@awak-app/simy-cli 0.5.72 → 0.5.73

package/README.md

@@ -2,6 +2,67 @@
 
 Local execution agent for SIMY Local Tasks and Agentic Loop runs.
 
+`simy --version` prints the installed CLI version without starting an agent or
+checking for updates. `simy doctor` performs read-only local connection and
+provider login checks and prints a JSON diagnostic report: installed/running
+versions, stale discovery records, session expiry, provider readiness, and recent
+structured SQM failures. It does not run a model, renew login, or restart tasks.
+Provider login is separate from model access; unverified model access remains
+`unknown`. Diagnostic journals retain stage/code/HTTP status only, use private
+files scoped to the Web/account/organization, rotate each writer's four 64 KiB
+segments, and keep lifetime failure counts after event rotation.
+
+The organization menu's **Allow Codex / Claude fallback** setting defaults to
+off and grants only read-only SQM analysis permission to use the other signed-in
+local provider. It is saved per Web/account/organization. Switching requires an
+explicit native model rejection or unavailable executable, a server that supports
+actual-provider provenance, and fresh actor/settings/lease checks. Authentication,
+quota, malformed output, arbitrary failures and coding tasks do not switch engines.
+Both providers share one analysis deadline. A deterministic provider failure pauses
+new claims and retains failed history until environment/settings change or **Retry
+analysis** is selected.
+
+Agentic Loop preflight accepts the task's existing requirement, acceptance
+criteria, required checks and charter context alongside the repository/provider.
+Its `startup` summary lists all blockers, unverified conditions and recovery
+actions together. Environment-only calls remain supported. Task checks reuse the
+execution charter policy; model entitlement, external test resources and assurance
+selection evidence are not guessed from a successful login or recommendation.
+
+### Explicit progress continuation
+
+Existing tasks use **Bounded** continuation. Select **Progress** before starting
+an Agentic Loop or Local Task, or when enrolling an App session, to treat the
+ordinary token/attempt/time limits as warning thresholds. New independently
+verified acceptance criteria permit further work within finite hard ceilings;
+repeated work without new proof pauses the same task. Approval requests, explicit
+stops and authorization failures still pause execution. Provider text claiming
+success does not count as verification.
+
+Local Tasks retain valid output after a soft token warning and stop at their
+explicit hard ceiling (at most 10 million reported provider tokens). The Agentic
+controller limits a whole progress run to its approved allocation, at most
+15 million tokens, and persists the allocation with canonical receipts. Missing
+reliable usage is charged conservatively against the reserved allocation. The
+Web controller supports this mode for the connected local Codex/Claude engines;
+unsupported cloud execution is refused before dispatch.
+
+In the App, pause a managed Delivery task before changing its continuation
+policy or extending its bounds. Extension keeps usage, task identity, acceptance
+and approvals, and requires a separate resume action. Delivery hard ceilings are
+24 hours and 50 continuations; host-observed Codex usage can also impose a token
+ceiling. A Claude live-hook session without reliable token telemetry explicitly
+uses time/continuation bounds instead of inventing usage.
+
+For interrupted Agentic work, `GET /v1/agentic-loop/:id/recovery` provides a
+read-only objective and checkout fingerprint. Reconcile an uncertain side effect
+with `POST` to that endpoint (`action: reconcile`, authoritative
+`operation_status`, `observation`, `evidence_refs`) before resuming. A paused
+progress run accepts explicit increases through
+`POST /v1/agentic-loop/:id/continuation-ceilings`; it never resets already used
+tokens or automatically resumes. Both routes require the original authorized
+account, organization and device throughout the request.
+
 ## Apple Silicon Mac app
 
 The macOS installer includes a native status window and menu bar app. Open
@@ -198,6 +259,29 @@ that task.
 
 ## Local Tasks
 
+Purpose runs may opt into independent local evidence by freezing user-approved
+`verification_recipes` in the server claim. `capabilities.purpose_verification`
+advertises the device's Ed25519 public key; the server pins that key for the
+attempt. Older clients without this capability cannot supply this evidence.
+`POST /v1/local-tasks/:id/verify-purpose` with `{}` checks the original plan only
+after the task succeeds. GET only reads the cached `task.host_verification`.
+File recipes independently check bounded regular files and optional JSON format;
+command recipes execute approved structured argv for Node tests or package
+`test`, `build`, `check`, and `lint` scripts. Commands require write permission,
+share a five-minute deadline, and retain only output hashes and exit status.
+Optional `protected_paths` bind check definitions to their pre-execution hashes.
+
+A receipt binds the actor, device, attempt, objective, recipe plan, local task,
+and before/after repository identity. Changed repositories, stopped or uncertain
+tasks, unknown prior check effects, and changed keys fail closed. Fresh POSTs are
+idempotent; an expired confirmed receipt may be refreshed explicitly on the same
+task, with at most three check rounds and no new Provider invocation. Checks
+never upload source files or raw command output. Exit zero proves the specified
+command's result; it does not prove product acceptance or that a package script
+has no external effects. This host evidence boundary does not isolate a malicious
+process running as the same OS user.
+
+
 CLI API contract v9 promotes non-coding work to a first-class **Local Task**
 runtime. A Local Task is separate from Agentic Loop: it runs one bounded Codex
 or Claude Code session without Git branches, pull requests, audits, or automatic

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@awak-app/simy-cli",
-  "version": "0.5.72",
+  "version": "0.5.73",
   "description": "Local SIMY executor for non-coding Local Tasks and Agentic Loop runs.",
   "type": "module",
   "bin": {
@@ -42,7 +42,7 @@
     "check:workflow-observation-sender": "node --check ./src/workflow-observation-sender.js",
     "check:sqm": "node --check ./src/sqm/canonical.js && node --check ./src/sqm/validation.js && node --check ./src/sqm/reliability.js && node --check ./src/sqm/repository.js && node --check ./src/sqm/checkers.js && node --check ./src/sqm/bundle-store.js && node --check ./src/sqm/proof.js && node --check ./src/sqm/evidence-client.js && node --check ./src/sqm/session.js && node --check ./src/sqm/index.js && node --check ./src/sqm/command.js && node --check ./src/sqm/agent-skill.js && node --check ./src/sqm/workflow-onboarding.js && node --check ./src/sqm/workflow-update.js && node --check ./src/sqm/history-store.js && node --check ./src/sqm/review-worker.js && node --check ./scripts/sqm-complete-loop-e2e.js && node --check ./scripts/sqm-local-review-e2e.js",
     "check:real-provider-syntax": "node --check ./scripts/real-codex-probe.js && node --check ./scripts/real-claude-probe.js && node --check ./scripts/real-provider-console-fixture.js && node --check ./scripts/real-provider-console-e2e.js && node --check ./scripts/claude-parity-runtime-e2e.js && node --check ./scripts/claude-plugin-runtime-e2e.js",
-    "check:syntax": "node --check ./src/index.js && node --check ./src/install-update.js && node --check ./src/agent.js && node --check ./src/connection-discovery.js && node --check ./src/codex-plugins/simy/hooks/connection-reminder.mjs && node --check ./src/codex-plugins/simy/hooks/dev-work.mjs && node --check ./src/dev-work-events.js && node --check ./src/browser.js && node --check ./src/web-api.js && node --check ./src/durable-local-task-worker.js && node --check ./src/durable-local-task-steps-contract.js && node --check ./src/bounded-local-task-subtasks-contract.js && node --check ./src/bounded-local-task-subtask-pool.js && node --check ./src/local-task-artifact-contract.js && node --check ./src/execution-capability-contract.js && node --check ./src/shutdown.js && node --check ./src/run-registry.js && node --check ./src/direct-executor.js && node --check ./src/execution-guardrail.js && node --check ./src/backend-executable.js && node --check ./src/provider-stream.js && node --check ./src/console/index.js && node --check ./src/console/app.js && node --check ./src/console/commands.js && node --check ./src/local-attachments.js && node --check ./src/repository-inventory.js && node --check ./src/session-store.js && node --check ./src/web-origin.js && node --check ./src/workspace-context.js && node --check ./src/orchestrator/index.js && node --check ./src/orchestrator/shared.js && node --check ./src/orchestrator/recovery.js && node --check ./src/orchestrator/risk.js && node --check ./src/orchestrator/contract.js && node --check ./src/orchestrator/completion-contract.js && node --check ./src/orchestrator/pr-preparation.js && node --check ./src/orchestrator/minimum-completion.js && node --check ./src/orchestrator/delivery/scripts/check-minimum-completion.mjs && node --check ./src/orchestrator/instruction.js && node --check ./src/orchestrator/problem-solving.js && node --check ./src/orchestrator/budget.js && node --check ./src/orchestrator/execution-io.js && node --check ./src/orchestrator/presentation.js && node --check ./src/orchestrator/result.js && node --check ./src/orchestrator/retry.js && node --check ./src/orchestrator/evidence.js && node --check ./src/orchestrator/independent-audit.js && node --check ./src/orchestrator/audit.js && node --check ./src/orchestrator/loop.js && node --check ./src/runner.js && node --check ./src/coding-run-store.js && node --check ./scripts/fixtures/completion-contract-fixture.js && node --check ./scripts/fixtures/fake-coding-backend.js && node --check ./scripts/e2e-terminal-evidence.js && node --check ./scripts/audit-gate-e2e.js && node --check ./scripts/completion-contract-e2e.js && node --check ./scripts/retry-circuit-e2e.js && node --check ./scripts/follow-up-charter-e2e.js && node --check ./scripts/task-routing-e2e.js && node --check ./scripts/executor-version-preflight-e2e.js && node --check ./scripts/blocked-recovery-e2e.js && node --check ./scripts/process-localization-e2e.js && node --check ./scripts/hypothesis-retries-e2e.js && node --check ./scripts/budget-enforcement-e2e.js && node --check ./scripts/local-task-default-workspace-e2e.js && node --check ./scripts/local-task-idempotency-e2e.js && node --check ./scripts/console-e2e-fixture.js && node --check ./scripts/console-e2e.js && node --check ./scripts/console-recovery-e2e-fixture.js && node --check ./scripts/console-recovery-e2e.js && node --check ./scripts/real-provider-console-fixture.js && node --check ./scripts/real-provider-console-e2e.js && node --check ./scripts/fullstack-cli-smoke.js && node --check ./scripts/check-package-contents.js && node --check ./src/codex-plugins/simy/hooks/delivery-loop.mjs && node --check ./src/delivery-hooks/command.js && node --check ./src/delivery-hooks/desktop.js && node --check ./src/delivery-hooks/supervisor.js && node --check ./src/session-supervision/automatic-continuation.js",
+    "check:syntax": "node --check ./src/purpose-local-proof.js && node --check ./src/device-signing-identity.js && node --check ./src/index.js && node --check ./src/install-update.js && node --check ./src/agent.js && node --check ./src/connection-discovery.js && node --check ./src/codex-plugins/simy/hooks/connection-reminder.mjs && node --check ./src/codex-plugins/simy/hooks/dev-work.mjs && node --check ./src/dev-work-events.js && node --check ./src/browser.js && node --check ./src/web-api.js && node --check ./src/durable-local-task-worker.js && node --check ./src/durable-local-task-steps-contract.js && node --check ./src/bounded-local-task-subtasks-contract.js && node --check ./src/bounded-local-task-subtask-pool.js && node --check ./src/local-task-artifact-contract.js && node --check ./src/execution-capability-contract.js && node --check ./src/shutdown.js && node --check ./src/run-registry.js && node --check ./src/direct-executor.js && node --check ./src/execution-guardrail.js && node --check ./src/backend-executable.js && node --check ./src/provider-stream.js && node --check ./src/console/index.js && node --check ./src/console/app.js && node --check ./src/console/commands.js && node --check ./src/local-attachments.js && node --check ./src/repository-inventory.js && node --check ./src/session-store.js && node --check ./src/web-origin.js && node --check ./src/workspace-context.js && node --check ./src/orchestrator/index.js && node --check ./src/orchestrator/shared.js && node --check ./src/orchestrator/recovery.js && node --check ./src/orchestrator/risk.js && node --check ./src/orchestrator/contract.js && node --check ./src/orchestrator/completion-contract.js && node --check ./src/orchestrator/pr-preparation.js && node --check ./src/orchestrator/minimum-completion.js && node --check ./src/orchestrator/delivery/scripts/check-minimum-completion.mjs && node --check ./src/orchestrator/instruction.js && node --check ./src/orchestrator/problem-solving.js && node --check ./src/orchestrator/budget.js && node --check ./src/orchestrator/execution-io.js && node --check ./src/orchestrator/presentation.js && node --check ./src/orchestrator/result.js && node --check ./src/orchestrator/retry.js && node --check ./src/orchestrator/evidence.js && node --check ./src/orchestrator/independent-audit.js && node --check ./src/orchestrator/audit.js && node --check ./src/orchestrator/loop.js && node --check ./src/runner.js && node --check ./src/coding-run-store.js && node --check ./scripts/fixtures/completion-contract-fixture.js && node --check ./scripts/fixtures/fake-coding-backend.js && node --check ./scripts/e2e-terminal-evidence.js && node --check ./scripts/audit-gate-e2e.js && node --check ./scripts/completion-contract-e2e.js && node --check ./scripts/retry-circuit-e2e.js && node --check ./scripts/follow-up-charter-e2e.js && node --check ./scripts/task-routing-e2e.js && node --check ./scripts/executor-version-preflight-e2e.js && node --check ./scripts/blocked-recovery-e2e.js && node --check ./scripts/process-localization-e2e.js && node --check ./scripts/hypothesis-retries-e2e.js && node --check ./scripts/budget-enforcement-e2e.js && node --check ./scripts/local-task-default-workspace-e2e.js && node --check ./scripts/local-task-idempotency-e2e.js && node --check ./scripts/console-e2e-fixture.js && node --check ./scripts/console-e2e.js && node --check ./scripts/console-recovery-e2e-fixture.js && node --check ./scripts/console-recovery-e2e.js && node --check ./scripts/real-provider-console-fixture.js && node --check ./scripts/real-provider-console-e2e.js && node --check ./scripts/fullstack-cli-smoke.js && node --check ./scripts/check-package-contents.js && node --check ./src/codex-plugins/simy/hooks/delivery-loop.mjs && node --check ./src/delivery-hooks/command.js && node --check ./src/delivery-hooks/desktop.js && node --check ./src/delivery-hooks/supervisor.js && node --check ./src/session-supervision/automatic-continuation.js",
     "check:package": "node ./scripts/check-package-contents.js",
     "build:macos-installer:dev": "bash packaging/build-macos-arm64.sh --unsigned-dev",
     "build:macos-installer": "bash packaging/build-macos-arm64.sh --production",

package/src/agent.js

@@ -1,8 +1,10 @@
+import { normalizePurposeVerification, assertPurposeActor, preparePurposeVerification, verifyPurposeLocalTask, purposeSigningKeyPath, purposeVerificationCapability, purposeWorkspaceBusy } from "./purpose-local-proof.js";
+import { startupRequirementChecks, summarizeStartupChecks } from "./startup-preflight.js";
 import { BROWSER_HANDOFF_TRANSPORT, browserHandoffPage, decryptBrowserHandoff } from "./browser-handoff.js";
 import { createAuthorizationWindow } from "./browser-authorization.js";
 import { createSqmHistoryStore, projectSqmHistory, sqmHistoryScope } from "./sqm/history-store.js";
 import { createDesktopAccessPolicy, desktopAccessReady, requireDesktopAccess } from './desktop-access.js';
-import { readSqmEnabled, writeSqmEnabled } from "./desktop-sqm-settings.js";
+import { readSqmEnabled, writeSqmEnabled, readSqmSettings, writeSqmEngineFallback } from "./desktop-sqm-settings.js";
 import { createServer } from "node:http";
 import { createHash, randomBytes } from "node:crypto";
 import { createReadStream } from "node:fs";
@@ -14,10 +16,13 @@ import {
   continueLocalCodingRunAfterRepositoryApproval,
   createRun,
   increaseLocalCodingRunProviderTokenBudget,
+  extendLocalCodingRunCeilings,
   LocalRunRegistry,
   markCodingRunMonitorFailure,
   isLocalRepositoryApprovalPending,
   pauseLocalCodingRun,
+  prepareLocalCodingRunRecovery,
+  reconcileLocalCodingRun,
   queueLocalCodingGuidance,
   recheckLocalCodingRun,
   publishRestoredLocalCodingRun,
@@ -30,6 +35,7 @@ import {
   toCompatibleCodingLoopState,
   toLedgerSnapshot,
 } from "./runner.js";
+import { assertRecoveryActor } from "./coding-run-recovery.js";
 import { codingRunsRoot, loadCodingRunSnapshots, saveCodingRunSnapshot } from "./coding-run-store.js";
 import {
   defaultSessionRoot,
@@ -196,7 +202,15 @@ export async function startAgent({
   // as its device (see /v1/auth/complete), so keep it after it is cleared.
   let reclaimableSession = null;
   if (session && sessionRequiresWebAuthorization(session, apiOrigin)) session = null;
-  let sqmEnabled = await readSqmEnabled(sessionRoot, apiOrigin, session);
+  const purposeKeyPath = dependencies.purposeSigningKeyPath || purposeSigningKeyPath(sessionRoot);
+  let purposeSigner = null;
+  const refreshPurposeSigner = async () => {
+    try { purposeSigner = await purposeVerificationCapability(purposeKeyPath, { create: true }); }
+    catch { purposeSigner = null; } // Invalid/private-key state fails closed without blocking ordinary tasks.
+  };
+  await refreshPurposeSigner();
+  let sqmEngineSettings = await readSqmSettings(sessionRoot, apiOrigin, session);
+  let sqmEnabled = sqmEngineSettings.enabled;
   let organizationItems = [];
   let organizationState = "unknown";
   let organizationRefresh = null;
@@ -251,7 +265,8 @@ export async function startAgent({
     const features = contractSourceCapabilities.features;
     availableCapabilities = withCliContract({
       ...contractSourceCapabilities,
-      ...(features ? { features: { ...features, ...(Object.hasOwn(features, "sqm_local_review")
+      ...(purposeSigner ? { purpose_verification: purposeSigner } : {}),
+      ...(features ? { features: { ...features, purpose_host_verification: Boolean(purposeSigner) && isSessionValid(session, Date.now(), apiOrigin), ...(Object.hasOwn(features, "sqm_local_review")
         ? { sqm_local_review: features.sqm_local_review === true && sqmEnabled && !pendingOrganization && desktopWorkAllowed() } : {}) } } : {}),
       ...(providerStatuses ? { provider_statuses: providerStatuses } : {}),
     }, installMode, {
@@ -576,6 +591,9 @@ export async function startAgent({
         const localRun = registry.get(snapshot.id);
         localRun.session = session;
         localRun.apiOrigin = apiOrigin;
+        // A newer ledger is not permission to replace an active local operation.
+        // Keep the owner, child, checkpoint and approval state until it settles.
+        if (localRun.operation || localRun.child) continue;
         const remoteTime = Date.parse(snapshot.updated_at || "") || 0;
         const localTime = Date.parse(localRun.snapshot.updated_at || "") || 0;
         if (remoteTime > localTime) {
@@ -874,11 +892,33 @@ export async function startAgent({
     sqmLocalReviewWorker = createWorker({
       api: sqmApi,
       beginWork: () => acquireUpdateWork(updateManager),
-      runReview: async (request, provider) => {
+      runReview: async (request, provider, context = {}) => {
         const scope = await authorizeDesktopWork();
-        const beforeDispatch = () => authorizeDesktopWork(scope);
-        return (dependencies.runSqmLocalReview || ((value, selected, options) => runStructuredProviderReview(selected, value, options)))(request, provider, { beforeDispatch, readOnlyAnalysis: true });
+        const actor = JSON.stringify([apiOrigin, session?.account_id, session?.organization_id, session?.device_id]);
+        const settings = await readSqmSettings(sessionRoot, apiOrigin, session);
+        const engineSettings = JSON.stringify([settings.enable_engine_fallback, settings.allowed_fallback_providers]);
+        const beforeDispatch = async () => {
+          await authorizeDesktopWork(scope);
+          const current = await readSqmSettings(sessionRoot, apiOrigin, session);
+          if (actor !== JSON.stringify([apiOrigin, session?.account_id, session?.organization_id, session?.device_id]) ||
+            engineSettings !== JSON.stringify([current.enable_engine_fallback, current.allowed_fallback_providers])) {
+            throw Object.assign(new Error("SQM actor or fallback settings changed."), { code: "desktop_access_required" });
+          }
+          await context.beforeLeaseDispatch?.();
+          await authorizeDesktopWork(scope);
+          const finalSettings = await readSqmSettings(sessionRoot, apiOrigin, session);
+          if (actor !== JSON.stringify([apiOrigin, session?.account_id, session?.organization_id, session?.device_id]) ||
+            engineSettings !== JSON.stringify([finalSettings.enable_engine_fallback, finalSettings.allowed_fallback_providers])) {
+            throw Object.assign(new Error("SQM actor or fallback settings changed."), { code: "desktop_access_required" });
+          }
+        };
+        return (dependencies.runSqmLocalReview || ((value, selected, options) => runStructuredProviderReview(selected, value, options)))(request, provider, {
+          beforeDispatch, readOnlyAnalysis: true, enableEngineFallback: settings.enable_engine_fallback,
+          allowedFallbackProviders: settings.allowed_fallback_providers, providerReportingSupported: context.providerReportingSupported === true,
+        });
       },
+      getEnvironmentKey: () => JSON.stringify([apiOrigin, session?.account_id, session?.organization_id, session?.device_id,
+        session?.expires_at, sqmEnabled, sqmEngineSettings, ["codex", "claude"].map(provider => [providerStatuses[provider]?.state, providerStatuses[provider]?.version])]),
       onChange: notifyDesktopStatusChange,
       getWorkspaceUrl: () => {
         const slug = desktopAccess.snapshot().organizationSlug;
@@ -1164,6 +1204,7 @@ export async function startAgent({
               : indexedRepository?.local_path || null,
           dependencies,
           providerStatus: (await readProviderStatuses())[backend],
+          request: body,
         });
         json(res, 200, environment);
         return;
@@ -1271,7 +1312,8 @@ export async function startAgent({
           authNonce = randomBytes(16).toString("base64url");
           pendingOrganization = null;
           desktopLoginPending = false;
-          sqmEnabled = await readSqmEnabled(sessionRoot, apiOrigin, session);
+          sqmEngineSettings = await readSqmSettings(sessionRoot, apiOrigin, session);
+          sqmEnabled = sqmEngineSettings.enabled;
           if (changingScope) sqmLocalReviewWorker?.resetHistory?.();
           organizationState = "unknown";
           organizationItems = [];
@@ -1321,6 +1363,35 @@ export async function startAgent({
           .pipe(res);
         return;
       }
+      const purposeVerifyMatch = url.pathname.match(/^\/v1\/local-tasks\/([^/]+)\/verify-purpose$/);
+      if (req.method === "POST" && purposeVerifyMatch) {
+        const actorScope = desktopScopeRevision();
+        const actorSession = session;
+        const task = localTaskRegistry.get(decodeURIComponent(purposeVerifyMatch[1]));
+        if (!task) { json(res, 404, { error: "Local Task not found" }); return; }
+        const body = await readBoundedJson(req, 1024).catch(() => { throw Object.assign(new Error("Purpose verification body must be an empty JSON object."), { code: "purpose_verification_plan_invalid", status: 400 }); });
+        if (!body || Array.isArray(body) || typeof body !== "object" || Object.keys(body).length) {
+          json(res, 400, { error: "Verification uses only the frozen task plan.", code: "purpose_verification_plan_invalid" }); return;
+        }
+        const selected = await ensureRepositoryIndexed(task.repository);
+        const approvedInventoryPath = findRepository(repositoryInventory, task.repository)?.local_path;
+        const authorize = async () => {
+          if (actorScope !== desktopScopeRevision() || !isSessionValid(session, Date.now(), apiOrigin) || shuttingDown ||
+              selected?.local_path !== task.workspace_path || findRepository(repositoryInventory, task.repository)?.local_path !== approvedInventoryPath ||
+              registry.list().some(run => run.request?.repository === task.repository && (run.operation || run.child || ["queued", "dispatching", "coding"].includes(run.status))) ||
+              localTaskRegistry.list().some(other => other !== task && other.workspace_path === task.workspace_path && ["queued", "running", "stopping"].includes(other.status))) {
+            throw Object.assign(new Error("Purpose verification scope or workspace changed."), { code: "purpose_verification_scope_changed", status: 403 });
+          }
+          await authorizeDesktopWork(requestAccessScope);
+          if (actorScope !== desktopScopeRevision()) throw Object.assign(new Error("Purpose verification actor changed."), { code: "purpose_verification_scope_changed", status: 403 });
+        };
+        await authorize();
+        await withUpdateWork(updateManager, async () => verifyPurposeLocalTask(task, {
+          session: actorSession, apiOrigin, keyPath: purposeKeyPath, authorize, persist: persistLocalTask,
+        }));
+        json(res, 200, { ok: true, task: localTaskSnapshot(task) });
+        return;
+      }
       const localTaskStreamMatch = url.pathname.match(/^\/v1\/local-tasks\/([^/]+)\/stream$/);
       if (req.method === "GET" && localTaskStreamMatch) {
         if (!isSessionValid(session, Date.now(), apiOrigin)) {
@@ -1391,7 +1462,13 @@ export async function startAgent({
           json(res, 401, { error: "simy session expired; run simy again" });
           return;
         }
+        const purposeStartScope = desktopScopeRevision();
         const { body, attachments } = await readCodingLoopStart(req);
+        const purposeVerification = normalizePurposeVerification(body.purpose_verification);
+        if (purposeVerification) {
+          assertPurposeActor(purposeVerification, session, apiOrigin, null, String(req.headers["idempotency-key"] || body.client_request_id || "").trim());
+          if (purposeStartScope !== desktopScopeRevision() || !purposeSigner || purposeSigner.key_id !== purposeVerification.key_id) throw Object.assign(new Error("Purpose verification signer or actor changed."), { code: "purpose_verification_scope_changed", status: 403 });
+        }
         const instruction = String(body.instruction || "").trim();
         if (!instruction) {
           json(res, 400, { error: "instruction is required" });
@@ -1510,11 +1587,14 @@ export async function startAgent({
             repository,
             expectedOutputs: body.expected_outputs,
             providerTokenBudget: body.provider_token_budget,
+            continuationPolicy: body.continuation_policy,
+            providerTokenCeiling: body.provider_token_ceiling,
             timeoutMs: body.timeout_ms,
             attachmentManifest: preparedAttachments,
             scenarioPack,
             workflowObservationBinding,
             workflowStartIntent,
+            purposeVerification,
           });
           hashLocalTaskIdempotencyKey(idempotencyKey);
         } catch (error) {
@@ -1636,20 +1716,36 @@ export async function startAgent({
             expectedOutputs: body.expected_outputs,
             scenarioPack,
             providerTokenBudget: body.provider_token_budget,
+            continuationPolicy: body.continuation_policy,
+            providerTokenCeiling: body.provider_token_ceiling,
             timeoutMs: body.timeout_ms,
             idempotencyKeyHash: hashLocalTaskIdempotencyKey(idempotencyKey),
             requestFingerprint,
             workflowObservationBinding,
             workflowStartIntent,
+            purposeVerification,
           });
+          if (purposeVerification) await preparePurposeVerification(task, { session, apiOrigin, clientRequestId: idempotencyKey, keyPath: purposeKeyPath,
+            authorize: async () => {
+              await authorizeDesktopWork(requestAccessScope);
+              if (purposeStartScope !== desktopScopeRevision() || !isSessionValid(session, Date.now(), apiOrigin)) throw Object.assign(new Error("Purpose verification actor changed."), { code: "purpose_verification_scope_changed", status: 403 });
+            } });
           await persistLocalTask(task);
           localTaskRegistry.create(task);
           requestClaim.complete(task);
           requestClaimCompleted = true;
           const runLocalTask =
             dependencies.runLocalTask ||
-            ((localTask) => startLocalTask(localTask, { session, apiOrigin }));
+            ((localTask) => startLocalTask(localTask, { session, apiOrigin, beforeSpawn: async () => {
+              if (purposeWorkspaceBusy(localTask.workspace_path)) throw Object.assign(new Error("Independent verification owns this workspace."), { code: "purpose_verification_workspace_busy" });
+              if (purposeVerification) {
+                await authorizeDesktopWork(requestAccessScope);
+                assertPurposeActor(purposeVerification, session, apiOrigin, localTask, idempotencyKey);
+                if (purposeStartScope !== desktopScopeRevision() || !isSessionValid(session, Date.now(), apiOrigin)) throw Object.assign(new Error("Purpose verification actor changed."), { code: "purpose_verification_scope_changed" });
+              }
+            } }));
           await authorizeDesktopWork(requestAccessScope);
+          if (purposeVerification && (purposeStartScope !== desktopScopeRevision() || !isSessionValid(session, Date.now(), apiOrigin))) throw Object.assign(new Error("Purpose verification actor changed."), { code: "purpose_verification_scope_changed", status: 403 });
           task.operation_promise = Promise.resolve(runLocalTask(task))
             .catch(async (error) => {
               task.status = "failed";
@@ -1917,6 +2013,10 @@ export async function startAgent({
                 : indexedRepository?.local_path || null,
             base_branch: typeof body.base_branch === "string" ? body.base_branch : "dev",
             max_attempts: body.max_attempts,
+            continuation_policy: body.continuation_policy,
+            provider_token_ceiling: body.provider_token_ceiling,
+            attempt_ceiling: body.attempt_ceiling,
+            max_no_progress_attempts: body.max_no_progress_attempts,
             retry_budget: body.retry_budget,
             token_budget: body.provider_token_budget ?? body.token_budget,
             ui_evidence_root:
@@ -2011,6 +2111,66 @@ export async function startAgent({
         return;
       }
 
+      const recoveryMatch = agenticLoopPath.match(/^\/v1\/agentic-loop\/([^/]+)\/recovery$/);
+      if (recoveryMatch && ["GET", "POST"].includes(req.method)) {
+        if (!isSessionValid(session, Date.now(), apiOrigin)) {
+          json(res, 401, { error: "Restore the authorized SIMY session before recovering this run." }); return;
+        }
+        const run = registry.get(decodeURIComponent(recoveryMatch[1]));
+        if (!run) { json(res, 404, { error: "run not found" }); return; }
+        try {
+          const recoveryScope = desktopScopeRevision();
+          const authorizeRecovery = async () => {
+            await authorizeDesktopWork(requestAccessScope);
+            if (recoveryScope !== desktopScopeRevision() || !isSessionValid(session, Date.now(), apiOrigin)) {
+              throw Object.assign(new Error('The authorized desktop session changed.'), { code: 'desktop_access_required' });
+            }
+            assertRecoveryActor({ ...run, session, apiOrigin });
+          };
+          assertRecoveryActor({ ...run, session, apiOrigin });
+          if (req.method === "GET") {
+            const recovery = await prepareLocalCodingRunRecovery(run);
+            await authorizeRecovery();
+            json(res, 200, { recovery });
+          } else {
+            const body = await readJson(req);
+            await authorizeRecovery();
+            if (body.action !== "reconcile") { json(res, 400, { error: "Choose reconcile after inspecting the interrupted operation." }); return; }
+            const recovery = await reconcileLocalCodingRun(run, {
+              operationStatus: body.operation_status, observation: body.observation, evidenceRefs: body.evidence_refs, authorize: authorizeRecovery,
+            });
+            json(res, 200, { recovery, run_id: run.id, resumed: false });
+          }
+        } catch (error) {
+          json(res, 409, { error: error instanceof Error ? error.message : "Recovery could not be verified.", code: error.code || "recovery_rejected" });
+        }
+        return;
+      }
+
+      const ceilingMatch = agenticLoopPath.match(/^\/v1\/agentic-loop\/([^/]+)\/continuation-ceilings$/);
+      if (req.method === "POST" && ceilingMatch) {
+        if (!isSessionValid(session, Date.now(), apiOrigin)) {
+          json(res, 401, { error: "Restore the authorized SIMY session before extending this run." }); return;
+        }
+        const run = registry.get(decodeURIComponent(ceilingMatch[1]));
+        if (!run) { json(res, 404, { error: "run not found" }); return; }
+        try {
+          const recoveryScope = desktopScopeRevision();
+          assertRecoveryActor({ ...run, session, apiOrigin });
+          const body = await readJson(req);
+          await authorizeDesktopWork(requestAccessScope);
+          if (recoveryScope !== desktopScopeRevision() || !isSessionValid(session, Date.now(), apiOrigin)) {
+            throw Object.assign(new Error('The authorized desktop session changed.'), { code: 'desktop_access_required' });
+          }
+          assertRecoveryActor({ ...run, session, apiOrigin });
+          const budget = await extendLocalCodingRunCeilings(run, body);
+          json(res, 200, { run_id: run.id, budget, resumed: false });
+        } catch (error) {
+          json(res, 409, { error: error instanceof Error ? error.message : "Continuation ceilings could not be extended.", code: error.code || "continuation_ceiling_rejected" });
+        }
+        return;
+      }
+
       const providerBudgetMatch = agenticLoopPath.match(
         /^\/v1\/agentic-loop\/([^/]+)\/provider-token-budget$/,
       );
@@ -2213,6 +2373,7 @@ export async function startAgent({
         json(res, 400, localTaskFileErrorPayload(err));
         return;
       }
+      if (String(err?.code || "").startsWith("purpose_verification_")) { json(res, err.status || 409, { error: err.message, code: err.code }); return; }
       json(res, 500, { error: err instanceof Error ? err.message : "internal error" });
     }
   });
@@ -2330,7 +2491,9 @@ export async function startAgent({
       tasks: desktopTaskStatus([registry, directRegistry, localTaskRegistry]),
       organizations: { state: organizationState, items: organizationItems, selectedId: session?.organization_id ?? null,
         switching: Boolean(pendingOrganization), canSwitch: auth.state === "signed_in" && !organizationBusy() },
-      sqm: { enabled: sqmEnabled, canEnable: auth.state === "signed_in" && Boolean(session?.account_id && session?.organization_id) && !pendingOrganization && desktopWorkAllowed(),
+      sqm: { enabled: sqmEnabled, engineFallbackEnabled: sqmEngineSettings.enable_engine_fallback,
+        allowedFallbackProviders: sqmEngineSettings.allowed_fallback_providers,
+        paused: sqmLocalReviewWorker?.snapshot?.().paused === true, pauseReason: sqmLocalReviewWorker?.snapshot?.().pause_reason ?? null, canEnable: auth.state === "signed_in" && Boolean(session?.account_id && session?.organization_id) && !pendingOrganization && desktopWorkAllowed(),
         activeJobId: sqmLocalReviewWorker?.snapshot?.().active_job_id ?? null,
         issue: sqmLocalReviewWorker?.snapshot?.().issue ?? null,
         providerPause: sqmLocalReviewWorker?.snapshot?.().provider_pause ?? null,
@@ -2382,10 +2545,22 @@ export async function startAgent({
       await writeSqmEnabled(sessionRoot, apiOrigin, settingsSession, enabled);
       if (session !== settingsSession) return;
       sqmEnabled = enabled;
+      sqmEngineSettings = { ...sqmEngineSettings, enabled };
+      sqmLocalReviewWorker?.retry?.();
       refreshAvailableCapabilities();
       notifyDesktopStatusChange();
       void heartbeat().catch(error => reportHeartbeatError(error, quiet));
     },
+    async setSqmEngineFallback({ enabled, allowedProviders }) {
+      await authorizeDesktopWork();
+      if (!isSessionValid(session, Date.now(), apiOrigin) || pendingOrganization) throw new Error("Sign in to an organization first.");
+      const settingsSession = session;
+      await writeSqmEngineFallback(sessionRoot, apiOrigin, settingsSession, { enabled, allowedProviders });
+      if (session !== settingsSession) return;
+      sqmEngineSettings = await readSqmSettings(sessionRoot, apiOrigin, session);
+      sqmLocalReviewWorker?.retry?.(); notifyDesktopStatusChange();
+    },
+    async retrySqmEnvironment() { await authorizeDesktopWork(); sqmLocalReviewWorker?.retry?.(); notifyDesktopStatusChange(); },
     getDesktopStatus,
     getDesktopTask: async (key) => {
       if (key?.startsWith("sqm:") || key?.startsWith("sqm-check:")) {
@@ -2831,6 +3006,7 @@ function capabilitiesFromInspections(codex, claude) {
       local_task_streaming: true,
       local_task_idempotency: true,
       local_task_model: true,
+      progress_continuation_policy: true,
       local_task_reconciliation: true,
       local_task_control_plane: true,
       pipeline_local_task_artifacts: true,
@@ -2862,6 +3038,7 @@ async function inspectCodingLoopEnvironment({
   localPath,
   dependencies,
   providerStatus,
+  request = {},
 }) {
   const checks = [
     {
@@ -2922,13 +3099,16 @@ async function inspectCodingLoopEnvironment({
   if (inspection.compatible) {
     checks.push({
       key: "provider_auth",
-      status: authFailure ? "failed" : "passed",
+      status: authFailure ? (providerStatus?.state === "signed_out" ? "failed" : "unknown") : "passed",
       summary: authFailure?.summary || `${backend === "claude" ? "Claude Code" : "Codex"} is signed in.`,
       ...(authFailure ? { recovery: authFailure.recovery } : {}),
     });
   }
+  checks.push(...startupRequirementChecks(request));
+  const startup = summarizeStartupChecks(checks);
   return {
-    ready: Boolean(repositoryPath && inspection.compatible && !authFailure),
+    ready: Boolean(repositoryPath && inspection.compatible && !authFailure && startup.blocker_keys.length === 0),
+    startup,
     repository,
     backend,
     execution_target: executionTarget,

package/src/backend-executable.js

@@ -1,5 +1,6 @@
 import spawn from "cross-spawn";
 import os from "node:os";
+import { readdir } from "node:fs/promises";
 import { readdirSync, statSync } from "node:fs";
 import path from "node:path";
 
@@ -26,11 +27,12 @@ export async function resolveBackendExecutable(backend, options = {}) {
 
 export async function inspectBackendExecutable(
   backend,
-  { spawnImpl = spawn, candidates = backendCandidates(backend), timeoutMs = 5_000, signal } = {},
+  { spawnImpl = spawn, candidates, timeoutMs = 5_000, signal, discoveryOptions = {} } = {},
 ) {
   const assertActive = () => { if (signal?.aborted) throw Object.assign(new Error('Backend inspection was cancelled.'), { code: 'cancelled' }); };
   assertActive();
   const policy = versionPolicy(backend);
+  candidates ??= [...backendCandidates(backend, discoveryOptions), ...await bundledClaudeCandidates(backend, discoveryOptions)];
   for (const candidate of [...new Set(candidates.filter(Boolean))]) {
     assertActive();
     const probe = await probeExecutable(candidate, { spawnImpl, timeoutMs, signal });
@@ -96,6 +98,24 @@ export function backendCandidates(backend, {
   ];
 }
 
+// Claude Desktop can install below <version>/<content hash>/claude.app while
+// an older PATH wrapper still points at <version>/claude.app. Discover only the
+// known app layout; never recursively execute arbitrary files from user folders.
+export async function bundledClaudeCandidates(backend, { platform = process.platform, home = os.homedir(), readDirectory = readdir } = {}) {
+  if (backend !== "claude" || platform !== "darwin") return [];
+  const root = path.posix.join(home, "Library", "Application Support", "Claude", "claude-code");
+  const entries = async target => { try { return await readDirectory(target, { withFileTypes: true }); } catch { return []; } };
+  const versions = (await entries(root)).filter(item => item.isDirectory() && /^\d+\.\d+\.\d+$/.test(item.name))
+    .sort((a, b) => compareVersions(b.name, a.name)).slice(0, 8);
+  const result = [];
+  for (const version of versions) {
+    const base = path.posix.join(root, version.name);
+    const hashes = (await entries(base)).filter(item => item.isDirectory() && /^[a-f0-9]{8,128}$/i.test(item.name)).slice(0, 8);
+    for (const hash of hashes) result.push(path.posix.join(base, hash.name, "claude.app", "Contents", "MacOS", "claude"));
+    result.push(path.posix.join(base, "claude.app", "Contents", "MacOS", "claude"));
+  }
+  return result;
+}
 // Desktop installs use either <version>/claude.app or <version>/<hash>/claude.app.
 // Only inspect these two known levels; never recurse through arbitrary home paths
 // or descend through version/hash directory symlinks. Existing explicit/PATH installs retain priority.

package/src/cli-contract.js

@@ -13,7 +13,7 @@ import {
 } from "./durable-local-task-steps-contract.js";
 import { buildExecutionCapabilityContract } from "./execution-capability-contract.js";
 
-export const CLI_VERSION = "0.5.72";
+export const CLI_VERSION = "0.5.73";
 export const CLI_API_CONTRACT_VERSION = 10;
 
 export function withCliContract(

package/src/codex-plugins/simy/hooks/README.md

@@ -31,6 +31,28 @@ keyed by provider and session ID and bound to the canonical working directory.
 Historical hook observations do not prove hooks are still enabled. See the
 bundled delivery-loop skill for checkpoint commands and explicit recovery.
 
+Managed tasks may explicitly select `--continuation-policy progress`. Existing
+tasks without that choice keep their bounded contract. Progress mode treats the
+ordinary time, continuation and token thresholds as warnings, with independent
+hard ceilings of 24 hours from the original start, 50 continuations and, on the
+host-observed Codex channel, 15 million provider tokens. Claude live hooks expose
+no trustworthy usage stream; their status explicitly reports that token limits
+are unsupported. Required approvals, missing observed usage, manual stops and
+unreconciled effects continue to block dispatch.
+
+The host records passing, failed and unevaluated frozen acceptance checks.
+Provider summaries, test counts, renamed evidence and restored old criteria
+cannot manufacture progress. Three attempts without new verified criteria, or
+30 minutes without verified progress at a dispatch/checkpoint boundary, pause
+the original run. Active tools are allowed to reach their checkpoint within the
+hard ceiling; activity alone does not prove progress.
+
+`delivery extend` raises explicitly selected ceilings within those global caps.
+It records the old and new bounds, preserves run identity, usage, checkpoints,
+criteria and approvals, and does not start a new provider turn. A separate resume
+is required after an exhausted run becomes paused. `delivery continuation_policy`
+requires an explicit choice and never infers progress consent from an extension.
+
 ## Project 実装状況 (dev-work) hook
 
 `dev-work.mjs` runs on Stop in both distributions. In a Git work tree with a

package/src/codex-plugins/simy/hooks/continuation-policy.mjs

@@ -0,0 +1,27 @@
+// Missing policy means the persisted legacy budget contract, never implicit consent.
+export const CONTINUATION_POLICIES = Object.freeze(["bounded", "progress"]);
+export const MAX_COMPLETION_TOKENS = 15_000_000;
+export const DEFAULT_COMPLETION_ATTEMPT_CEILING = 50;
+
+export function continuationMode(value) {
+  if (value === undefined || value === null) return "bounded";
+  if (!CONTINUATION_POLICIES.includes(value)) throw new Error("Invalid continuation policy.");
+  return value;
+}
+
+export function progressLimits(input = {}) {
+  return {
+    continuation_policy: continuationMode(input.continuation_policy),
+    provider_token_ceiling: finiteLimit(input.provider_token_ceiling, MAX_COMPLETION_TOKENS, MAX_COMPLETION_TOKENS),
+    attempt_ceiling: finiteLimit(input.attempt_ceiling, DEFAULT_COMPLETION_ATTEMPT_CEILING, 100),
+    max_no_progress_attempts: finiteLimit(input.max_no_progress_attempts, 3, 20),
+  };
+}
+
+function finiteLimit(value, fallback, maximum) {
+  if (value === undefined || value === null) return fallback;
+  if (!Number.isSafeInteger(value) || value < 1 || value > maximum) {
+    throw new Error(`Continuation ceiling must be an integer between 1 and ${maximum}.`);
+  }
+  return value;
+}