
-                    pendingForgettingRefs = firstRunReport.forgettingCandidates.map((e) => e.ref);
-                }
-                if (persist) {
-                    appendEvent({
-                        eventType: "improve_salience_first_run",
-                        ref: undefined,
-                        metadata: {
-                            candidateCount: salienceMap.size,
-                            note: "first WS-1 salience run — partial reconstruction of old combinedEligibilityScore ordering for candidate pool (stash-wide ordering not available); WS-1 step 7",
-                            forgettingCandidates: firstRunReport.forgettingCandidates.length,
-                            topDrops: firstRunReport.forgettingCandidates.slice(0, 10).map((e) => ({
-                                ref: e.ref,
-                                oldRank: e.oldRank,
-                                newRank: e.newRank,
-                            })),
-                        },
-                    }, eventsCtx);
-                }
-            }
-            else {
-                // Scenario B: subsequent run — compare stash-wide old vs. new ranks.
-                //
-                // Build new scores by merging the full table with this run's updates.
-                // Refs in salienceMap override their stored value; refs not in this run
-                // retain their stored value unchanged. This gives a complete stash-wide
-                // picture of what the new ordering looks like after this run.
-                const mergedNewScores = new Map(existingAllScores);
-                for (const [ref, vector] of salienceMap) {
-                    // Chunk-5 flip F5e — key this run's fresh scores by the WRITE key so
-                    // they overwrite (never duplicate) the same asset's normalized stored row.
-                    mergedNewScores.set(wk(ref), vector.rankScore);
-                }
-                // Assign 1-indexed rank positions sorted by score desc (tie-break: ref asc).
-                const oldRanks = toRankPositions(existingAllScores);
-                const newRanks = toRankPositions(mergedNewScores);
-                const report = buildRankChangeReport(oldRanks, newRanks);
-                if (report.forgettingCandidates.length > 0) {
-                    warn(`[improve/salience] WS-1 rank-change report: ${report.forgettingCandidates.length} asset(s) fell from top-200 to below position 500. ` +
-                        `Top drops: ${report.forgettingCandidates
-                            .slice(0, 5)
-                            .map((e) => `${e.ref} (#${e.oldRank}→#${e.newRank})`)
-                            .join(", ")}`);
-                    // Collect refs for protective consolidation pass (plan §WS-1 step 7).
-                    // These are force-included in the candidate pool (mergedRefs) after
-                    // this try block, bypassing cooldown/signal-delta gating.
-                    // Chunk-5 flip F5e — map an in-pool candidate's write-key spelling
-                    // back to its bare `r.ref` so applyForgettingSafety re-stamps the
-                    // existing pool ref. Other stored spellings stay qualified here;
-                    // the downstream admission boundary resolves them only when they
-                    // match an exact current-plan item_ref.
-                    pendingForgettingRefs = report.forgettingCandidates.map((e) => refByWriteKey.get(e.ref) ?? e.ref);
-                }
-                if (persist) {
-                    appendEvent({
-                        eventType: "improve_salience_rank_change",
-                        ref: undefined,
-                        metadata: {
-                            stashSize: existingAllScores.size,
-                            totalChanged: report.allChanges.length,
-                            forgettingCandidates: report.forgettingCandidates.length,
-                            topDrops: report.forgettingCandidates.slice(0, 10).map((e) => ({
-                                ref: e.ref,
-                                oldRank: e.oldRank,
-                                newRank: e.newRank,
-                            })),
-                        },
-                    }, eventsCtx);
-                }
-            }
-            if (persist) {
-                for (const [ref, vector] of salienceMap) {
-                    // Persist salience under item_ref when resolved, else the conceptId.
-                    upsertAssetSalience(stateDb, wk(ref), vector, nowForSalience);
-                }
-            }
-        }, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
-    }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort: salience persistence failure never blocks ranking
+function updateOutcomeScores(args) {
+    const { mergedRefs, feedback, eventsCtx, persist } = args;
+    const now = Date.now();
+    const inputsFor = (r, accepted) => {
+        const fb = feedback.get(r.ref) ?? { positive: 0, negative: 0 };
+        return {
+            ref: keyOf(r),
+            currentRetrievalCount: args.retrievalCounts.get(r.ref) ?? 0,
+            lastRetrievedAt: args.lastUseMsByRef.get(r.ref) ?? 0,
+            acceptedChangeCount: accepted,
+            negativeFeedbackCount: fb.negative,
+            valence: computeValenceScore(fb).valence,
+            utilityScore: args.utilityMap.get(r.ref),
+            now,
+        };
+    };
+    const out = new Map();
+    if (!persist && !eventsCtx?.db) {
+        const projected = new Map(mergedRefs.map((r) => [r.ref, projectAssetOutcome(undefined, inputsFor(r, 0)).outcomeScore]));
+        const max = Math.min(OUTCOME_SCORE_MAX, Math.max(0, ...projected.values()));
+        for (const [ref, score] of projected)
+            out.set(ref, outcomeScoreToSalience(score, max));
+        return out;
     }
-    return pendingForgettingRefs;
-}
-/**
- * The protective forgetting-safety injection (plan §WS-1 step 7). Returns the
- * (possibly extended) mergedRefs; re-stamps lane attribution in precedence
- * order on the SHARED ref objects and eligibilitySourceByRef map.
- */
-export function applyForgettingSafety(args) {
-    const { pendingForgettingRefs, scope, eligibleRefs, allowFallbacks, eligibilitySourceByRef, highSalienceRefs, proactiveRefs, signalFiltered, } = args;
-    let mergedRefs = args.mergedRefs;
-    // ── Protective consolidation pass (plan §WS-1 step 7) ─────────────────────
-    // Forgetting candidates detected in scenario B are force-injected into
-    // mergedRefs here, BEFORE the effectiveScore sort, bypassing cooldown and
-    // signal-delta gating. The lane may only reuse exact objects from this
-    // invocation's post-cleanup/post-validation plan. Stale, out-of-scope, and
-    // differently-qualified durable state must never synthesize executable work.
-    if (pendingForgettingRefs.length > 0 && scope.mode !== "ref" && allowFallbacks) {
-        const existingRefSet = new Set(mergedRefs.map((r) => r.ref));
-        const eligibleByRef = new Map(eligibleRefs.map((candidate) => [candidate.ref, candidate]));
-        const eligibleByItemRef = new Map();
-        for (const candidate of eligibleRefs) {
-            if (candidate.itemRef)
-                eligibleByItemRef.set(candidate.itemRef, candidate);
-        }
-        const newForgettingRefs = [];
-        const forgettingRefSet = new Set();
-        for (const stateRef of pendingForgettingRefs) {
-            const boundary = stateRef.indexOf("//");
-            const candidate = eligibleByItemRef.get(stateRef) ?? (boundary < 0 ? eligibleByRef.get(bareImproveRef(stateRef)) : undefined);
-            if (!candidate || forgettingRefSet.has(candidate.ref))
-                continue;
-            forgettingRefSet.add(candidate.ref);
-            if (!existingRefSet.has(candidate.ref)) {
-                newForgettingRefs.push(candidate);
-                existingRefSet.add(candidate.ref);
-            }
-            // Always stamp the lane in the attribution map (overwrites weaker lanes;
-            // stronger reactive signals — scope/signal-delta/proactive — are written
-            // after this block so they take precedence).
-            eligibilitySourceByRef.set(candidate.ref, "forgetting-safety");
-        }
-        if (newForgettingRefs.length > 0) {
-            mergedRefs = dedupeRefs([...mergedRefs, ...newForgettingRefs]);
+    withRunState(eventsCtx, persist, (db) => {
+        const accepted = new Map();
+        try {
+            const rows = listStateProposals(db, {
+                status: "accepted",
+                ...(args.primaryStashDir ? { stashDir: args.primaryStashDir } : {}),
+            });
+            for (const p of rows)
+                accepted.set(p.ref, (accepted.get(p.ref) ?? 0) + 1);
         }
-        // Re-stamp attribution for any refs whose lane needs updating.
-        // Precedence (weakest → strongest, each overwrites the previous):
-        //   proactive < forgetting-safety < signal-delta
-        // Scope mode is already excluded by the outer guard (`scope.mode !== "ref"`).
-        // forgetting-safety sits above proactive so that a ref flagged as a
-        // forgetting candidate is always visible to S5/WS-5 as such, even when it
-        // was also due for a proactive maintenance run. signal-delta overrides
-        // forgetting-safety because a ref with fresh feedback is reactive and
-        // doesn't need the protective pass label for measurement purposes.
-        for (const r of highSalienceRefs)
-            eligibilitySourceByRef.set(r.ref, "high-salience");
-        for (const r of proactiveRefs)
-            eligibilitySourceByRef.set(r.ref, "proactive");
-        // Apply forgetting-safety OVER proactive and high-salience (already
-        // stamped in the loop above via
-        // `eligibilitySourceByRef.set(ref, "forgetting-safety")`). No-op here: the
-        // set() calls above for proactive/high-salience overwrite the earlier
-        // forgetting-safety stamp — so we re-apply forgetting-safety now for those
-        // refs that are both forgetting candidates AND in another fallback lane.
-        for (const ref of forgettingRefSet) {
-            eligibilitySourceByRef.set(ref, "forgetting-safety");
+        catch {
+            // Accepted counts stay 0.
         }
-        // signal-delta is the strongest reactive signal and overrides forgetting-safety.
-        for (const r of signalFiltered)
-            eligibilitySourceByRef.set(r.ref, "signal-delta");
-        // Update eligibilitySource on the ref objects themselves for any refs whose
-        // lane changed (covers both new stubs and pre-existing refs).
+        const raw = new Map();
+        const byKey = new Map();
         for (const r of mergedRefs) {
-            r.eligibilitySource = eligibilitySourceByRef.get(r.ref) ?? "unknown";
-        }
-    }
-    return mergedRefs;
-}
-/**
- * Pass: eligibility-filter — replay selection (#610), the no-op dampener sort,
- * coverage gaps, the disk-existence guard, the --limit slice, and the summary
- * info emits.
- */
-async function filterEligibility(args) {
-    const { scope, options, replayEligibleRefs, eventsCtx, salienceMap, eligibilitySourceByRef, distillOnlyRefs, persist, } = args;
-    const { signalAndRetrievalRefs, signalFiltered } = args.summary;
-    const validationFailureRefs = args.validationFailureRefs;
-    const replay = applyReplaySelection({
-        scope,
-        options,
-        plannedRefs: replayEligibleRefs,
-        eventsCtx,
-        mergedRefs: args.mergedRefs,
-        salienceMap,
-        eligibilitySourceByRef,
-        persist,
-    });
-    const mergedRefs = replay.mergedRefs;
-    const { replayRefSet, replayBudget } = replay;
-    // Build no-op map for consolidation-selection dampener (plan §WS-1 step 8).
-    // Reads consecutive_no_ops from the SAME pinned db handle used elsewhere in
-    // this function.  The effective score is used ONLY for processing/selection
-    // order — the persisted rank_score in asset_salience is never mutated here.
-    const noOpMap = new Map();
-    try {
-        const noOpDb = eventsCtx?.db ?? (persist && eventsCtx?.dbPath ? openStateDatabase(eventsCtx.dbPath) : null);
-        if (noOpDb) {
-            const ownsNoOpDb = !eventsCtx?.db;
             try {
-                for (const r of mergedRefs) {
-                    noOpMap.set(r.ref, readConsecutiveNoOpsForImproveRef(noOpDb, r.ref, r.itemRef));
-                }
+                const inputs = inputsFor(r, accepted.get(r.ref) ?? 0);
+                const result = persist
+                    ? updateAssetOutcome(db, inputs)
+                    : projectAssetOutcome(getAssetOutcome(db, inputs.ref), inputs);
+                raw.set(r.ref, result.outcomeScore);
+                byKey.set(inputs.ref, result.outcomeScore);
             }
-            finally {
-                if (ownsNoOpDb)
-                    noOpDb.close();
-            }
-        }
-    }
-    catch {
-        // best-effort: dampener failure never blocks selection
-    }
-    // Sort by effective selection score (desc), with explicit ref-string tie-break
-    // for determinism.  The effective score applies the consolidation-selection
-    // dampener: assets that have been repeatedly skipped (consecutive_no_ops >=
-    // THRESHOLD) are penalised by FACTOR so they sort after peers with similar
-    // rankScore.  The persisted rank_score is left unchanged — this is the whole
-    // point of the dampener (stable assets stay fully retrievable).
-    //
-    // WIRING NOTE (plan §WS-1 step 8 / "consolidation-selection" disambiguation):
-    // "consolidation-selection" in the plan refers to THIS reflect/distill
-    // eligibility ordering — i.e. which assets are chosen for the reflect/distill
-    // LLM pass — NOT to akmConsolidate (the cluster-merge phase at ~line 1994,
-    // which runs earlier and never reads noOpMap).  The no-op counter originates
-    // from no-change reflect / quality-rejected distill outcomes; the dampener
-    // suppresses repeated LLM attempts on those same assets without touching their
-    // persisted rank_score (so they remain fully retrievable).
-    //
-    // This is the only ranking path. The eligibilitySource lanes (signal-delta /
-    // proactive / high-salience) survive as labels set above.
-    const effectiveScore = (ref) => {
-        const rankScore = salienceMap.get(ref)?.rankScore ?? 0;
-        const noOps = noOpMap.get(ref) ?? 0;
-        return noOps >= SALIENCE_NO_OP_DAMPEN_THRESHOLD ? rankScore * SALIENCE_NO_OP_DAMPEN_FACTOR : rankScore;
-    };
-    const sorted = [...mergedRefs].sort((a, b) => {
-        const scoreA = effectiveScore(a.ref);
-        const scoreB = effectiveScore(b.ref);
-        if (scoreB !== scoreA)
-            return scoreB - scoreA;
-        // Stable tie-break: deterministic regardless of input ordering.
-        return a.ref < b.ref ? -1 : a.ref > b.ref ? 1 : 0;
-    });
-    // Phase 0: surface coverage gaps from zero-result search queries
-    let coverageGaps = [];
-    try {
-        const dbForGaps = persist
-            ? openExistingDatabase()
-            : openReadonlyExistingDatabase(undefined, { isolatedSnapshot: true });
-        if (dbForGaps) {
-            try {
-                coverageGaps = getZeroResultSearches(dbForGaps);
-            }
-            finally {
-                closeDatabase(dbForGaps);
+            catch {
+                // This ref keeps its stored score.
             }
         }
-    }
-    catch (err) {
-        rethrowIfTestIsolationError(err);
-        // best-effort
-    }
-    const diskCheck = await dropRefsMissingOnDisk({ sorted, options, eventsCtx, persist });
-    const assetMissingOnDisk = diskCheck.assetMissingOnDisk;
-    const actionableRefs = diskCheck.actionableRefs;
-    // Re-split actionableRefs (sorted) into reflect-path vs distill-only-path while
-    // preserving sort order. distillOnlyRefs participate in the sort so --limit
-    // picks them by score, not by arbitrary position.
-    // ── Phase 5: --limit applies to the post-cooldown actionable set ──────────
-    //
-    // #610 ADDITIVITY: replay-lane refs are budgeted SEPARATELY from the --limit
-    // fresh slice. Without this split, a high-rankScore replay ref could sort above
-    // a fresh ref in the single combined slice and STEAL its slot (violating AC2).
-    // We partition into the replay lane vs the rest, apply --limit to the
-    // non-replay (fresh) refs only, then APPEND up to `replayBudget` replay refs
-    // after the fresh slice. Sort order within each partition is preserved.
-    //
-    // Default replayBudget=0 reduces this to the exact pre-#610 expression: with no
-    // replay refs, `nonReplayLoop === allLoopRefs`, so `baseLoop === old slice` and
-    // `replayLoop.slice(0, 0) === []` — byte-identical.
-    const selection = selectEffectiveImproveRefs({
-        rankedRefs: actionableRefs,
-        distillOnlyRefs,
-        limit: options.limit,
-        replayBudget,
-    });
-    const loopRefs = selection.loopRefs;
-    const distillOnlyRefsResult = selection.distillOnlyRefs;
-    if (signalAndRetrievalRefs.length > 0) {
-        info(`[improve] ${signalAndRetrievalRefs.length} refs with usage signals (${signalFiltered.length} feedback${replayRefSet.size > 0 ? `, ${replayRefSet.size} replay` : ""})`);
-    }
-    if (validationFailureRefs.size > 0) {
-        info(`[improve] ${validationFailureRefs.size} with validation failures excluded`);
-    }
-    if (persist && assetMissingOnDisk.length > 0) {
-        info(`[improve] ${assetMissingOnDisk.length} candidates dropped — file not on disk`);
-    }
-    const deferredCount = actionableRefs.length - loopRefs.length;
-    info(`[improve] ${actionableRefs.length} actionable; ${loopRefs.length} will be processed` +
-        (options.limit && deferredCount > 0 ? ` (--limit ${options.limit} applied; ${deferredCount} deferred)` : ""));
-    return {
-        loopRefs,
-        actionableRefs,
-        distillOnlyRefs: distillOnlyRefsResult,
-        coverageGaps,
-        limitRemoved: selection.limitRemoved,
-        missingDiskCount: assetMissingOnDisk.length,
-        replayBudget,
-        preDiskRefs: sorted,
-    };
-}
-/** The #610 bounded, additive replay-selection lane (eligibility-filter). */
-function applyReplaySelection(args) {
-    const { scope, options, plannedRefs, eventsCtx, salienceMap, eligibilitySourceByRef, persist } = args;
-    let mergedRefs = args.mergedRefs;
-    // ── REPLAY SELECTION layer (#610) ─────────────────────────────────────────
-    // Bounded, ADDITIVE replay budget: up to `replayBudget` top-salience refs are
-    // revisited even with zero reactive signal (no feedback, no retrieval) and
-    // regardless of cooldown — exactly like the forgetting-safety lane, replay is
-    // injected AFTER cooldown/signal-delta partitioning so it bypasses those gates.
-    //
-    // Strictly additive: the replay slice is appended AFTER the --limit fresh slice
-    // (see the loopRefs partition below), so it can never shrink the fresh-ref set.
-    // Replay is the WEAKEST lane — it only stamps refs no other lane already claimed,
-    // and budget is spent only on refs not already in mergedRefs (so a stronger lane
-    // never has its budget wasted or its label overwritten).
-    //
-    // Default replayBudget=0 ⇒ this whole block is a no-op (no DB open, no event,
-    // no mergedRefs mutation), preserving byte-identical pre-#610 selection behavior.
-    const replayBudget = (options.config ?? loadConfig()).improve?.salience?.replayBudget ?? 0;
-    const replayRefSet = new Set();
-    if (replayBudget > 0 && scope.mode !== "ref" && !options.requireFeedbackSignal) {
+        // Normalize stash-wide (every row, this run's overlaid), within the writer's bound.
+        let max = 0;
         try {
-            if (!persist && !eventsCtx?.db)
-                return { mergedRefs, replayRefSet, replayBudget };
-            withStateDb((replayDb) => {
-                const alreadyInPool = new Set(mergedRefs.map((r) => r.ref));
-                const storedRankScores = getAllRankScores(replayDb);
-                const plannedByRef = new Map(plannedRefs.map((planned) => [planned.ref, planned]));
-                const plannedByItemRef = new Map();
-                for (const planned of plannedRefs) {
-                    if (planned.itemRef)
-                        plannedByItemRef.set(planned.itemRef, planned);
-                }
-                // Replay can only revisit an entry selected into THIS invocation's
-                // source/type plan. Match durable item_ref rows by exact provenance;
-                // legacy bare rows may match the current plan's concept ref. Folding
-                // both spellings onto the planned ref also prevents duplicate budget
-                // spend when old and current state rows coexist.
-                const allRankScores = new Map();
-                for (const [stateRef, score] of storedRankScores) {
-                    const boundary = stateRef.indexOf("//");
-                    if (options.sourceName && (boundary < 0 || stateRef.slice(0, boundary) !== options.sourceName))
-                        continue;
-                    const planned = plannedByItemRef.get(stateRef) ?? (boundary < 0 ? plannedByRef.get(bareImproveRef(stateRef)) : undefined);
-                    if (!planned)
-                        continue;
-                    const previous = allRankScores.get(planned.ref);
-                    if (previous === undefined || score > previous)
-                        allRankScores.set(planned.ref, score);
-                }
-                // Candidate universe = every current-plan salience match NOT already in the
-                // pool, ordered by rank_score desc with a deterministic ref-string tie-break
-                // (mirrors the main sort). Converged refs (consecutive_no_ops >= dampener
-                // threshold) are fully EXCLUDED — a stronger skip than the dampener (which
-                // only halves order).
-                let convergedSkipped = 0;
-                const candidates = [];
-                for (const [ref, rankScore] of allRankScores) {
-                    if (alreadyInPool.has(ref))
-                        continue;
-                    const planned = plannedByRef.get(ref);
-                    if (!planned)
-                        continue;
-                    const noOps = readConsecutiveNoOpsForImproveRef(replayDb, ref, planned.itemRef);
-                    if (noOps >= SALIENCE_NO_OP_DAMPEN_THRESHOLD) {
-                        convergedSkipped++;
-                        continue;
-                    }
-                    candidates.push({ planned, rankScore });
-                }
-                candidates.sort((a, b) => b.rankScore !== a.rankScore
-                    ? b.rankScore - a.rankScore
-                    : a.planned.ref < b.planned.ref
-                        ? -1
-                        : a.planned.ref > b.planned.ref
-                            ? 1
-                            : 0);
-                const candidatePool = candidates.length;
-                const selected = candidates.slice(0, replayBudget);
-                const newReplayRefs = [];
-                for (const { planned } of selected) {
-                    const ref = planned.ref;
-                    replayRefSet.add(ref);
-                    newReplayRefs.push({
-                        ...planned,
-                        eligibilitySource: "replay",
-                    });
-                    // Seed the salienceMap so the sort/effectiveScore can rank the replay ref.
-                    if (!salienceMap.has(ref)) {
-                        salienceMap.set(ref, {
-                            encoding: 0,
-                            outcome: 0,
-                            retrieval: 0,
-                            rankScore: allRankScores.get(ref) ?? 0,
-                        });
-                    }
-                }
-                if (newReplayRefs.length > 0) {
-                    mergedRefs = dedupeRefs([...mergedRefs, ...newReplayRefs]);
-                    // Replay is the WEAKEST lane: stamp 'replay' ONLY for refs not already
-                    // keyed by a stronger lane.
-                    for (const ref of replayRefSet) {
-                        if (!eligibilitySourceByRef.has(ref))
-                            eligibilitySourceByRef.set(ref, "replay");
-                    }
-                    for (const r of mergedRefs) {
-                        r.eligibilitySource = eligibilitySourceByRef.get(r.ref) ?? "unknown";
-                    }
-                }
-                // Aggregated observability event (never per-ref).
-                if (persist) {
-                    appendEvent({
-                        eventType: "improve_replay_selected",
-                        ref: undefined,
-                        metadata: {
-                            count: newReplayRefs.length,
-                            budget: replayBudget,
-                            convergedSkipped,
-                            candidatePool,
-                        },
-                    }, eventsCtx);
-                }
-            }, { path: eventsCtx?.dbPath, borrowed: eventsCtx?.db });
+            const scores = new Map(getAllAssetOutcomes(db).map((row) => [row.asset_ref, row.outcome_score]));
+            for (const [key, score] of byKey)
+                scores.set(key, score);
+            for (const score of scores.values())
+                if (score > max)
+                    max = score;
+            max = Math.min(max, OUTCOME_SCORE_MAX);
         }
-        catch (err) {
-            rethrowIfTestIsolationError(err);
-            // best-effort: if DB unavailable, replayRefSet stays empty
+        catch {
+            max = 0;
         }
-    }
-    return { mergedRefs, replayRefSet, replayBudget };
+        for (const [ref, score] of raw)
+            out.set(ref, outcomeScoreToSalience(score, max));
+        const missing = mergedRefs.filter((r) => !raw.has(r.ref));
+        if (missing.length > 0) {
+            const refByKey = new Map(missing.map((r) => [keyOf(r), r.ref]));
+            for (const [key, score] of getOutcomeScoresByRef(db, [...refByKey.keys()])) {
+                out.set(refByKey.get(key) ?? key, outcomeScoreToSalience(score, max));
+            }
+        }
+    });
+    return out;
 }
-/** The final disk-existence guard + its aggregated audit event (eligibility-filter). */
-async function dropRefsMissingOnDisk(args) {
-    const { sorted, options, eventsCtx, persist } = args;
-    // actionableRefs is the post-cooldown, post-validation, post-signal, post-sort
-    // set — i.e. the genuinely processable refs in priority order. Note: this is
-    // a semantic shift from earlier code where actionableRefs was the pre-cooldown
-    // sorted set; the new meaning matches reality and is documented on
-    // ImprovePreparationResult.actionableRefs.
-    //
-    // Final guard: drop any candidate whose backing file is no longer on disk.
-    // Phase 1 validation captures missing files at the start of preparation, but
-    // the gap between that check and dispatch can be minutes on large stashes —
-    // long enough for a checkpoint / git checkout / external cleanup to delete
-    // the asset. Empirically (improve-critical-review 2026-05-20) the single
-    // biggest reject category was "Asset no longer exists on disk" (604/1407 =
-    // 43%), meaning reflect/distill was producing proposals against deleted refs.
-    // A cheap existsSync per surviving candidate eliminates that wasted work.
-    const assetMissingOnDisk = [];
-    const existsCheckedActionable = [];
+/** Drop candidates whose file vanished since planning, with one aggregate event. */
+async function dropRefsMissingOnDisk(sorted, options, eventsCtx, persist) {
+    const actionableRefs = [];
+    const missing = [];
     for (const candidate of sorted) {
-        // #591: prefer the path pre-resolved at planning time (synchronous
-        // existsSync) over a serial async DB lookup per ref.
         const filePath = candidate.filePath && fs.existsSync(candidate.filePath)
             ? candidate.filePath
             : await findAssetFilePath(candidate.ref, options.stashDir);
-        if (filePath && fs.existsSync(filePath)) {
-            existsCheckedActionable.push(candidate);
-        }
-        else {
-            assetMissingOnDisk.push(candidate.ref);
-        }
+        if (filePath && fs.existsSync(filePath))
+            actionableRefs.push(candidate);
+        else
+            missing.push(candidate.ref);
     }
-    // #592 audit: one summary event instead of one per missing ref. Normally
-    // tiny, but a stash deletion racing the run could make this O(n) sequential
-    // state.db writes. `refs` is capped so the metadata row stays bounded.
-    if (persist && assetMissingOnDisk.length > 0) {
-        appendEvent({
-            eventType: "improve_skipped",
-            ref: undefined,
-            metadata: {
-                reason: "asset_missing_on_disk",
-                count: assetMissingOnDisk.length,
-                refs: assetMissingOnDisk.slice(0, 50),
-            },
-        }, eventsCtx);
+    if (persist && missing.length > 0) {
+        recordImproveSkip(eventsCtx, undefined, {
+            reason: "asset_missing_on_disk",
+            count: missing.length,
+            refs: missing.slice(0, 50),
+        });
     }
-    return { actionableRefs: existsCheckedActionable, assetMissingOnDisk };
+    return { actionableRefs, missing };
 }