v1.86.0.0 feat: route outside reviews by harness (#2850)

* feat: add a restricted and supervised Claude Code runner

Preserve configured authentication and models while enforcing tool access, strict completion JSON, bounded output and process cleanup. Cover argv, failure handling, session metadata and Windows process containment.

* feat: route outside reviews by harness and migrate wrapper installs

Use Claude Code from Codex and Codex from other supported hosts, with shared invocation rendering, positive gate validation and per-phase provenance. Rename /claude to /claude-code, repair managed shared and copied installations safely, and generate native Kiro skills. Add installed-workflow, failure-injection and live cross-harness regression coverage.

* test: recognize CEO mode labels without terminal spacing

The paid workflow rendered SCOPEEXPANSION at option 4, but its driver required a literal space. Match the leading mode title without cursor-spacing artifacts and ignore adjacent preview text. Preserve missing-target failures and downstream posture assertions.

* test: isolate plan-count fixtures before starting review workflows

Seed the complete test plan in a private git repository before launching Claude, so a bare slash command cannot review the live workspace while a delayed fixture message remains queued. Preserve count thresholds, parsers and budgets. Add initial-context and installed-discovery tests, and retain startup/terminal diagnostics on failed evaluations.

* test: stabilize review fixtures and Claude eval startup

Preserve source boundaries in workflow judge inputs, isolate CEO mode plans, and wait for interactive trust input readiness. Keep startup failure evidence and retain existing models, budgets, and assertions.

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* test: classify collapsed review modes and isolate seeded findings

Keep review questions out of the setup count when terminal cursor positioning removes spaces. State existing webhook safeguards so the five-finding control measures its seeded defects without accidental extra security and concurrency gaps. Preserve question bands and the paired control.

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* test: isolate browser daemon state across free shards

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* test: stabilize native review counting and interactive navigation

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* chore: prepare v1.82.0.0 release

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* fix: eliminate browser and process-cleanup test flakes

Pin every CI surface to Bun 1.4.0 to avoid extra-stdio finalizers closing
reused live sockets. Add an isolated GC/listener regression that fails on
Bun 1.3.13, and prevent coordinated rollback to an affected CI runtime.

Check renderer cleanup against the render's own staging directory so
concurrent renders cannot invalidate the assertion. Make the no-pgrep
process-tree walk tolerate disappearing /proc entries, and synchronize
its test fixture through child readiness and pipe EOF instead of sleeps.

Validation: 9,157 passed, 31 skipped, zero failures across 556 files with
retries disabled. Build, all-host generation freshness, and skill checks
passed. All three races have failing-before/passing-after regressions.

* fix: count completed native review questions in evals

* fix: drive review navigation from confirmed native choices

* fix: require complete section-loading eval reports

* test: isolate telemetry HTTP transport from local assertions

* fix: keep review input on the active native question

* test: let tunnel revocation daemon choose an available port

* test: allocate available ports for pairing and watchdog fixtures

* fix: stabilize planning eval navigation and phase reporting

* test: isolate installed runtime paths in planning evals

* test: stabilize review evidence and concurrent refresh fixtures

* fix: resolve design findings before editing the plan

* fix: honor and persist disabled outside plan reviews

* fix: preserve planning decisions and terminal evidence

Load installed host reviews at autoplan phase entry and wait for completed
reviewers and saved artifacts. Reuse approved remedies while preserving
individual finding decisions.

Drive interactive evals from the current terminal viewport, bind native
questions across scrolling, and require complete native report evidence.
Cover captured stale menus, permission lifecycles, setup classification,
and disabled-review tool availability with deterministic regressions.

Advance release metadata and the upgrade migration to the unclaimed
1.83.0.0 slot.

* fix: drive native review questions and preserve current plans

Use the native single-choice keyboard protocol and current terminal viewport,
with per-question navigation inside packets and completed-call coverage.
Keep permissions, multi-select menus, and Submit controls distinct.

Send Autoplan reviewers the amended implementation plan, keep its review record
separate, and supply retained application contracts in the chain fixture.
Clarify individual DevEx decisions and complete CEO fix options; use one active
plan destination for the section-loading report.

* fix: preserve complete plan-review decisions

* fix: recognize native plan dialogs and reviewer controls

* fix: preserve review decisions and phase completion

* fix: recognize completed reviews without losing findings

* fix: preserve review continuity and native eval completion

* test: fix native review completion and eval retry isolation

* test: handle native review menus and complete eval fixtures

* test: fix native review setup, completion, and isolation failures

* test: limit native skill discovery to runtime assets

* fix: bind Autoplan reviews to full ordered phase inputs

* test: fix planning eval routing, counting, and timeout handling

* chore: advance queued release to v1.84.0.0

* fix: preserve complete review inputs and planning decisions

* fix: reconcile review approvals and preserve phase obligations

* fix: preserve review obligations and unblock eval permissions

Carry recorded Autoplan requirements into blind phase inputs, require Eng
review approvals before exit, and exercise combined asynchronous flows in
CEO reviews. Correct native finding and handoff classification and unblock
repeated report edits using scoped request identities.

* fix: retain plan requirements and complete native review dialogs

* fix: complete native review prompts and retain plan references

* fix: preserve review inputs and classify native eval evidence

* fix: check competing completion orders in CEO reviews

* fix: recognize review decisions and require phase methodology

Require the current phase methodology before Autoplan snapshots. Correct
substantive decision, closed handoff, and cache-finding classification, and
honor the recommended implementation approach in native review dialogs.

Add captured-transcript regressions without changing review thresholds,
provider models, retries, or deadlines.

* test: bind native review decisions and close completed handoffs

* fix: complete review dialogs and verify methodology delivery

* fix: preserve review evidence and unblock native eval prompts

* fix: handle native review question completions

* fix: recognize native review narration and controls

* fix: count native review decisions and isolate eval fixtures

* test: verify seeded review coverage and current artifact permissions

* test: isolate model and brain-aware skill renders

* fix: repair native workflow evaluation and clarify review steps

* fix: stabilize workflow eval evidence and review guidance

* test: repair native workflow observation and fixture isolation

* fix: recognize completed workflow evidence and owned skill reads

* test: repair seeded workflow delivery and completion evidence

* test: recognize current review evidence across native forms

* test: handle native review variants and permission redraws

* fix: honor review preferences and recognize native eval evidence

* test: recognize completed review decisions and queued permissions

* test: match current review contracts and partial-line edits

* test: recognize completed workflow evidence and bounded human waits

* fix: preserve review entry gates and native eval interactions

* fix: recognize native workflow evidence and preserve review gates

* test: recognize current review evidence and preconfigure workflow fixtures

* test: recognize completed review findings and scoped artifact permissions

* fix: stabilize native workflow review and permission evidence

* fix: recognize current review evidence and scoped edit confirmations

Clarify Design and engineering review entry instructions and Design scoring.
Recognize required legacy coverage and public Autoplan completion recaps.
Bind the pending Edit confirmation to its exact file, ordered digest, and
one-request approval when a preceding command display remains visible.
Keep reviews within their existing size limits and preserve scope gates
when extracting workflow fixtures from either supported preamble header.

Keep failure outcomes, review thresholds, provider choices, and eval budgets.

* fix: recover review workflow progress and eval evidence

* fix: recognize valid review evidence and scope selection

* test: fix review evidence parsing and repeated artifact prompts

* test: recognize valid review decisions and pending native cards

* fix(plan-eng-review): keep final navigation consistent with approved tasks

* test: recognize valid review evidence and bind legacy diff requests

* fix: stabilize review eval evidence and harness repair guidance

* docs: update project documentation for v1.85.0.0

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* test: fix Windows CI fixtures and credential scan

Rebase captured JSON values and filesystem evidence using the appropriate
path convention. Compile native fake CLIs on Windows and synchronize pipe
holder readiness, with cleanup retained when assertions fail.

Assemble synthetic credential fixtures at runtime so the added-line scan
keeps enforcing the same gate without flagging its own rejection controls.

Discover generated skills directly for the empty-find regression check,
avoiding a recursive scan through saved evaluation artifacts and dependencies.

* fix: preserve source renders on Windows

Compare canonical generator paths using native separators so an output
sidecar pointing at the source cannot overwrite its skill or metadata.
Keep the regression fixture isolated from the real checkout and expose
freshness diagnostics before asserting subprocess status.

Detach Windows drain-test pipe holders from the fake provider's automatic
child cleanup while preserving the enclosing runner job and its assertions.

* fix: clarify outside review fallback and CEO decisions

Render one applicable own-harness fallback path and retain native review,
disabled policy, and missing-coverage semantics. Align report field names
and mode labels, and make the existing per-cut scope approval explicit.

Regenerate skill outputs and keep the workflow judge's model, thresholds,
and retry policy unchanged.

* chore: move release to free version slot (v1.86.0.0)

PR #2852 now claims v1.85.0.0. Align the release metadata and
rename migration so upgrades from that version still receive it.

Co-Authored-By: OpenAI Codex <noreply@openai.com>

* fix: include engineering review prerequisites and restore branch context

* fix: recognize coverage diagrams and clarify design review instructions

* fix: preserve file identities and join Windows test processes

---------

Co-authored-by: OpenAI Codex <noreply@openai.com>

test/skill-e2e-plan-design-finding-count.test.ts

@@ -10,6 +10,8 @@
 
 import { test } from 'bun:test';
 import { describeE2ETier } from './helpers/e2e-gate';
+import { isDesignCountFirstReview, isDesignCountSetup, isDesignCompletionHandoff, pickDesignCountQuestion } from './helpers/design-count-review';
+import { isDesignArtifactGeneration } from './helpers/design-artifact-question';
 import * as fs from 'node:fs';
 import * as os from 'node:os';
 import * as path from 'node:path';
@@ -25,11 +27,178 @@ const N = 5;
 const FLOOR = N - 1;
 const CEILING = N + 2;
 
+// Existing interaction behavior belongs to the surrounding fixture, not the
+// five intentionally inconsistent visual requirements under review.
+const existingInteractionStates = [
+  'The existing router protects dirty edits on every in-app exit, including',
+  'persistent app navigation, using the same Cancel confirmation dialog.',
+  'Register the browser-native beforeunload warning only while the form is dirty;',
+  'remove it when clean. Confirmed in-app navigation uses the existing destination',
+  'heading focus behavior; Keep editing returns focus to the attempted exit.',
+  'During Save or Export, both request buttons use aria-disabled=true plus an',
+  'explicit click/keyboard activation guard, rather than the HTML disabled attribute.',
+  'They remain focusable and keep the existing disabled appearance. Reset and',
+  'Cancel use HTML disabled during the request. Do not move focus while pending',
+  'or after success. On a network error, focus the operation-specific Retry only',
+  'if focus is still on the request trigger; never steal focus the user moved.',
+  'The existing InlineStatus text stays unchanged while Save is pending:',
+  'Unsaved changes for a dirty form, otherwise its saved timestamp or initial',
+  'blank text. Pending feedback belongs to the request button; do not repeat',
+  'Saving… in the status live region. Success and failure use the outcomes above.',
+  'When clean and idle, Reset is disabled because it has nothing to discard,',
+  'and Cancel navigates back immediately without a confirmation. When dirty',
+  'and idle, Reset and Cancel use their existing discard confirmations. Their',
+  '44px geometry is unchanged; the disabled style is separate from pending feedback.',
+];
+
+// A known surrounding design prevents missing layout/journey/state contracts
+// from becoming legitimate extra findings unrelated to the five seeded gaps.
+const designSystem = [
+  '# Approved account-settings design system',
+  '',
+  'Reuse the existing single-column settings shell: persistent app navigation,',
+  'page title and description, action group, then Profile and Notifications fieldsets.',
+  'The existing description is “Manage your display name, email address, and notification preferences.”',
+  'Preserve that description verbatim.',
+  'The form has a 640px maximum width. Reuse existing Button, Field, InlineStatus,',
+  'ErrorSummary and ConfirmationDialog components; no new component family is needed.',
+  'Profile contains Display name (text) and Email (email). Notifications contains',
+  'Weekly digest and Product tips switches. Existing server defaults are the',
+  'account name/email, Weekly digest on, and Product tips off; labels/order stay fixed.',
+  'The page title “Account settings” is h1. Profile and Notifications are h2',
+  'headings that label their fieldsets via aria-labelledby; no heading level is skipped.',
+  'The existing DOM and visual order are:',
+  '```text',
+  'Persistent app navigation',
+  'main: Account settings (h1) + description',
+  '  Save | Reset | Cancel | Export',
+  '  InlineStatus',
+  '  Profile (h2): Display name, Email',
+  '  Notifications (h2): Weekly digest, Product tips',
+  '```',
+  '',
+  'Save is the only filled primary action (#1d4ed8 with white text). Reset, Cancel',
+  'and Export are neutral ghost buttons; destructive intent is explained in the',
+  'existing confirmation dialog, whose default is Cancel. All targets are at least 44px.',
+  'Spacing uses an 8px base: sections 32px, field groups 24px, label-to-input 8px.',
+  'Typography has two roles: 16px body/form labels/helper text, 20px section headings.',
+  'The existing app font family is system-ui, sans-serif, inherited by form controls.',
+  'Use error.text #991b1b on error.surface #fef2f2 with an icon and explicit text.',
+  'All text must meet WCAG AA contrast; never communicate status through color alone.',
+  'Focus-visible is the existing 2px solid #1d4ed8 outline, offset 2px on white;',
+  'its measured contrast exceeds 3:1. Reuse it on all controls and dialog actions.',
+  '',
+  'The established pending-action pattern is an inline spinner beside “Saving…”',
+  'inside the disabled Save button, aria-busy=true, with reduced-motion support.',
+  'Save and Export are mutually exclusive: disable both while either is pending.',
+  'Reset and Cancel are also disabled while either request is pending; all four',
+  'header actions return to their idle/dirty-state behavior when it settles. Export uses the existing inline spinner',
+  'beside “Exporting…” inside its disabled button, aria-busy=true, with reduced-motion support.',
+  'After Save finishes, Export downloads the latest successfully saved preferences.',
+  'Success uses the persistent inline status “Saved” plus the save time (aria-live=polite).',
+  'The existing formatter renders “Saved at HH:mm” in the user’s local 24-hour time.',
+  'Editing a field away from its saved value changes that same persistent status',
+  'to “Unsaved changes” (role=status, aria-live=polite, aria-atomic=true); text',
+  'communicates the dirty state without relying on color or button enabled state.',
+  'Reverting all edits or confirming Reset restores “Saved at HH:mm” for the',
+  'last successful save. Failed saves retain “Unsaved changes” alongside the error.',
+  'Before the first successful save, the unchanged form has blank status text;',
+  'editing shows “Unsaved changes”, and reverting or confirming Reset restores blank text.',
+  'Reset dialog: “Discard unsaved changes?” / “Your saved preferences will be restored.”',
+  'Its actions are “Keep editing” (default) and “Discard changes”. Cancel navigation',
+  'uses the same dialog with “Keep editing” (default) and “Discard and leave”.',
+  'Closing a dialog while staying on the form returns focus to its Reset or Cancel',
+  'trigger; confirmed navigation uses the existing destination-main-heading focus behavior.',
+  'Validation errors appear beside fields and in a linked error summary; focus the',
+  'first invalid field. Network errors preserve edits and show Retry in that status area.',
+  'Retry is a sibling button beside the status text, outside its live region.',
+  'Visible text stays “Retry”; its aria-label is “Retry save” or “Retry export” for that operation.',
+  'The read-failure control follows the same pattern with aria-label “Retry loading”.',
+  ...existingInteractionStates,
+  'Export failures use that same error/retry area without changing unsaved fields;',
+  'Retry repeats Export. A successful download clears the Export error only.',
+  'At 640px and below, the three secondary actions share one equal-column row',
+  'below full-width Save, fitting 320px while preserving order and 44px targets.',
+].join('\n');
+
 const planDesign5Findings = (planPath: string) => [
   `Please review this plan thoroughly. As you go, write your plan-mode plan to ${planPath} (use Edit/Write to that exact path).`,
+  // The skill explicitly supports this opt-out. The cadence test measures
+  // review findings; generated-variant preference questions have separate evals.
+  'Text-only review; skip mockups. Review all seven design dimensions.',
+  'Native design review only: skip optional outside voices. Finish after the design review; I will handle subsequent reviews manually.',
   '',
   '# Plan: Settings Page UI redesign',
   '',
+  '## Existing product and accepted behavior',
+  'This updates an existing account-settings form using the checked-in DESIGN.md.',
+  'The shell and components already exist. Profile and Notifications are the only',
+  'sections, with visible headings and associated field labels. The page header',
+  'contains the title, a short description, and Save/Reset/Cancel/Export actions.',
+  'The existing description is “Manage your display name, email address, and notification preferences.”',
+  'Preserve that description verbatim.',
+  'Preserve the approved single-column structure and component behavior.',
+  'The page title “Account settings” is h1. Profile and Notifications are h2',
+  'headings that label their fieldsets via aria-labelledby; no heading level is skipped.',
+  'The existing DOM and visual order are:',
+  '```text',
+  'Persistent app navigation',
+  'main: Account settings (h1) + description',
+  '  Save | Reset | Cancel | Export',
+  '  InlineStatus',
+  '  Profile (h2): Display name, Email',
+  '  Notifications (h2): Weekly digest, Product tips',
+  '```',
+  '',
+  'Journey: a user arrives from account navigation wanting to adjust preferences,',
+  'edits the labeled fields, saves, and reads the inline Saved timestamp before',
+  'leaving. The feedback preserves confidence that their preferences were stored.',
+  'The persistent InlineStatus has role=status, aria-live=polite, aria-atomic=true.',
+  'After success it reads “Saved at HH:mm” in the user’s local 24-hour time.',
+  'Editing away from a saved value changes its text to “Unsaved changes”, so',
+  'dirty state never relies on color or the Save button being enabled. Reverting',
+  'all edits or confirming Reset restores the last successful save timestamp.',
+  'Before any successful save, unchanged values show blank status text; editing',
+  'shows “Unsaved changes”, and reverting or confirming Reset restores blank text.',
+  'Failed saves retain “Unsaved changes” alongside the error message.',
+  'Initial loading uses the existing form skeleton. A new account sees useful',
+  'default preferences as specified in DESIGN.md rather than an empty page. Read failures show Retry.',
+  'Save is atomic: all fields persist together or none do, so partial success is',
+  'not exposed. Field validation, network failure, and successful-save feedback',
+  'use the exact existing DESIGN.md patterns. Preserve unsaved values after errors.',
+  'Disable repeat Save submissions while pending. Save and Export are mutually',
+  'exclusive: disable both while either is pending. After Save finishes, Export',
+  'downloads the latest successfully saved preferences. Reset restores saved values only',
+  'after confirmation; Cancel confirms discarding dirty edits before returning to',
+  'the previous page; Export downloads the current saved preferences as JSON.',
+  'Reset and Cancel are disabled while Save or Export is pending; all four header',
+  'actions return to their idle/dirty-state behavior when it settles. While preparing Export, use the existing inline',
+  'spinner beside “Exporting…” inside its disabled button, aria-busy=true, with reduced-motion support.',
+  'An Export failure uses the existing inline error/retry area and preserves',
+  'unsaved fields. Retry repeats Export; success clears only that Export error.',
+  'Retry controls are siblings beside the status text, outside its live region.',
+  'Visible text stays “Retry”; its aria-label is “Retry save” or “Retry export” for that operation.',
+  'The read-failure control follows the same pattern with aria-label “Retry loading”.',
+  ...existingInteractionStates,
+  '',
+  'Responsive behavior: above 640px keep the header action group in one row; at',
+  '640px and below, place full-width Save first and the three secondary actions',
+  'in one equal-column row below it, preserving DOM/tab order. The form fits 320px',
+  'without horizontal scroll, including the secondary actions and their 44px targets.',
+  'All controls have visible focus rings and 44px targets. Use semantic fieldsets,',
+  'labels, a main landmark and heading order; errors link through aria-describedby.',
+  'Focus-visible on every control and dialog action is the existing 2px solid',
+  '#1d4ed8 outline, offset 2px on white, with measured contrast above 3:1.',
+  'Dialogs trap focus; Escape cancels. Closing while staying on the form restores',
+  'focus to the Reset or Cancel trigger. Confirmed navigation uses the existing',
+  'destination-main-heading focus behavior. Export remains a',
+  'clearly labeled button. Respect reduced motion. No additional visual exploration',
+  'or component replacement is part of this established form update.',
+  'Retain the existing system-ui, sans-serif font family, including on form controls.',
+  '',
+  '## Planned implementation gaps',
+  'The proposed form still has the following inconsistencies with that design:',
+  '',
   '## Visual Hierarchy',
   'The "Save" button is rendered with the same size, weight, and color as',
   'three other buttons in the page header (Reset, Cancel, Export). Nothing',
@@ -67,11 +236,17 @@ describeE2E('/plan-design-review per-finding AskUserQuestion count (periodic)',
           skillName: 'plan-design-review',
           slashCommand: '/plan-design-review',
           followUpPrompt: planDesign5Findings(planPath),
+          expectedPlanPath: planPath,
           isLastStep0AUQ: designStep0Boundary,
+          isFirstReviewAUQ: isDesignCountFirstReview,
+          isSetupAUQ: isDesignCountSetup,
+          isCompletionHandoffAUQ: isDesignCompletionHandoff,
+          isArtifactGenerationAUQ: isDesignArtifactGeneration,
+          fixtureFiles: { 'DESIGN.md': designSystem },
+          // Design's explicit opt-in is separate from codex_reviews. Keep
+          // this native-cadence fixture within its declared review scope.
+          pickAUQ: pickDesignCountQuestion,
           reviewCountCeiling: CEILING + 1,
-          // LIVE-REPO CWD: PTY session needs the repo cwd — gstack skill
-          // registry + hermetic pre-trusted dir (hermetic-env trustedDirs).
-          cwd: process.cwd(),
           timeoutMs: 1_500_000,
           env: { QUESTION_TUNING: 'false', EXPLAIN_LEVEL: 'default' },
         });
@@ -94,16 +269,22 @@ describeE2E('/plan-design-review per-finding AskUserQuestion count (periodic)',
         if (obs.reviewCount < FLOOR) {
           throw new Error(
             `BAND FAIL (below floor): reviewCount=${obs.reviewCount} < FLOOR=${FLOOR}.\n` +
-              `Likely batching regression. Review-phase fingerprints:\n` +
+              `outcome=${obs.outcome} step0=${obs.step0Count} elapsed=${obs.elapsedMs}ms\n` +
+              `summary: ${obs.summary}\n` +
+              `All captured fingerprints (including Step 0):\n` +
               obs.fingerprints
-                .filter((f) => !f.preReview)
-                .map((f) => `  - "${f.promptSnippet.slice(0, 80)}"`)
-                .join('\n'),
+                .map(
+                  (f) =>
+                    `  - preReview=${f.preReview} sig=${f.signature.slice(0, 12)} prompt="${f.promptSnippet}"`,
+                )
+                .join('\n') +
+              `\n--- evidence (last 3KB) ---\n${obs.evidence}`,
           );
         }
         if (obs.reviewCount > CEILING) {
           throw new Error(
-            `BAND FAIL (above ceiling): reviewCount=${obs.reviewCount} > CEILING=${CEILING}.`,
+            `BAND FAIL (above ceiling): reviewCount=${obs.reviewCount} > CEILING=${CEILING}.\n` +
+              `Captured observation:\n${JSON.stringify(obs, null, 2)}`,
           );
         }