attocode 0.2.4 → 0.2.5

package/dist/src/integrations/swarm/swarm-execution.js

@@ -0,0 +1,1021 @@
+/**
+ * Swarm Execution — Task dispatch loop, wave management, and completion handling.
+ *
+ * Extracted from swarm-orchestrator.ts (Phase 3a).
+ * Contains: executeWaves, executeWave, dispatchTask, handleTaskCompletion.
+ */
+import { getTaskTypeConfig } from './types.js';
+import { selectAlternativeModel } from './model-selector.js';
+import { evaluateWorkerOutput, runPreFlightChecks, checkArtifacts, checkArtifactsEnhanced, runConcreteChecks } from './swarm-quality-gate.js';
+import { classifySwarmFailure } from './failure-classifier.js';
+import { isHollowCompletion, FAILURE_INDICATORS, hasFutureIntentLanguage } from './swarm-helpers.js';
+import { reviewWave, saveCheckpoint, emitBudgetUpdate, getEffectiveRetries, getSwarmProgressSummary, getModelHealthSummary, extractFileArtifacts, } from './swarm-lifecycle.js';
+import { tryResilienceRecovery, rescueCascadeSkipped, assessAndAdapt, shouldAutoSplit, judgeSplit, recordRateLimit, isCircuitBreakerActive, decreaseStagger, getStaggerMs, } from './swarm-recovery.js';
+// ─── Wave Execution ─────────────────────────────────────────────────────
+/**
+ * Execute all waves in sequence, with review after each.
+ */
+export async function executeWaves(ctx, recoveryState, getStatus) {
+    let waveIndex = ctx.taskQueue.getCurrentWave();
+    const totalWaves = ctx.taskQueue.getTotalWaves();
+    const dispatchLeaseStaleMs = ctx.config.dispatchLeaseStaleMs ?? 5 * 60 * 1000;
+    while (waveIndex < totalWaves && !ctx.cancelled) {
+        const activeTaskIds = new Set(ctx.workerPool.getActiveWorkerStatus().map(w => w.taskId));
+        const recovered = ctx.taskQueue.reconcileStaleDispatched({
+            staleAfterMs: dispatchLeaseStaleMs,
+            activeTaskIds,
+        });
+        if (recovered.length > 0) {
+            ctx.logDecision('lease-recovery', `Recovered ${recovered.length} stale dispatched task(s)`, recovered.join(', '));
+        }
+        const readyTasks = ctx.taskQueue.getReadyTasks();
+        const queueStats = ctx.taskQueue.getStats();
+        // F18: Skip empty waves
+        if (readyTasks.length === 0 && queueStats.running === 0 && queueStats.ready === 0) {
+            ctx.logDecision('wave-skip', `Skipping waves ${waveIndex + 1}-${totalWaves}: no dispatchable tasks remain`, `Stats: ${queueStats.completed} completed, ${queueStats.failed} failed, ${queueStats.skipped} skipped`);
+            break;
+        }
+        ctx.emit({
+            type: 'swarm.wave.start',
+            wave: waveIndex + 1,
+            totalWaves,
+            taskCount: readyTasks.length,
+        });
+        // Dispatch tasks up to concurrency limit
+        await executeWave(ctx, recoveryState, readyTasks, getStatus);
+        // Wave complete stats
+        const afterStats = ctx.taskQueue.getStats();
+        const waveCompleted = afterStats.completed - (queueStats.completed);
+        const waveFailed = afterStats.failed - (queueStats.failed);
+        const waveSkipped = afterStats.skipped - (queueStats.skipped);
+        ctx.emit({
+            type: 'swarm.wave.complete',
+            wave: waveIndex + 1,
+            totalWaves,
+            completed: waveCompleted,
+            failed: waveFailed,
+            skipped: waveSkipped,
+        });
+        // Wave failure recovery: if ALL tasks in a wave failed, retry with adapted context
+        if (waveCompleted === 0 && waveFailed > 0 && readyTasks.length > 0) {
+            ctx.emit({ type: 'swarm.wave.allFailed', wave: waveIndex + 1 });
+            ctx.logDecision('wave-recovery', `Entire wave ${waveIndex + 1} failed (${waveFailed} tasks)`, 'Checking if budget allows retry with adapted strategy');
+            const budgetRemaining = ctx.budgetPool.hasCapacity();
+            const failedWaveTasks = readyTasks.filter(t => {
+                const task = ctx.taskQueue.getTask(t.id);
+                return task && task.status === 'failed' && task.attempts < (ctx.config.workerRetries + 1);
+            });
+            if (budgetRemaining && failedWaveTasks.length > 0) {
+                for (const t of failedWaveTasks) {
+                    const task = ctx.taskQueue.getTask(t.id);
+                    if (!task)
+                        continue;
+                    task.status = 'ready';
+                    task.retryContext = {
+                        previousFeedback: 'All tasks in this batch failed. Try a fundamentally different approach — the previous strategy did not work.',
+                        previousScore: 0,
+                        attempt: task.attempts,
+                        previousModel: task.assignedModel,
+                        swarmProgress: getSwarmProgressSummary(ctx),
+                    };
+                }
+                ctx.logDecision('wave-recovery', `Re-queued ${failedWaveTasks.length} tasks with adapted retry context`, 'Budget allows retry');
+                await executeWave(ctx, recoveryState, failedWaveTasks.map(t => ctx.taskQueue.getTask(t.id)).filter(t => t.status === 'ready'), getStatus);
+            }
+        }
+        // F5: Adaptive re-decomposition signal
+        const waveTotal = waveCompleted + waveFailed + waveSkipped;
+        const waveSuccessRate = waveTotal > 0 ? waveCompleted / waveTotal : 0;
+        if (waveSuccessRate < 0.5 && waveTotal >= 2) {
+            ctx.logDecision('decomposition-quality', `Wave ${waveIndex + 1} success rate ${(waveSuccessRate * 100).toFixed(0)}% (${waveCompleted}/${waveTotal})`, 'Low success rate may indicate decomposition quality issues');
+        }
+        // V2: Review wave outputs
+        const review = await reviewWave(ctx, waveIndex);
+        if (review && review.fixupTasks.length > 0) {
+            await executeWave(ctx, recoveryState, review.fixupTasks, getStatus);
+        }
+        // Rescue cascade-skipped tasks that can still run
+        const rescued = rescueCascadeSkipped(ctx);
+        if (rescued.length > 0) {
+            ctx.logDecision('cascade-rescue', `Rescued ${rescued.length} cascade-skipped tasks after wave ${waveIndex + 1}`, rescued.map(t => t.id).join(', '));
+            await executeWave(ctx, recoveryState, rescued, getStatus);
+        }
+        // Reset quality circuit breaker at wave boundary
+        if (recoveryState.qualityGateDisabledModels.size > 0) {
+            recoveryState.qualityGateDisabledModels.clear();
+            recoveryState.perModelQualityRejections.clear();
+            ctx.logDecision('quality-circuit-breaker', `Re-enabled quality gates for all models at wave ${waveIndex + 1} boundary`, 'Each wave gets a fresh quality evaluation window');
+        }
+        // F3: Log budget reallocation after wave completion
+        const budgetStats = ctx.budgetPool.getStats();
+        ctx.logDecision('budget-reallocation', `After wave ${waveIndex + 1}: ${budgetStats.tokensRemaining} tokens remaining (${(budgetStats.utilization * 100).toFixed(0)}% utilized)`, '');
+        ctx.budgetPool.reallocateUnused(budgetStats.tokensRemaining);
+        // F21: Mid-swarm situational assessment
+        await assessAndAdapt(ctx, recoveryState, waveIndex);
+        // V2: Checkpoint after each wave
+        saveCheckpoint(ctx, `wave-${waveIndex}`);
+        // Advance to next wave
+        if (!ctx.taskQueue.advanceWave())
+            break;
+        waveIndex++;
+    }
+}
+/**
+ * Execute a single wave's tasks with concurrency control.
+ */
+export async function executeWave(ctx, recoveryState, tasks, getStatus) {
+    // Dispatch initial batch with stagger
+    let taskIndex = 0;
+    while (taskIndex < tasks.length && ctx.workerPool.availableSlots > 0 && !ctx.cancelled) {
+        if (isCircuitBreakerActive(recoveryState, ctx)) {
+            const waitMs = recoveryState.circuitBreakerUntil - Date.now();
+            if (waitMs > 0)
+                await new Promise(resolve => setTimeout(resolve, waitMs));
+            continue;
+        }
+        const task = tasks[taskIndex];
+        await dispatchTask(ctx, recoveryState, task, getStatus);
+        taskIndex++;
+        if (taskIndex < tasks.length && ctx.workerPool.availableSlots > 0) {
+            await new Promise(resolve => setTimeout(resolve, getStaggerMs(recoveryState)));
+        }
+    }
+    // Process completions and dispatch more tasks as slots open
+    while (ctx.workerPool.activeCount > 0 && !ctx.cancelled) {
+        const completed = await ctx.workerPool.waitForAny();
+        if (!completed)
+            break;
+        await handleTaskCompletion(ctx, recoveryState, completed.taskId, completed.result, completed.startedAt, getStatus);
+        emitBudgetUpdate(ctx);
+        ctx.emit({ type: 'swarm.status', status: getStatus() });
+        // Dispatch more tasks if slots available and tasks remain
+        while (taskIndex < tasks.length && ctx.workerPool.availableSlots > 0 && !ctx.cancelled) {
+            const task = tasks[taskIndex];
+            if (task.status === 'ready') {
+                await dispatchTask(ctx, recoveryState, task, getStatus);
+                if (taskIndex + 1 < tasks.length && ctx.workerPool.availableSlots > 0) {
+                    await new Promise(resolve => setTimeout(resolve, getStaggerMs(recoveryState)));
+                }
+            }
+            taskIndex++;
+        }
+        // Also check for cross-wave ready tasks to fill slots
+        if (ctx.workerPool.availableSlots > 0 && !isCircuitBreakerActive(recoveryState, ctx)) {
+            const moreReady = ctx.taskQueue.getAllReadyTasks()
+                .filter(t => !ctx.workerPool.getActiveWorkerStatus().some(w => w.taskId === t.id));
+            for (let i = 0; i < moreReady.length; i++) {
+                if (ctx.workerPool.availableSlots <= 0)
+                    break;
+                await dispatchTask(ctx, recoveryState, moreReady[i], getStatus);
+                if (i + 1 < moreReady.length && ctx.workerPool.availableSlots > 0) {
+                    await new Promise(resolve => setTimeout(resolve, getStaggerMs(recoveryState)));
+                }
+            }
+        }
+    }
+    // F20: Re-dispatch pass — after all workers finish, budget may have been freed
+    if (!ctx.cancelled && ctx.budgetPool.hasCapacity()) {
+        const stillReady = ctx.taskQueue.getAllReadyTasks()
+            .filter(t => !ctx.workerPool.getActiveWorkerStatus().some(w => w.taskId === t.id));
+        if (stillReady.length > 0) {
+            ctx.logDecision('budget-redispatch', `Budget freed after wave — re-dispatching ${stillReady.length} ready task(s)`, `Budget: ${JSON.stringify(ctx.budgetPool.getStats())}`);
+            for (const task of stillReady) {
+                if (ctx.workerPool.availableSlots <= 0 || !ctx.budgetPool.hasCapacity())
+                    break;
+                await dispatchTask(ctx, recoveryState, task, getStatus);
+                if (ctx.workerPool.availableSlots > 0) {
+                    await new Promise(resolve => setTimeout(resolve, getStaggerMs(recoveryState)));
+                }
+            }
+            while (ctx.workerPool.activeCount > 0 && !ctx.cancelled) {
+                const completed = await ctx.workerPool.waitForAny();
+                if (!completed)
+                    break;
+                await handleTaskCompletion(ctx, recoveryState, completed.taskId, completed.result, completed.startedAt, getStatus);
+                emitBudgetUpdate(ctx);
+                ctx.emit({ type: 'swarm.status', status: getStatus() });
+            }
+        }
+    }
+}
+// ─── Task Dispatch ──────────────────────────────────────────────────────
+/**
+ * Dispatch a single task to a worker.
+ */
+export async function dispatchTask(ctx, recoveryState, task, _getStatus) {
+    const worker = ctx.workerPool.selectWorker(task);
+    if (!worker) {
+        ctx.logDecision('no-worker', `${task.id}: no worker for type ${task.type}`, '');
+        if (task.attempts > 0) {
+            const syntheticTaskResult = { success: false, output: '', tokensUsed: 0, costUsed: 0, durationMs: 0, model: 'none' };
+            const syntheticSpawn = { success: false, output: '', metrics: { tokens: 0, duration: 0, toolCalls: 0 } };
+            if (await tryResilienceRecovery(ctx, recoveryState, task, task.id, syntheticTaskResult, syntheticSpawn)) {
+                return;
+            }
+        }
+        ctx.taskQueue.markFailedWithoutCascade(task.id, 0);
+        ctx.taskQueue.triggerCascadeSkip(task.id);
+        ctx.emit({
+            type: 'swarm.task.failed',
+            taskId: task.id,
+            error: `No worker available for task type: ${task.type}`,
+            attempt: task.attempts,
+            maxAttempts: 0,
+            willRetry: false,
+            failureMode: 'error',
+        });
+        return;
+    }
+    try {
+        // Pre-dispatch auto-split for critical-path bottlenecks
+        if (shouldAutoSplit(ctx, task)) {
+            try {
+                const splitResult = await judgeSplit(ctx, task);
+                if (splitResult.shouldSplit && splitResult.subtasks) {
+                    task.status = 'dispatched';
+                    ctx.taskQueue.replaceWithSubtasks(task.id, splitResult.subtasks);
+                    ctx.emit({
+                        type: 'swarm.task.resilience',
+                        taskId: task.id,
+                        strategy: 'auto-split',
+                        succeeded: true,
+                        reason: `Pre-dispatch split into ${splitResult.subtasks.length} parallel subtasks`,
+                        artifactsFound: 0,
+                        toolCalls: 0,
+                    });
+                    return;
+                }
+            }
+            catch (err) {
+                ctx.logDecision('auto-split', `${task.id}: split judge failed — ${err.message}`, '');
+            }
+        }
+        ctx.totalDispatches++;
+        const dispatchedModel = task.assignedModel ?? worker.model;
+        ctx.taskQueue.markDispatched(task.id, dispatchedModel);
+        if (task.assignedModel && task.assignedModel !== worker.model) {
+            ctx.logDecision('failover', `Dispatching ${task.id} with failover model ${task.assignedModel} (worker default: ${worker.model})`, 'Retry model override is active');
+        }
+        await ctx.workerPool.dispatch(task, worker);
+        ctx.emit({
+            type: 'swarm.task.dispatched',
+            taskId: task.id,
+            description: task.description,
+            model: dispatchedModel,
+            workerName: worker.name,
+            toolCount: worker.allowedTools?.length ?? -1,
+            tools: worker.allowedTools,
+            retryContext: task.retryContext,
+            fromModel: task.retryContext ? task.retryContext.previousModel : undefined,
+            attempts: task.attempts,
+        });
+    }
+    catch (error) {
+        const errorMsg = error.message;
+        // F20: Budget exhaustion is NOT a task failure
+        if (errorMsg.includes('Budget pool exhausted')) {
+            task.status = 'ready';
+            ctx.logDecision('budget-pause', `Cannot dispatch ${task.id}: budget exhausted — task kept ready for potential re-dispatch`, `Budget stats: ${JSON.stringify(ctx.budgetPool.getStats())}`);
+            return;
+        }
+        ctx.errors.push({
+            taskId: task.id,
+            phase: 'dispatch',
+            message: errorMsg,
+            recovered: false,
+        });
+        ctx.logDecision('dispatch-error', `${task.id}: dispatch failed: ${errorMsg.slice(0, 100)}`, `attempts: ${task.attempts}`);
+        if (task.attempts > 0) {
+            const syntheticTaskResult = { success: false, output: '', tokensUsed: 0, costUsed: 0, durationMs: 0, model: 'none' };
+            const syntheticSpawn = { success: false, output: '', metrics: { tokens: 0, duration: 0, toolCalls: 0 } };
+            if (await tryResilienceRecovery(ctx, recoveryState, task, task.id, syntheticTaskResult, syntheticSpawn)) {
+                ctx.errors[ctx.errors.length - 1].recovered = true;
+                return;
+            }
+        }
+        ctx.taskQueue.markFailedWithoutCascade(task.id, 0);
+        ctx.taskQueue.triggerCascadeSkip(task.id);
+        ctx.emit({
+            type: 'swarm.task.failed',
+            taskId: task.id,
+            error: errorMsg,
+            attempt: task.attempts,
+            maxAttempts: 1 + ctx.config.workerRetries,
+            willRetry: false,
+            failureMode: 'error',
+        });
+    }
+}
+// ─── Task Completion Handling ───────────────────────────────────────────
+/**
+ * Handle a completed task: quality gate, bookkeeping, retry logic, model health, failover.
+ */
+export async function handleTaskCompletion(ctx, recoveryState, taskId, spawnResult, startedAt, _getStatus) {
+    const task = ctx.taskQueue.getTask(taskId);
+    if (!task)
+        return;
+    // Guard: task was terminally resolved while its worker was running
+    if ((task.status === 'skipped' || task.status === 'failed') && !task.pendingCascadeSkip)
+        return;
+    // V7: Global dispatch cap
+    const maxDispatches = ctx.config.maxDispatchesPerTask ?? 5;
+    if (task.attempts >= maxDispatches) {
+        const durationMs = Date.now() - startedAt;
+        const taskResult = ctx.workerPool.toTaskResult(spawnResult, task, durationMs);
+        ctx.totalTokens += taskResult.tokensUsed;
+        ctx.totalCost += taskResult.costUsed;
+        if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+            return;
+        }
+        ctx.taskQueue.markFailedWithoutCascade(taskId, 0);
+        ctx.taskQueue.triggerCascadeSkip(taskId);
+        ctx.emit({
+            type: 'swarm.task.failed',
+            taskId,
+            error: `Dispatch cap reached (${maxDispatches} attempts)`,
+            attempt: task.attempts,
+            maxAttempts: maxDispatches,
+            willRetry: false,
+            failureMode: task.failureMode,
+        });
+        ctx.logDecision('dispatch-cap', `${taskId}: hard cap reached (${task.attempts}/${maxDispatches})`, 'No more retries — resilience recovery also failed');
+        return;
+    }
+    const durationMs = Date.now() - startedAt;
+    const taskResult = ctx.workerPool.toTaskResult(spawnResult, task, durationMs);
+    // Track model usage
+    const model = task.assignedModel ?? 'unknown';
+    const usage = ctx.modelUsage.get(model) ?? { tasks: 0, tokens: 0, cost: 0 };
+    usage.tasks++;
+    usage.tokens += taskResult.tokensUsed;
+    usage.cost += taskResult.costUsed;
+    ctx.modelUsage.set(model, usage);
+    ctx.totalTokens += taskResult.tokensUsed;
+    ctx.totalCost += taskResult.costUsed;
+    if (taskResult.budgetUtilization) {
+        ctx.logDecision('budget-utilization', `${taskId}: token ${taskResult.budgetUtilization.tokenPercent}%, iter ${taskResult.budgetUtilization.iterationPercent}%`, `model=${model}, tokens=${taskResult.tokensUsed}, duration=${durationMs}ms`);
+    }
+    // V10: Emit per-attempt event
+    ctx.emit({
+        type: 'swarm.task.attempt',
+        taskId,
+        attempt: task.attempts,
+        model,
+        success: spawnResult.success,
+        durationMs,
+        toolCalls: spawnResult.metrics.toolCalls ?? 0,
+        failureMode: !spawnResult.success ? task.failureMode : undefined,
+        qualityScore: taskResult.qualityScore,
+        output: taskResult.output.slice(0, 500),
+    });
+    if (!spawnResult.success) {
+        return handleFailedCompletion(ctx, recoveryState, task, taskId, spawnResult, taskResult, model, durationMs, startedAt, maxDispatches);
+    }
+    // V6: Hollow completion detection
+    if (isHollowCompletion(spawnResult, task.type, ctx.config)) {
+        return handleHollowCompletion(ctx, recoveryState, task, taskId, spawnResult, taskResult, model, maxDispatches);
+    }
+    // F4: Task had pendingCascadeSkip but produced non-hollow results
+    if (task.pendingCascadeSkip) {
+        const cachedReport = checkArtifacts(task);
+        const preFlight = runPreFlightChecks(task, taskResult, ctx.config, cachedReport);
+        if (preFlight && !preFlight.passed) {
+            task.pendingCascadeSkip = undefined;
+            task.status = 'skipped';
+            ctx.logDecision('cascade-skip', `${taskId}: pending cascade skip honored (pre-flight failed: ${preFlight.feedback})`, '');
+            ctx.emit({ type: 'swarm.task.skipped', taskId, reason: `cascade skip honored — output failed pre-flight: ${preFlight.feedback}` });
+            return;
+        }
+        task.pendingCascadeSkip = undefined;
+        task.status = 'dispatched';
+        ctx.logDecision('cascade-skip', `${taskId}: pending cascade skip overridden — worker produced valid output`, '');
+    }
+    // Record model health on success
+    ctx.healthTracker.recordSuccess(model, durationMs);
+    decreaseStagger(recoveryState);
+    // Run quality gate if enabled
+    const effectiveRetries = getEffectiveRetries(ctx, task);
+    const recentRLCount = recoveryState.recentRateLimits.filter(t => t > Date.now() - 30_000).length;
+    const isLastAttempt = task.attempts >= (effectiveRetries + 1);
+    const shouldRunQualityGate = ctx.config.qualityGates
+        && !recoveryState.qualityGateDisabledModels.has(model)
+        && !isLastAttempt
+        && Date.now() >= recoveryState.circuitBreakerUntil
+        && recentRLCount < 2;
+    const cachedArtifactReport = checkArtifacts(task);
+    if (shouldRunQualityGate) {
+        const rejected = await runQualityGate(ctx, recoveryState, task, taskId, spawnResult, taskResult, model, effectiveRetries, cachedArtifactReport);
+        if (rejected)
+            return;
+    }
+    // F7: When quality gate was skipped, still run pre-flight + concrete checks
+    if (!shouldRunQualityGate && ctx.config.qualityGates) {
+        const preFlight = runPreFlightChecks(task, taskResult, ctx.config, cachedArtifactReport);
+        if (preFlight && !preFlight.passed) {
+            taskResult.qualityScore = preFlight.score;
+            taskResult.qualityFeedback = preFlight.feedback;
+            ctx.qualityRejections++;
+            const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+            if (canRetry) {
+                ctx.retries++;
+            }
+            else {
+                ctx.logDecision('preflight-reject', `${taskId}: pre-flight failed: ${preFlight.feedback}`, '');
+                if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                    return;
+                }
+                ctx.taskQueue.triggerCascadeSkip(taskId);
+            }
+            ctx.emit({
+                type: 'swarm.quality.rejected',
+                taskId,
+                score: preFlight.score,
+                feedback: preFlight.feedback,
+                artifactCount: 0,
+                outputLength: taskResult.output.length,
+                preFlightReject: true,
+            });
+            return;
+        }
+        // F2: Run concrete validation when pre-flight passes but gate was skipped
+        if (ctx.config.enableConcreteValidation !== false) {
+            const concreteResult = runConcreteChecks(task, taskResult);
+            if (!concreteResult.passed) {
+                taskResult.qualityScore = 2;
+                taskResult.qualityFeedback = `Concrete validation failed: ${concreteResult.issues.join('; ')}`;
+                ctx.qualityRejections++;
+                const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+                if (canRetry) {
+                    ctx.retries++;
+                }
+                else {
+                    ctx.logDecision('concrete-reject', `${taskId}: concrete validation failed: ${concreteResult.issues.join('; ')}`, '');
+                    if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                        return;
+                    }
+                    ctx.taskQueue.triggerCascadeSkip(taskId);
+                }
+                ctx.emit({
+                    type: 'swarm.quality.rejected',
+                    taskId,
+                    score: 2,
+                    feedback: taskResult.qualityFeedback,
+                    artifactCount: 0,
+                    outputLength: taskResult.output.length,
+                    preFlightReject: false,
+                });
+                return;
+            }
+        }
+    }
+    // Final completion guard: block "narrative success" for action tasks
+    const completionGuard = ctx.config.completionGuard ?? {};
+    const rejectFutureIntentOutputs = completionGuard.rejectFutureIntentOutputs ?? true;
+    const requireConcreteArtifactsForActionTasks = completionGuard.requireConcreteArtifactsForActionTasks ?? true;
+    const typeConfig = getTaskTypeConfig(task.type, ctx.config);
+    const artifactReport = checkArtifactsEnhanced(task, taskResult);
+    const filesOnDisk = artifactReport.files.filter(f => f.exists && f.sizeBytes > 0).length;
+    const hasConcreteArtifacts = filesOnDisk > 0 || (taskResult.filesModified?.length ?? 0) > 0;
+    const isActionTask = !!typeConfig.requiresToolCalls;
+    if (rejectFutureIntentOutputs && hasFutureIntentLanguage(taskResult.output ?? '')) {
+        taskResult.qualityScore = 1;
+        taskResult.qualityFeedback = 'Completion rejected: output indicates pending, unexecuted work';
+        const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+        if (canRetry) {
+            ctx.retries++;
+        }
+        else {
+            if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                return;
+            }
+            ctx.taskQueue.triggerCascadeSkip(taskId);
+        }
+        ctx.emit({
+            type: 'swarm.quality.rejected',
+            taskId,
+            score: 1,
+            feedback: taskResult.qualityFeedback,
+            artifactCount: filesOnDisk,
+            outputLength: taskResult.output.length,
+            preFlightReject: true,
+            filesOnDisk,
+        });
+        return;
+    }
+    if (requireConcreteArtifactsForActionTasks && isActionTask && !hasConcreteArtifacts) {
+        taskResult.qualityScore = 1;
+        taskResult.qualityFeedback = 'Completion rejected: action task produced no concrete artifacts';
+        const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+        if (canRetry) {
+            ctx.retries++;
+        }
+        else {
+            if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                return;
+            }
+            ctx.taskQueue.triggerCascadeSkip(taskId);
+        }
+        ctx.emit({
+            type: 'swarm.quality.rejected',
+            taskId,
+            score: 1,
+            feedback: taskResult.qualityFeedback,
+            artifactCount: filesOnDisk,
+            outputLength: taskResult.output.length,
+            preFlightReject: true,
+            filesOnDisk,
+        });
+        return;
+    }
+    // Task passed — mark completed
+    ctx.taskQueue.markCompleted(taskId, taskResult);
+    ctx.hollowStreak = 0;
+    recoveryState.taskTimeoutCounts.delete(taskId);
+    // H6: Post findings to blackboard
+    if (ctx.blackboard && taskResult.findings) {
+        try {
+            for (const finding of taskResult.findings) {
+                ctx.blackboard.post(`swarm-worker-${taskId}`, {
+                    topic: `swarm.task.${task.type}`,
+                    content: finding,
+                    type: 'progress',
+                    confidence: (taskResult.qualityScore ?? 3) / 5,
+                    tags: ['swarm', task.type],
+                    relatedFiles: task.targetFiles,
+                });
+            }
+        }
+        catch {
+            ctx.errors.push({
+                taskId,
+                phase: 'execution',
+                message: 'Failed to post findings to blackboard',
+                recovered: true,
+            });
+        }
+    }
+    ctx.emit({
+        type: 'swarm.task.completed',
+        taskId,
+        success: true,
+        tokensUsed: taskResult.tokensUsed,
+        costUsed: taskResult.costUsed,
+        durationMs: taskResult.durationMs,
+        qualityScore: taskResult.qualityScore,
+        qualityFeedback: taskResult.qualityFeedback,
+        output: taskResult.output,
+        closureReport: taskResult.closureReport,
+        toolCalls: spawnResult.metrics.toolCalls,
+    });
+}
+// ─── Internal Helpers ───────────────────────────────────────────────────
+/**
+ * Handle the case where a worker fails (spawnResult.success === false).
+ */
+async function handleFailedCompletion(ctx, recoveryState, task, taskId, spawnResult, taskResult, model, durationMs, startedAt, maxDispatches) {
+    const failure = classifySwarmFailure(spawnResult.output, spawnResult.metrics.toolCalls);
+    const { failureClass, retryable, errorType, failureMode, reason } = failure;
+    const isTimeout = failureMode === 'timeout';
+    const isRateLimited = failureClass === 'rate_limited';
+    const isSpendLimit = failureClass === 'provider_spend_limit';
+    const isNonRetryable = !retryable;
+    ctx.healthTracker.recordFailure(model, errorType);
+    ctx.emit({ type: 'swarm.model.health', record: { model, ...getModelHealthSummary(ctx, model) } });
+    task.failureMode = failureMode;
+    if (isRateLimited) {
+        recordRateLimit(recoveryState, ctx);
+    }
+    // F25a: Consecutive timeout tracking
+    if (isTimeout) {
+        const count = (recoveryState.taskTimeoutCounts.get(taskId) ?? 0) + 1;
+        recoveryState.taskTimeoutCounts.set(taskId, count);
+        const timeoutLimit = ctx.config.consecutiveTimeoutLimit ?? 3;
+        ctx.logDecision('timeout-tracking', `${taskId}: consecutive timeout ${count}/${timeoutLimit}`, '');
+        if (count >= timeoutLimit) {
+            let failoverSucceeded = false;
+            if (ctx.config.enableModelFailover) {
+                const capability = getTaskTypeConfig(task.type, ctx.config).capability ?? 'code';
+                const alternative = selectAlternativeModel(ctx.config.workers, model, capability, ctx.healthTracker);
+                if (alternative) {
+                    ctx.emit({
+                        type: 'swarm.model.failover',
+                        taskId,
+                        fromModel: model,
+                        toModel: alternative.model,
+                        reason: 'consecutive-timeouts',
+                    });
+                    task.assignedModel = alternative.model;
+                    recoveryState.taskTimeoutCounts.set(taskId, 0);
+                    ctx.logDecision('failover', `Timeout failover ${taskId}: ${model} → ${alternative.model}`, `${count} consecutive timeouts`);
+                    failoverSucceeded = true;
+                }
+            }
+            if (!failoverSucceeded) {
+                task.failureMode = 'timeout';
+                const timeoutTaskResult = ctx.workerPool.toTaskResult(spawnResult, task, Date.now() - startedAt);
+                if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, timeoutTaskResult, spawnResult)) {
+                    recoveryState.taskTimeoutCounts.delete(taskId);
+                    return;
+                }
+                ctx.taskQueue.markFailedWithoutCascade(taskId, 0);
+                ctx.taskQueue.triggerCascadeSkip(taskId);
+                ctx.emit({
+                    type: 'swarm.task.failed',
+                    taskId,
+                    error: `${count} consecutive timeouts — no alternative model available`,
+                    attempt: task.attempts,
+                    maxAttempts: maxDispatches,
+                    willRetry: false,
+                    failureMode: 'timeout',
+                    failureClass: 'timeout',
+                    retrySuppressed: true,
+                    retryReason: 'Consecutive timeout limit reached with no alternative model',
+                });
+                ctx.logDecision('timeout-early-fail', `${taskId}: ${count} consecutive timeouts, no alt model — resilience recovery also failed`, '');
+                recoveryState.taskTimeoutCounts.delete(taskId);
+                return;
+            }
+        }
+    }
+    else {
+        recoveryState.taskTimeoutCounts.delete(taskId);
+    }
+    // V2: Model failover on retryable rate limits
+    if (isRateLimited && ctx.config.enableModelFailover) {
+        const capability = getTaskTypeConfig(task.type, ctx.config).capability ?? 'code';
+        const alternative = selectAlternativeModel(ctx.config.workers, model, capability, ctx.healthTracker);
+        if (alternative) {
+            ctx.emit({
+                type: 'swarm.model.failover',
+                taskId,
+                fromModel: model,
+                toModel: alternative.model,
+                reason: errorType,
+            });
+            task.assignedModel = alternative.model;
+            ctx.logDecision('failover', `Switched ${taskId} from ${model} to ${alternative.model}`, `${errorType} error`);
+        }
+    }
+    // V5/V7: Store error context so retry gets different prompt
+    if (!(isRateLimited || isSpendLimit)) {
+        const timeoutSeconds = isTimeout ? Math.round(durationMs / 1000) : 0;
+        task.retryContext = {
+            previousFeedback: isTimeout
+                ? `Previous attempt timed out after ${timeoutSeconds}s. You must complete this task more efficiently — work faster, use fewer tool calls, and produce your result sooner.`
+                : spawnResult.output.slice(0, 2000),
+            previousScore: 0,
+            attempt: task.attempts,
+            previousModel: model,
+            previousFiles: taskResult.filesModified,
+            swarmProgress: getSwarmProgressSummary(ctx),
+        };
+        ctx.sharedContextEngine.reportFailure(taskId, {
+            action: task.description.slice(0, 200),
+            error: spawnResult.output.slice(0, 500),
+        });
+    }
+    // V7: Reset hollow streak on non-hollow failure
+    ctx.hollowStreak = 0;
+    // Worker failed — use higher retry limit for rate limit errors
+    const baseRetries = getEffectiveRetries(ctx, task);
+    const retryLimit = isNonRetryable
+        ? 0
+        : isRateLimited
+            ? Math.min(ctx.config.rateLimitRetries ?? 3, baseRetries + 1)
+            : baseRetries;
+    const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, retryLimit);
+    if (isNonRetryable) {
+        ctx.logDecision('retry-suppressed', `${taskId}: ${failureClass}`, reason);
+    }
+    if (canRetry) {
+        ctx.retries++;
+        if (isRateLimited) {
+            const baseDelay = ctx.config.retryBaseDelayMs ?? 5000;
+            const cooldownMs = Math.min(baseDelay * Math.pow(2, task.attempts - 1), 30000);
+            ctx.taskQueue.setRetryAfter(taskId, cooldownMs);
+            ctx.logDecision('rate-limit-cooldown', `${taskId}: ${errorType} cooldown ${cooldownMs}ms, model ${model}`, '');
+        }
+    }
+    else if (!isRateLimited) {
+        if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+            return;
+        }
+        ctx.taskQueue.triggerCascadeSkip(taskId);
+    }
+    else {
+        ctx.taskQueue.triggerCascadeSkip(taskId);
+    }
+    ctx.emit({
+        type: 'swarm.task.failed',
+        taskId,
+        error: spawnResult.output.slice(0, 200),
+        attempt: task.attempts,
+        maxAttempts: 1 + ctx.config.workerRetries,
+        willRetry: canRetry,
+        toolCalls: spawnResult.metrics.toolCalls,
+        failoverModel: task.assignedModel !== model ? task.assignedModel : undefined,
+        failureMode: task.failureMode,
+        failureClass,
+        retrySuppressed: isNonRetryable,
+        retryReason: reason,
+    });
+}
+/**
+ * Handle hollow completion — workers that "succeed" without doing any work.
+ */
+async function handleHollowCompletion(ctx, recoveryState, task, taskId, spawnResult, taskResult, model, maxDispatches) {
+    // F4: Hollow result + pendingCascadeSkip
+    if (task.pendingCascadeSkip) {
+        task.pendingCascadeSkip = undefined;
+        task.status = 'skipped';
+        ctx.totalHollows++;
+        ctx.logDecision('cascade-skip', `${taskId}: pending cascade skip honored (hollow completion)`, '');
+        ctx.emit({ type: 'swarm.task.skipped', taskId, reason: 'cascade skip honored — hollow completion' });
+        return;
+    }
+    task.failureMode = 'hollow';
+    ctx.healthTracker.recordHollow(model);
+    const admitsFailure = spawnResult.success && FAILURE_INDICATORS.some(f => (spawnResult.output ?? '').toLowerCase().includes(f));
+    task.retryContext = {
+        previousFeedback: admitsFailure
+            ? 'Previous attempt reported success but admitted failure (e.g., "budget exhausted", "unable to complete"). You MUST execute tool calls and produce concrete output this time.'
+            : 'Previous attempt produced no meaningful output. Try again with a concrete approach.',
+        previousScore: 1,
+        attempt: task.attempts,
+        previousModel: model,
+        previousFiles: taskResult.filesModified,
+        swarmProgress: getSwarmProgressSummary(ctx),
+    };
+    ctx.sharedContextEngine.reportFailure(taskId, {
+        action: task.description.slice(0, 200),
+        error: 'Hollow completion: worker produced no meaningful output',
+    });
+    // Model failover for hollow completions
+    if (ctx.config.enableModelFailover) {
+        const capability = getTaskTypeConfig(task.type, ctx.config).capability ?? 'code';
+        const alternative = selectAlternativeModel(ctx.config.workers, model, capability, ctx.healthTracker);
+        if (alternative) {
+            ctx.emit({
+                type: 'swarm.model.failover',
+                taskId,
+                fromModel: model,
+                toModel: alternative.model,
+                reason: 'hollow-completion',
+            });
+            task.assignedModel = alternative.model;
+            ctx.logDecision('failover', `Hollow failover ${taskId}: ${model} → ${alternative.model}`, 'Model produced hollow completion');
+        }
+    }
+    const hollowRetries = getEffectiveRetries(ctx, task);
+    const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, hollowRetries);
+    if (canRetry) {
+        ctx.retries++;
+    }
+    else {
+        if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+            return;
+        }
+        ctx.taskQueue.triggerCascadeSkip(taskId);
+    }
+    ctx.emit({
+        type: 'swarm.task.failed',
+        taskId,
+        error: 'Hollow completion: worker used no tools',
+        attempt: task.attempts,
+        maxAttempts: 1 + ctx.config.workerRetries,
+        willRetry: canRetry,
+        toolCalls: spawnResult.metrics.toolCalls,
+        failoverModel: task.assignedModel !== model ? task.assignedModel : undefined,
+        failureMode: 'hollow',
+    });
+    ctx.hollowStreak++;
+    ctx.totalHollows++;
+    ctx.logDecision('hollow-completion', `${taskId}: worker completed with 0 tool calls (streak: ${ctx.hollowStreak}, total hollows: ${ctx.totalHollows}/${ctx.totalDispatches})`, canRetry ? 'Marking as failed for retry' : 'Retries exhausted — hard fail');
+    // B2: Hollow streak handling
+    const HOLLOW_STREAK_THRESHOLD = 3;
+    if (ctx.hollowStreak >= HOLLOW_STREAK_THRESHOLD) {
+        const uniqueModels = new Set(ctx.config.workers.map(w => w.model));
+        const singleModel = uniqueModels.size === 1;
+        const onlyModel = [...uniqueModels][0];
+        const modelUnhealthy = singleModel && !ctx.healthTracker.getAllRecords().find(r => r.model === onlyModel)?.healthy;
+        if (singleModel && modelUnhealthy) {
+            if (ctx.config.enableHollowTermination) {
+                ctx.logDecision('early-termination', `Terminating swarm: ${ctx.hollowStreak} consecutive hollow completions on sole model ${onlyModel}`, 'Single-model swarm with unhealthy model — enableHollowTermination is on');
+                skipRemainingTasksInternal(ctx, `Single-model hollow streak (${ctx.hollowStreak}x on ${onlyModel})`);
+            }
+            else {
+                ctx.logDecision('stall-mode', `${ctx.hollowStreak} consecutive hollows on sole model ${onlyModel} — entering stall mode`, 'Will attempt model failover or simplified retry on next dispatch');
+                ctx.hollowStreak = 0;
+            }
+        }
+    }
+    // V7: Multi-model hollow ratio
+    const minDispatches = ctx.config.hollowTerminationMinDispatches ?? 8;
+    const threshold = ctx.config.hollowTerminationRatio ?? 0.55;
+    if (ctx.totalDispatches >= minDispatches) {
+        const ratio = ctx.totalHollows / ctx.totalDispatches;
+        if (ratio > threshold) {
+            if (ctx.config.enableHollowTermination) {
+                ctx.logDecision('early-termination', `Terminating swarm: hollow ratio ${(ratio * 100).toFixed(0)}% (${ctx.totalHollows}/${ctx.totalDispatches})`, `Exceeds threshold ${(threshold * 100).toFixed(0)}% after ${minDispatches}+ dispatches — enableHollowTermination is on`);
+                skipRemainingTasksInternal(ctx, `Hollow ratio ${(ratio * 100).toFixed(0)}% — models cannot execute tasks`);
+            }
+            else if (!recoveryState.hollowRatioWarned) {
+                recoveryState.hollowRatioWarned = true;
+                ctx.logDecision('stall-warning', `Hollow ratio ${(ratio * 100).toFixed(0)}% (${ctx.totalHollows}/${ctx.totalDispatches})`, 'High hollow rate but continuing — tasks may still recover via resilience');
+            }
+        }
+    }
+}
+/**
+ * Run the quality gate (LLM judge) on a task completion.
+ * Returns true if the task was rejected (caller should return), false if it passed.
+ */
+async function runQualityGate(ctx, recoveryState, task, taskId, spawnResult, taskResult, model, effectiveRetries, cachedArtifactReport) {
+    const judgeModel = ctx.config.hierarchy?.judge?.model
+        ?? ctx.config.qualityGateModel ?? ctx.config.orchestratorModel;
+    const judgeConfig = {
+        model: judgeModel,
+        persona: ctx.config.hierarchy?.judge?.persona,
+    };
+    ctx.emit({ type: 'swarm.role.action', role: 'judge', action: 'quality-gate', model: judgeModel, taskId });
+    const fileArtifacts = extractFileArtifacts(ctx, task, taskResult);
+    const baseThreshold = ctx.config.qualityThreshold ?? 3;
+    const qualityThreshold = task.isFoundation ? Math.max(2, baseThreshold - 1) : baseThreshold;
+    const quality = await evaluateWorkerOutput(ctx.provider, judgeModel, task, taskResult, judgeConfig, qualityThreshold, (resp, purpose) => ctx.trackOrchestratorUsage(resp, purpose), fileArtifacts, ctx.config, cachedArtifactReport);
+    taskResult.qualityScore = quality.score;
+    taskResult.qualityFeedback = quality.feedback;
+    // F11: Foundation tasks that barely pass
+    if (quality.passed && task.isFoundation && quality.score <= baseThreshold - 1) {
+        const concreteResult = runConcreteChecks(task, taskResult);
+        if (!concreteResult.passed) {
+            quality.passed = false;
+            quality.feedback += ` [F11: foundation task barely passed (${quality.score}/${baseThreshold}) but concrete validation failed: ${concreteResult.issues.join('; ')}]`;
+            ctx.logDecision('foundation-concrete-gate', `${taskId}: foundation task scored ${quality.score} (relaxed threshold ${qualityThreshold}) but concrete checks failed — rejecting`, concreteResult.issues.join('; '));
+        }
+    }
+    if (!quality.passed) {
+        // F7: Gate error fallback
+        if (quality.gateError && (ctx.config.enableConcreteValidation !== false)) {
+            const concreteResult = runConcreteChecks(task, taskResult);
+            if (concreteResult.passed) {
+                ctx.logDecision('gate-error-fallback', `${taskId}: gate error but concrete checks passed — tentatively accepting`, quality.gateErrorMessage ?? 'unknown');
+                taskResult.qualityScore = quality.score;
+                taskResult.qualityFeedback = `${quality.feedback} [concrete validation passed — tentative accept]`;
+                recoveryState.perModelQualityRejections.delete(model);
+                return false; // passed
+            }
+            else {
+                ctx.logDecision('gate-error-fallback', `${taskId}: gate error AND concrete checks failed — rejecting`, `Concrete issues: ${concreteResult.issues.join('; ')}`);
+                ctx.qualityRejections++;
+                task.failureMode = 'quality';
+                ctx.healthTracker.recordQualityRejection(model, quality.score);
+                ctx.emit({ type: 'swarm.model.health', record: { model, ...getModelHealthSummary(ctx, model) } });
+                ctx.hollowStreak = 0;
+                task.retryContext = {
+                    previousFeedback: `Gate error + concrete validation failed: ${concreteResult.issues.join('; ')}`,
+                    previousScore: quality.score,
+                    attempt: task.attempts,
+                    previousModel: model,
+                    previousFiles: taskResult.filesModified,
+                    swarmProgress: getSwarmProgressSummary(ctx),
+                };
+                const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+                if (canRetry) {
+                    ctx.retries++;
+                }
+                else {
+                    if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                        return true;
+                    }
+                    ctx.taskQueue.triggerCascadeSkip(taskId);
+                }
+                ctx.emit({
+                    type: 'swarm.quality.rejected',
+                    taskId,
+                    score: quality.score,
+                    feedback: quality.feedback,
+                    artifactCount: fileArtifacts.length,
+                    outputLength: taskResult.output.length,
+                    preFlightReject: false,
+                    filesOnDisk: checkArtifactsEnhanced(task, taskResult).files.filter(f => f.exists && f.sizeBytes > 0).length,
+                });
+                return true;
+            }
+        }
+        else if (!quality.gateError) {
+            // Normal quality rejection
+            ctx.qualityRejections++;
+            task.failureMode = 'quality';
+            ctx.healthTracker.recordQualityRejection(model, quality.score);
+            ctx.emit({ type: 'swarm.model.health', record: { model, ...getModelHealthSummary(ctx, model) } });
+            ctx.hollowStreak = 0;
+            // F7: Per-model circuit breaker
+            if (!quality.preFlightReject) {
+                const QUALITY_CIRCUIT_BREAKER_THRESHOLD = 5;
+                const modelRejections = (recoveryState.perModelQualityRejections.get(model) ?? 0) + 1;
+                recoveryState.perModelQualityRejections.set(model, modelRejections);
+                if (modelRejections >= QUALITY_CIRCUIT_BREAKER_THRESHOLD) {
+                    recoveryState.qualityGateDisabledModels.add(model);
+                    ctx.logDecision('quality-circuit-breaker', `Switched model ${model} to pre-flight-only mode after ${modelRejections} rejections`, 'Skipping LLM judge but keeping pre-flight checks mandatory');
+                }
+            }
+            task.retryContext = {
+                previousFeedback: quality.feedback,
+                previousScore: quality.score,
+                attempt: task.attempts,
+                previousModel: model,
+                previousFiles: taskResult.filesModified,
+                swarmProgress: getSwarmProgressSummary(ctx),
+            };
+            ctx.sharedContextEngine.reportFailure(taskId, {
+                action: task.description.slice(0, 200),
+                error: `Quality gate rejection (score ${quality.score}): ${quality.feedback.slice(0, 300)}`,
+            });
+            // V5: Model failover on quality rejection
+            if (quality.score < qualityThreshold && ctx.config.enableModelFailover && !quality.artifactAutoFail) {
+                const capability = getTaskTypeConfig(task.type, ctx.config).capability ?? 'code';
+                const alternative = selectAlternativeModel(ctx.config.workers, model, capability, ctx.healthTracker);
+                if (alternative) {
+                    ctx.emit({
+                        type: 'swarm.model.failover',
+                        taskId,
+                        fromModel: model,
+                        toModel: alternative.model,
+                        reason: `quality-score-${quality.score}`,
+                    });
+                    task.assignedModel = alternative.model;
+                    ctx.logDecision('failover', `Quality failover ${taskId}: ${model} → ${alternative.model}`, `Score ${quality.score}/5`);
+                }
+            }
+            const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+            if (canRetry) {
+                ctx.retries++;
+            }
+            else {
+                if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                    return true;
+                }
+                ctx.taskQueue.triggerCascadeSkip(taskId);
+            }
+            ctx.emit({
+                type: 'swarm.quality.rejected',
+                taskId,
+                score: quality.score,
+                feedback: quality.feedback,
+                artifactCount: fileArtifacts.length,
+                outputLength: taskResult.output.length,
+                preFlightReject: quality.preFlightReject,
+                filesOnDisk: checkArtifactsEnhanced(task, taskResult).files.filter(f => f.exists && f.sizeBytes > 0).length,
+            });
+            return true;
+        }
+        else {
+            // gateError=true but concrete validation disabled
+            ctx.qualityRejections++;
+            task.failureMode = 'quality';
+            ctx.hollowStreak = 0;
+            task.retryContext = {
+                previousFeedback: quality.feedback,
+                previousScore: quality.score,
+                attempt: task.attempts,
+                previousModel: model,
+                previousFiles: taskResult.filesModified,
+                swarmProgress: getSwarmProgressSummary(ctx),
+            };
+            const canRetry = ctx.taskQueue.markFailedWithoutCascade(taskId, effectiveRetries);
+            if (canRetry) {
+                ctx.retries++;
+            }
+            else {
+                if (await tryResilienceRecovery(ctx, recoveryState, task, taskId, taskResult, spawnResult)) {
+                    return true;
+                }
+                ctx.taskQueue.triggerCascadeSkip(taskId);
+            }
+            ctx.emit({
+                type: 'swarm.quality.rejected',
+                taskId,
+                score: quality.score,
+                feedback: quality.feedback,
+                artifactCount: fileArtifacts.length,
+                outputLength: taskResult.output.length,
+                preFlightReject: false,
+                filesOnDisk: checkArtifactsEnhanced(task, taskResult).files.filter(f => f.exists && f.sizeBytes > 0).length,
+            });
+            return true;
+        }
+    }
+    // Quality passed — reset per-model rejection counter
+    recoveryState.perModelQualityRejections.delete(model);
+    return false; // passed
+}
+/** Internal helper for skipRemainingTasks within execution context */
+function skipRemainingTasksInternal(ctx, reason) {
+    for (const task of ctx.taskQueue.getAllTasks()) {
+        if (task.status === 'pending' || task.status === 'ready') {
+            task.status = 'skipped';
+            ctx.emit({ type: 'swarm.task.skipped', taskId: task.id, reason });
+        }
+    }
+}
+//# sourceMappingURL=swarm-execution.js.map