@principles/pd-cli 1.152.11 → 1.152.13

package/src/commands/diagnose.ts

@@ -1,756 +0,0 @@
-/**
- * pd diagnose run/status commands — Diagnostician execution and status inspection.
- *
- * Usage:
- *   pd diagnose status --task-id <taskId> --workspace <path>
- *   pd diagnose run --task-id <taskId> --workspace <path>
- */
-import {
-  RuntimeStateManager,
-  SqliteHistoryQuery,
-  SqliteContextAssembler,
-  SqliteDiagnosticianCommitter,
-  SqliteTrajectoryLocator,
-  SqliteSourceTraceLocator,
-  StoreEventEmitter,
-  storeEmitter,
-  SplitDiagnosticianRunner,
-  DiagRootCauseRunner,
-  DiagDistillerRunner,
-  DiagRouterRunner,
-  DefaultDiagRootCauseValidator,
-  DefaultDiagDistillerValidator,
-  type DiagnosticianRunnerLike,
-  TestDoubleRuntimeAdapter,
-  PDRuntimeError,
-  isRuntimeConfigError,
-  CandidateIntakeService,
-  run as diagnoseRun,
-  status as diagnoseStatus,
-  PrincipleTreeLedgerAdapter,
-  buildDreamerSeedFromCandidate,
-  findExistingDreamerTask,
-  CANDIDATE_KIND_TO_ROUTE,
-  ROUTE_CHANNEL_MAP,
-  MVP_ENABLED_CHANNELS,
-} from '@principles/core/runtime-v2';
-import type { PDRuntimeAdapter, OutputLanguage } from '@principles/core/runtime-v2';
-import { resolveWorkspaceDir } from '../resolve-workspace.js';
-import { readOutputLanguageFromWorkspace } from '../config-reader.js';
-import { loadPdConfig, resolvePromptFullPipelineSeedMode } from '../services/pd-config-loader.js';
-import { SPLIT_PIPELINE_TOTAL_TIMEOUT_MS, resolveDiagnosticianCapability } from '@principles/core/runtime-v2';
-import { resolveRuntimeAdapterFromConfig, ConfigResolutionError } from '../services/runtime-adapter-resolver.js';
-import { resolveRuntimeFromPdConfig } from '../services/resolve-runtime-from-pd-config.js';
-import { checkAdmissionGate } from './admission-gate.js';
-import { resolveSourcePainIdFromDiagnostician } from './candidate.js';
-import * as path from 'path';
-
-function validateStalledThreshold(val: unknown): number | undefined {
-  if (val === undefined) {
-    return undefined;
-  }
-  const str = String(val).trim();
-  if (!/^[1-9]\d*$/.test(str)) {
-    throw new Error('stalled-threshold must be a positive integer.');
-  }
-  const num = parseInt(str, 10);
-  if (isNaN(num)) {
-    throw new Error('stalled-threshold must be a positive integer.');
-  }
-  return num;
-}
-
-interface DiagnoseStatusOptions {
-  taskId: string;
-  workspace?: string;
-  json?: boolean;
-  stalledThreshold?: unknown;
-}
-
-interface DiagnoseRunOptions {
-  taskId: string;
-  workspace?: string;
-  json?: boolean;
-  runtime?: string;
-  openclawLocal?: boolean;
-  openclawGateway?: boolean;
-  agent?: string;
-  provider?: string;
-  model?: string;
-  apiKeyEnv?: string;
-  baseUrl?: string;
-  maxRetries?: number;
-  timeoutMs?: number;
-  intake?: boolean;
-}
-
-/**
- * pd diagnose status --task-id <taskId> [--workspace <path>] [--json]
- *
- * Inspects the current status of a diagnostician task.
- */
-export async function handleDiagnoseStatus(opts: DiagnoseStatusOptions): Promise<void> {
-  let stalledThresholdSeconds: number | undefined;
-  try {
-    stalledThresholdSeconds = validateStalledThreshold(opts.stalledThreshold);
-  } catch (err) {
-    const msg = err instanceof Error ? err.message : String(err);
-    if (opts.json) {
-      console.log(JSON.stringify({
-        ok: false,
-        reason: 'invalid_stalled_threshold',
-        nextAction: 'Provide a valid positive integer for --stalled-threshold (e.g., --stalled-threshold 300).',
-      }));
-    } else {
-      console.error(`error: ${msg}`);
-      console.error('nextAction: Provide a valid positive integer for --stalled-threshold (e.g., --stalled-threshold 300).');
-    }
-    process.exit(1);
-    return;
-  }
-
-  const workspaceDir = resolveWorkspaceDir(opts.workspace);
-  const stateManager = new RuntimeStateManager({ workspaceDir });
-
-  try {
-    await stateManager.initialize();
-    const result = await diagnoseStatus({
-      taskId: opts.taskId,
-      stateManager,
-      stalledThresholdSeconds,
-    });
-
-    if (!result) {
-      console.error(`Task not found: ${opts.taskId}`);
-      process.exit(1);
-      return;
-    }
-
-    if (opts.json) {
-      console.log(JSON.stringify(result, null, 2));
-      return;
-    }
-
-    console.log(`\nDiagnostician Task: ${result.taskId}\n`);
-    console.log(`  Status:       ${result.status}`);
-    console.log(`  Attempts:     ${result.attemptCount} / ${result.maxAttempts}`);
-    if (result.commitId) {
-      console.log(`  Result Ref:   commit://${result.commitId}`);
-      console.log(`  Commit ID:    ${result.commitId}`);
-      console.log(`  Artifact ID:  ${result.artifactId ?? 'N/A'}`);
-      console.log(`  Candidates:   ${result.candidateCount ?? 0}`);
-    }
-    if (result.lastError) {
-      console.log(`  Last Error:   ${result.lastError}`);
-    }
-    if (result.reason) {
-      console.log(`  Reason:       ${result.reason}`);
-    }
-    if (result.age !== undefined && result.age !== null) {
-      console.log(`  Age:          ${result.age}s`);
-    }
-    if (result.nextAction) {
-      console.log(`  Next Action:  ${result.nextAction}`);
-    }
-    console.log('');
-  } finally {
-    await stateManager.close();
-  }
-}
-
-/**
- * PRI-431 Step 1d: Build diagnostician-specific test-double payload.
- * Extracted from inline resolution in handleDiagnoseRun to use with shared resolver.
- */
-function buildDiagnosticianTestDouble(taskId: string): PDRuntimeAdapter {
-  return new TestDoubleRuntimeAdapter({
-    onPollRun: (_runId: string) => ({
-      runId: _runId,
-      status: 'succeeded',
-      startedAt: new Date().toISOString(),
-      endedAt: new Date().toISOString(),
-    }),
-    onFetchOutput: (_runId: string) => ({
-      runId: _runId,
-      payload: {
-        valid: true,
-        diagnosisId: `diag-cli-${Date.now()}`,
-        taskId,
-        summary: 'CLI test diagnosis — validate tool arguments before execution',
-        rootCause: 'Test root cause — missing argument validation',
-        violatedPrinciples: [],
-        evidence: [],
-        recommendations: [
-          { kind: 'principle', description: 'Always validate tool arguments before execution to prevent silent failures' },
-          { kind: 'rule', description: 'Use schema validation for external inputs' },
-        ],
-        confidence: 0.9,
-      },
-    }),
-  });
-}
-
-/**
- * pd diagnose run --task-id <taskId> [--workspace <path>] [--json]
- *
- * Executes the diagnostician runner for a task.
- */
-export async function handleDiagnoseRun(opts: DiagnoseRunOptions): Promise<void> {
-  const workspaceDir = resolveWorkspaceDir(opts.workspace);
-  // PRI-720: resolve the full-chain override once per invocation.
-  const seedPipelineMode = resolvePromptFullPipelineSeedMode(workspaceDir);
-
-  // Validate mutually exclusive flags (HG-03)
-  if (opts.openclawLocal && opts.openclawGateway) {
-    console.error('error: --openclaw-local and --openclaw-gateway are mutually exclusive');
-    process.exit(1);
-    return;
-  }
-
-  // PRI-638: this is the CLI's ONLY capability disable check. It reads the
-  // canonical authority (internalAgents.agents.diagnostician.enabled) through
-  // the same resolver the runtime factory uses, so `pd diagnose` can never
-  // mislabel a deliberate Owner kill switch as a missing runtime, a provider
-  // failure or a malformed config. No runtime adapter is constructed and no
-  // provider is contacted on this path.
-  const configLoadResult = loadPdConfig(workspaceDir);
-  const capability = resolveDiagnosticianCapability(
-    configLoadResult.ok ? configLoadResult.effective : configLoadResult.defaults,
-  );
-  if (!capability.available) {
-    const disabledResult = {
-      ok: false,
-      status: 'failed',
-      reason: capability.reason,
-      message: capability.message,
-      nextAction: capability.nextAction,
-    };
-    if (opts.json) {
-      console.log(JSON.stringify(disabledResult, null, 2));
-    } else {
-      console.error(`error: ${capability.message}`);
-      console.error(`reason: ${capability.reason}`);
-      console.error(`nextAction: ${capability.nextAction}`);
-    }
-    process.exit(1);
-    return;
-  }
-
-  // Resolve runtime kind. P1 fix (mirrors pain-retry.ts): pd diagnose run
-  // must NOT default to test-double. Without --runtime, the split pipeline
-  // would validate the test-double's stale DiagnosticianOutputV1-shaped
-  // payload against DiagRootCauseOutputV1Schema and fail with
-  // max_attempts_exceeded — silently producing failed/fake diagnostic data
-  // in a real workspace. Fall back to .pd/config.yaml first; refuse if
-  // nothing is bound. (rc-9-no-silent-fallback / ERR-002; EP-03, EP-04)
-  let runtimeKind = opts.runtime;
-  if (!runtimeKind) {
-    const resolved = resolveRuntimeFromPdConfig(workspaceDir);
-    if (!isRuntimeConfigError(resolved.result) && resolved.result.runtimeKind) {
-      ({ runtimeKind } = resolved.result);
-    }
-  }
-
-  if (!runtimeKind) {
-    const missingRuntime = {
-      ok: false,
-      reason: 'missing_runtime',
-      message:
-        'No --runtime specified and no .pd/config.yaml runtime binding found. pd diagnose run must not default to test-double to prevent fake data in real workspaces.',
-      nextAction: `Specify --runtime explicitly: pd diagnose run --task-id ${opts.taskId} --runtime pi-ai --provider <provider> --model <model> --apiKeyEnv <ENV>`,
-    };
-    if (opts.json) {
-      console.log(JSON.stringify(missingRuntime, null, 2));
-    } else {
-      console.error(`error: ${missingRuntime.message}`);
-      console.error(`Next action: ${missingRuntime.nextAction}`);
-    }
-    process.exit(1);
-    return;
-  }
-
-  const stateManager = new RuntimeStateManager({ workspaceDir });
-
-  try {
-    await stateManager.initialize();
-
-    // Build context assembler from internal stores
-    const sqliteConn = stateManager.connection;
-    const {taskStore} = stateManager;
-    const {runStore} = stateManager;
-    const historyQuery = new SqliteHistoryQuery(sqliteConn);
-    const trajectoryLocator = new SqliteTrajectoryLocator(sqliteConn);
-    const sourceTraceLocator = new SqliteSourceTraceLocator(taskStore, trajectoryLocator);
-    const contextAssembler = new SqliteContextAssembler(taskStore, historyQuery, runStore, { sourceTraceLocator });
-
-    // Select runtime adapter based on --runtime flag (CLI-02)
-    // PRI-431: migrated to shared resolveRuntimeAdapterFromConfig
-    let runtimeAdapter: PDRuntimeAdapter;
-    // Capture config info for telemetry (resolver consumes config internally)
-    let telemetryConfig: { provider?: string; model?: string; openclawMode?: string } = {};
-    try {
-      if (runtimeKind === 'test-double') {
-        runtimeAdapter = resolveRuntimeAdapterFromConfig({
-          runtimeKind,
-          workspaceDir,
-          allowTestDouble: true,
-          testDoublePayloadBuilder: () => buildDiagnosticianTestDouble(opts.taskId),
-        });
-      } else if (runtimeKind === 'openclaw-cli') {
-        const flagMode: 'local' | 'gateway' | undefined = opts.openclawLocal
-          ? 'local'
-          : opts.openclawGateway
-            ? 'gateway'
-            : undefined;
-        runtimeAdapter = resolveRuntimeAdapterFromConfig({
-          runtimeKind,
-          workspaceDir,
-          openclawMode: flagMode,
-          agentId: opts.agent ?? 'main',
-          onConfigResolved: (resolved) => {
-            if (!isRuntimeConfigError(resolved.result)) {
-              telemetryConfig.openclawMode = flagMode ?? resolved.result.openclawMode;
-            } else {
-              telemetryConfig.openclawMode = flagMode;
-            }
-          },
-        });
-        // TELE-01: runtime_adapter_selected — user explicitly chose openclaw-cli runtime
-        storeEmitter.emitTelemetry({
-          eventType: 'runtime_adapter_selected',
-          traceId: opts.taskId,
-          timestamp: new Date().toISOString(),
-          sessionId: 'pd-cli-diagnose',
-          agentId: 'openclaw-cli-adapter',
-          payload: {
-            runtimeKind: 'openclaw-cli',
-            runtimeMode: telemetryConfig.openclawMode ?? 'local',
-          },
-        });
-      } else if (runtimeKind === 'pi-ai') {
-        runtimeAdapter = resolveRuntimeAdapterFromConfig({
-          runtimeKind,
-          workspaceDir,
-          configOptional: true,
-          validateApiKeyEnv: true,
-          piAiOverrides: {
-            provider: opts.provider,
-            model: opts.model,
-            apiKeyEnv: opts.apiKeyEnv,
-            baseUrl: opts.baseUrl,
-            maxRetries: opts.maxRetries,
-          },
-          timeoutMs: opts.timeoutMs,
-          onConfigResolved: (resolved) => {
-            for (const w of resolved.legacyWarnings) console.warn(`[pd diagnose] ${w}`);
-            if (isRuntimeConfigError(resolved.result)) {
-              console.warn(`[pd diagnose] .pd/config.yaml resolution failed: ${resolved.result.message}. Using CLI flags if provided.`);
-            } else {
-              telemetryConfig.provider = resolved.result.provider;
-              telemetryConfig.model = resolved.result.model;
-            }
-          },
-        });
-        // TELE: runtime_adapter_selected telemetry
-        const telemetryProvider = opts.provider ?? telemetryConfig.provider;
-        const telemetryModel = opts.model ?? telemetryConfig.model;
-        storeEmitter.emitTelemetry({
-          eventType: 'runtime_adapter_selected',
-          traceId: opts.taskId,
-          timestamp: new Date().toISOString(),
-          sessionId: 'pd-cli-diagnose',
-          agentId: 'pi-ai-adapter',
-          payload: { runtimeKind: 'pi-ai', provider: telemetryProvider, model: telemetryModel, baseUrlPresent: !!opts.baseUrl },
-        });
-      } else {
-        const unsupportedResult = {
-          ok: false,
-          reason: `unsupported_runtime_kind: ${runtimeKind}`,
-          nextAction: 'Use one of: openclaw-cli, test-double, pi-ai',
-        };
-        if (opts.json) {
-          console.log(JSON.stringify(unsupportedResult, null, 2));
-        } else {
-          console.error(`error: unknown runtime kind '${runtimeKind}' (supported: openclaw-cli, test-double, pi-ai)`);
-          console.error(`Next action: ${unsupportedResult.nextAction}`);
-        }
-        process.exit(1);
-        return;
-      }
-    } catch (err) {
-      if (err instanceof ConfigResolutionError) {
-        // Preserve original error format for openclaw-cli missing mode (backward compat)
-        if (err.kind === 'missing-fields' && err.missing?.includes('openclawMode')) {
-          if (opts.json) {
-            console.log(JSON.stringify({
-              ok: false,
-              reason: 'missing_openclaw_mode',
-              message: 'runtimeKind is openclaw-cli but no mode resolved',
-              nextAction: 'Provide --openclaw-local or --openclaw-gateway, or set openclawMode in .pd/config.yaml',
-            }));
-          } else {
-            console.error('error: runtimeKind is openclaw-cli but no mode resolved');
-            console.error('nextAction: Provide --openclaw-local or --openclaw-gateway, or set openclawMode in .pd/config.yaml');
-          }
-          process.exit(1);
-          return;
-        }
-        // Default error formatting for other config resolution errors
-        if (opts.json) {
-          console.log(JSON.stringify({ ok: false, reason: err.kind, message: err.message, missing: err.missing, nextAction: err.nextAction ?? 'Check .pd/config.yaml and retry' }));
-        } else {
-          console.error(`error: ${err.message}`);
-          if (err.nextAction) {
-            console.error(`nextAction: ${err.nextAction}`);
-          }
-        }
-        process.exit(1);
-        return;
-      }
-      throw err;
-    }
-
-    const eventEmitter = new StoreEventEmitter();
-    const committer = new SqliteDiagnosticianCommitter(sqliteConn);
-
-    // PRI-336: Read outputLanguage from workspace config
-    const outputLangResult = readOutputLanguageFromWorkspace(workspaceDir);
-    const outputLanguage: OutputLanguage | undefined = outputLangResult.outputLanguage;
-
-    // PRI-638: implementation selection is gone. The split pipeline is the only
-    // Diagnostician implementation in the tree, so it always runs with its
-    // documented 3-stage budget; `diagnostician_split_pipeline` no longer
-    // selects a runner nor disables capability.
-    const pipelineTimeoutMs = SPLIT_PIPELINE_TOTAL_TIMEOUT_MS;
-    // BUG-1 (PRI-442): extract effectiveConfig so ADR-0019 LLM rate-limit
-    // degradation (isDegradationEnabled in base-peer-runner) can read the
-    // diagnostician_llm_degradation feature flag. Without this, the runners
-    // receive no effectiveConfig and degradation silently never fires.
-    // CR-1 (CodeRabbit P2, rc-9): warn when config load failed so the
-    // fallback to defaults is observable — no silent degradation.
-    if (!configLoadResult.ok) {
-      const errSummary = configLoadResult.errors
-        .map((e) => `${e.path}: ${e.reason}`)
-        .join('; ');
-      console.warn(
-        `[pd diagnose] .pd/config.yaml at ${configLoadResult.configPath} is malformed — falling back to default feature flags. Errors: ${errSummary}. Next action: ${configLoadResult.errors[0]?.nextAction ?? 'fix config errors and retry'}`,
-      );
-    }
-    const effectiveConfig = configLoadResult.ok ? configLoadResult.effective : configLoadResult.defaults;
-
-    const resolvedKind = typeof runtimeAdapter.kind === 'function' ? runtimeAdapter.kind() : runtimeKind;
-    const perStageTimeoutMs = pipelineTimeoutMs / 3;
-    const rootCauseRunner = new DiagRootCauseRunner(
-      { stateManager, runtimeAdapter, eventEmitter, artifactStore: stateManager.piArtifactStore, validator: new DefaultDiagRootCauseValidator(), contextAssembler },
-      { owner: 'pd-cli-diagnose', runtimeKind: resolvedKind, outputLanguage, timeoutMs: perStageTimeoutMs, effectiveConfig },
-    );
-    const distillerRunner = new DiagDistillerRunner(
-      { stateManager, runtimeAdapter, eventEmitter, artifactStore: stateManager.piArtifactStore, validator: new DefaultDiagDistillerValidator() },
-      { owner: 'pd-cli-diagnose', runtimeKind: resolvedKind, outputLanguage, timeoutMs: perStageTimeoutMs, effectiveConfig },
-    );
-    const routerRunner = new DiagRouterRunner(
-      { stateManager, runtimeAdapter, eventEmitter, artifactStore: stateManager.piArtifactStore, committer },
-      { owner: 'pd-cli-diagnose', runtimeKind: resolvedKind, outputLanguage, timeoutMs: perStageTimeoutMs, effectiveConfig },
-    );
-
-    const runner: DiagnosticianRunnerLike = new SplitDiagnosticianRunner({
-      rootCauseRunner,
-      distillerRunner,
-      routerRunner,
-      stateManager,
-      committer,
-      perStageTimeoutMs,
-    });
-
-    if (!opts.json) {
-      console.log(`\nRunning diagnostician for task: ${opts.taskId}`);
-      console.log(`Workspace: ${workspaceDir}\n`);
-    }
-
-    // ERR-067 fix: loop on `retried` status — the SplitDiagnosticianRunner
-    // handles per-stage retry internally, but the CLI must also loop in case
-    // the top-level result is `retried` (e.g., from a non-split pipeline).
-    const retryPolicy = stateManager.getRetryPolicy();
-    let result = await diagnoseRun({
-      taskId: opts.taskId,
-      stateManager,
-      runner,
-    });
-
-    let retryLoopCount = 0;
-    const maxRetryLoops = 10; // Safety limit
-    while (result.status === 'retried' && retryLoopCount < maxRetryLoops) {
-      retryLoopCount++;
-      const task = await stateManager.getTask(opts.taskId);
-      const backoffMs = task ? retryPolicy.calculateBackoff(task.attemptCount) : 30_000;
-
-      if (!opts.json) {
-        console.log(`  Retry ${retryLoopCount}: waiting ${Math.round(backoffMs / 1000)}s before re-running...`);
-      }
-
-      await new Promise((resolve) => setTimeout(resolve, backoffMs));
-
-      result = await diagnoseRun({
-        taskId: opts.taskId,
-        stateManager,
-        runner,
-      });
-    }
-
-    // P0-1 fix: convert non-terminal `retried` to `failed` with reason
-    if (result.status === 'retried') {
-      result = {
-        ...result,
-        status: 'failed',
-        failureReason: `Max retry loops (${maxRetryLoops}) exceeded without reaching terminal state. ${result.failureReason ?? ''}`,
-      };
-    }
-
-    if (result.status !== 'succeeded') {
-      if (opts.json) {
-        console.log(JSON.stringify(result, null, 2));
-      } else {
-        console.log(`\nResult:`);
-        console.log(`  Status:         ${result.status}`);
-        console.log(`  Task ID:        ${result.taskId}`);
-        if (result.errorCategory) {
-          console.log(`  Error Category: ${result.errorCategory}`);
-        }
-        if (result.failureReason) {
-          console.log(`  Failure Reason: ${result.failureReason}`);
-        }
-        console.log(`  Attempt Count:  ${result.attemptCount}`);
-        console.log('');
-      }
-      process.exit(1);
-      return;
-    }
-
-    const candidates = await stateManager.getCandidatesByTaskId(opts.taskId);
-    const intakeResults: { candidateId: string; ledgerEntryId?: string; status: string; error?: string; nextAction?: string }[] = [];
-    let intakeFailed = false;
-
-    if (opts.intake === false) {
-      for (const candidate of candidates) {
-        intakeResults.push({
-          candidateId: candidate.candidateId,
-          status: 'skipped',
-        });
-      }
-    } else {
-      const ledgerAdapter = new PrincipleTreeLedgerAdapter({ stateDir: path.join(workspaceDir, '.state') });
-      const intakeService = new CandidateIntakeService({ stateManager, ledgerAdapter });
-
-      for (const candidate of candidates) {
-        // PRI-503: admission gate check — refuse non-admitted candidates before
-        // intake, mirroring candidate.ts (intake/repair/backfill). Without this,
-        // `pd diagnose --intake` could bypass the gate and intake unreviewed
-        // candidates. cli-5-failure-no-mutation: refusal path does not call
-        // intake(), does not update candidate status, does not write ledger.
-        // cli-6-output-next-action: refusal result carries admissionBlock.nextAction.
-        const admissionBlock = checkAdmissionGate(candidate);
-        if (admissionBlock) {
-          intakeFailed = true;
-          intakeResults.push({
-            candidateId: candidate.candidateId,
-            status: 'intake_failed',
-            error: `Admission gate refused: ${admissionBlock.reason}`,
-            nextAction: admissionBlock.nextAction,
-          });
-          continue;
-        }
-        try {
-          const entry = await intakeService.intake(candidate.candidateId);
-          if (candidate.status !== 'consumed') {
-            await stateManager.updateCandidateStatus(candidate.candidateId, { status: 'consumed' });
-          }
-          intakeResults.push({
-            candidateId: candidate.candidateId,
-            ledgerEntryId: entry.id,
-            status: 'consumed',
-          });
-        } catch (intakeErr: unknown) {
-          intakeFailed = true;
-          const intakeErrorMessage = intakeErr instanceof Error ? intakeErr.message : String(intakeErr);
-          intakeResults.push({
-            candidateId: candidate.candidateId,
-            status: 'intake_failed',
-            error: intakeErrorMessage,
-            nextAction: `pd candidate intake --candidate-id ${candidate.candidateId} --workspace "${workspaceDir}"`,
-          });
-        }
-      }
-    }
-
-    // Defect-004 fix: seed dreamer task for each consumed candidate whose
-    // recommendation_kind routes to an MVP-enabled internalization channel.
-    // Without this, the chain stops at 'consumed' and never reaches internalization
-    // (11 of 18 consumed candidates hit this in production). Mirrors the pattern
-    // in PainSignalBridge.onDiagnosisComplete (pain-signal-bridge.ts L310-338).
-    //
-    // DEFECT-005 (PRI-514): the dreamer seed loop is gated ONLY by `opts.intake`,
-    // NOT by a global `intakeFailed` flag. A single defer candidate whose intake
-    // is refused by the admission gate must NOT poison dreamer seed for sibling
-    // candidates that were successfully consumed (EP-03 / ERR-089 sibling-branch
-    // defect). Per-candidate eligibility is checked inside the loop via the
-    // intake result's status. `intakeFailed` is retained solely for the
-    // nextAction/exit-code signaling below (partial failure still exits non-zero).
-    if (opts.intake !== false) {
-      for (const candidate of candidates) {
-        const kind = candidate.recommendationKind;
-        if (kind === 'defer' || kind === 'implementation') continue;
-        const intakeResult = intakeResults.find((r) => r.candidateId === candidate.candidateId);
-        if (!intakeResult || intakeResult.status !== 'consumed') continue;
-        try {
-          const route = CANDIDATE_KIND_TO_ROUTE[kind ?? ''];
-          if (!route) continue;
-          const channel = ROUTE_CHANNEL_MAP[route];
-          const ready = !!channel && MVP_ENABLED_CHANNELS.has(channel);
-          // BUG-2 (PRI-442): resolve sourcePainId from the diagnostician task
-          // chain (rc-6 lineage consistency). The candidate itself doesn't carry
-          // sourcePainId; it must be resolved from the diagnostician task's
-          // diagnosticJson. ERR-004: never invent lineage.
-          const sourcePainId = await resolveSourcePainIdFromDiagnostician(stateManager, candidate);
-          const seed = buildDreamerSeedFromCandidate(candidate, { route, ready, sourcePainId: sourcePainId ?? undefined, pipelineMode: seedPipelineMode });
-          // eslint-disable-next-line no-restricted-syntax -- 'in' required for discriminated union narrowing (BridgeTaskSeed | BridgeDecision)
-          if ('decision' in seed) continue; // not_internalizable or invalid — skip
-          // PRI-720 C6: candidate-level dedup (demotion may change the
-          // derived channel suffix vs a pre-existing chain).
-          const existingTask = await findExistingDreamerTask((id) => stateManager.getTask(id), candidate.candidateId);
-          if (!existingTask) {
-            await stateManager.createTask({
-              taskId: seed.taskId,
-              taskKind: seed.taskKind,
-              inputRef: '',
-              status: seed.status,
-              attemptCount: seed.attemptCount,
-              maxAttempts: seed.maxAttempts,
-              diagnosticJson: seed.diagnosticJson,
-            });
-            intakeResults.push({
-              candidateId: candidate.candidateId,
-              status: 'dreamer_seeded',
-              nextAction: `pd task show ${seed.taskId} | pd dreamer run --task-id ${seed.taskId}`,
-            });
-          }
-        } catch (seedErr) {
-          intakeResults.push({
-            candidateId: candidate.candidateId,
-            status: 'dreamer_seed_failed',
-            error: seedErr instanceof Error ? seedErr.message : String(seedErr),
-            nextAction: `pd candidate internalize --candidate-id ${candidate.candidateId}`,
-          });
-        }
-      }
-    }
-
-    if (opts.json) {
-      const candidateIds = candidates.map((c) => c.candidateId);
-      // CodeRabbit review fix: reflect dreamer seed status in nextAction so
-      // the owner knows whether internalization has already started or still
-      // needs manual candidate internalize. Previously the message always
-      // said "NOT started automatically" even when dreamer tasks were seeded.
-      const dreamerSeededCount = intakeResults.filter((r) => r.status === 'dreamer_seeded').length;
-      const dreamerSeedFailedCount = intakeResults.filter((r) => r.status === 'dreamer_seed_failed').length;
-      const internalizeNextAction = dreamerSeededCount > 0
-        ? `Dreamer tasks seeded automatically for ${dreamerSeededCount} candidate(s). To continue, run the dreamer tasks shown in intake.candidates[].${dreamerSeedFailedCount > 0 ? ` (${dreamerSeedFailedCount} candidate(s) failed seeding — see intake.candidates[] for retry guidance.)` : ''}`
-        : candidateIds.length > 0
-          ? `Candidates generated but internalization has NOT started automatically. To begin internalization, run:\n  ${candidateIds.map((id) => `pd candidate internalize --candidate-id ${id} --workspace "${workspaceDir}"`).join('\n  ')}`
-          : 'No candidates were generated from this diagnosis.';
-      const jsonOutput = {
-        ...result,
-        intake: {
-          enabled: opts.intake !== false,
-          candidates: intakeResults,
-        },
-        nextAction: internalizeNextAction,
-      };
-      console.log(JSON.stringify(jsonOutput, null, 2));
-      if (intakeFailed) {
-        process.exit(1);
-      }
-      return;
-    }
-
-    console.log(`\nResult:`);
-    console.log(`  Status:         ${result.status}`);
-    console.log(`  Task ID:        ${result.taskId}`);
-    if (result.contextHash) {
-      console.log(`  Context Hash:  ${result.contextHash.substring(0, 16)}...`);
-    }
-    if (result.output) {
-      console.log(`  Diagnosis ID:   ${result.output.diagnosisId}`);
-      console.log(`  Summary:        ${result.output.summary}`);
-      if (result.output.recommendations) {
-        const principleCount = result.output.recommendations.filter((r: { kind: string }) => r.kind === 'principle').length;
-        if (principleCount > 0) {
-          console.log(`  Principles:     ${principleCount} candidate(s) generated`);
-        }
-      }
-    }
-    console.log(`  Attempt Count:  ${result.attemptCount}`);
-
-    if (intakeResults.length > 0) {
-      console.log(`\n  Candidate Intake:`);
-      for (const ir of intakeResults) {
-        if (ir.status === 'consumed') {
-          console.log(`    ${ir.candidateId}: consumed (ledger: ${ir.ledgerEntryId})`);
-        } else if (ir.status === 'skipped') {
-          console.log(`    ${ir.candidateId}: skipped (--no-intake)`);
-        } else if (ir.status === 'intake_failed') {
-          console.log(`    ${ir.candidateId}: INTAKE FAILED — ${ir.error}`);
-          console.log(`      Next action: pd candidate intake --candidate-id ${ir.candidateId} --workspace "${workspaceDir}"`);
-        } else if (ir.status === 'dreamer_seeded') {
-          // CodeRabbit review fix: surface dreamer seed success in TTY output
-          console.log(`    ${ir.candidateId}: dreamer seeded — ${ir.nextAction}`);
-        } else if (ir.status === 'dreamer_seed_failed') {
-          // CodeRabbit review fix: surface dreamer seed failure in TTY output
-          console.log(`    ${ir.candidateId}: DREAMER SEED FAILED — ${ir.error}`);
-          console.log(`      Next action: ${ir.nextAction}`);
-        }
-      }
-    }
-
-    if (opts.intake === false && candidates.length > 0) {
-      console.log(`\n  Note: --no-intake was set. Candidates remain at 'pending'.`);
-      console.log(`  To intake manually:`);
-      for (const c of candidates) {
-        console.log(`    pd candidate intake --candidate-id ${c.candidateId} --workspace "${workspaceDir}"`);
-      }
-    }
-
-    if (candidates.length > 0) {
-      console.log(`\n  Next Action:`);
-      console.log(`  Candidates generated but internalization has NOT started automatically.`);
-      console.log(`  To begin internalization:`);
-      for (const c of candidates) {
-        console.log(`    pd candidate internalize --candidate-id ${c.candidateId} --workspace "${workspaceDir}"`);
-      }
-    }
-
-    console.log('');
-
-    if (intakeFailed) {
-      process.exit(1);
-    }
-  } catch (error: unknown) {
-    const message = error instanceof Error ? error.message : String(error);
-    let errorCategory = 'execution_failed';
-    if (error instanceof PDRuntimeError) {
-      errorCategory = error.category;
-    }
-    if (opts.json) {
-      console.log(JSON.stringify({
-        status: 'failed',
-        errorCategory,
-        message,
-        runtimeKind,
-      }, null, 2));
-    } else {
-      console.error(`error: ${message} (${errorCategory})`);
-    }
-    process.exit(1);
-  } finally {
-    await stateManager.close();
-  }
-}