v1.87.6.0 fix: make checks reliable and everyday validation faster (#2898)

* fix: acknowledge seeded plans before invoking review skills

* fix: distinguish current plan input from conversation history

* fix: keep hermetic plan reviews on manual permissions

* fix: distinguish tool discovery from file permission ownership

* fix: preserve initial plan mode in observation tests

* fix: wait for scope decisions before writing review findings

* fix: carry autoplan decisions consistently into review artifacts

* test: retain native failure context in periodic assertions

* fix: advance active file permissions before queued questions

* fix: finish red-team attempts before retry and cleanup

* fix: finalize plan format captures and judges before retry

* fix: cancel setup-gbrain SDK attempts before fixture cleanup

* test: select periodic consumers of the bounded attempt helper

* fix native Bash permission cards and queued questions

* fix: preserve independent decisions and review scope

Keep CEO approach, engineering scope and outside-review choices from approving independent remedies together. Carry declared contracts through DX polish and resolve new gaps before editing the plan. Regenerate every host and retain existing stop boundaries.

Validation: 654 focused tests passed across nine files; all-host generation passed. Full free and periodic validation pending.

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

* fix: require approval before design plan amendments

Align the Design review philosophy and rating recipe with its section protocol: resolve one proposed fix, then apply only that approved decision and retain honest scores for declined fixes.

Validation: 469 focused tests passed across four files; all-host generation passed.

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

* fix: observe native question completion before transcript persistence

Match owned completion hooks to submitted choices, reject conflicting or late answers, and retain bounded failure evidence.

* test: recognize review posture in acknowledged native questions

Require the selected mode acknowledgement, a completed follow-up question, and its current decoded display while preserving existing posture assertions.

* fix: preserve settled CEO choices and isolate pending remedies

Resolve established approach gates with cited authority and keep independent fixes out of unrelated option commitments and plan amendments.

* fix: carry approved DX choices through later review steps

Choose documentation approaches within the accepted scope and map resolved confusion points without reopening them through a bulk menu.

* test: handle native settings-file edit prompts

Keep one-time owned-file approvals and retain the actual sampled Autoplan permission frame with its matching barrier state.

* test: accept standard CEO reply directives with tuning footers

Recognize the exact trailing preference footer and letter-list directive while preserving current-display and exact acknowledgement checks.

* test: scope split reviewers to their generated plan artifacts

* test: observe native Bash permissions and invocation results

* test: handle owned Bash prompts during mode preference checks

* test: preserve synchronous subprocess rejection in Codex fixture

* Fix periodic review handoff navigation

Recognize review-first and explicit manual-next-step labels while preserving exact action families, manual preference, and ambiguous-menu rejection.

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

* Bind pending file permissions to distinct current targets

Allow one captured file request to own the complete current dialog while unrelated file work is pending. Preserve same-path ambiguity, exact input ownership, and one-time grant checks.

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

* Make paired CEO verification choices genuinely unresolved

Start the positive control with proposed manual checks so its unchanged oracle measures two new coverage decisions. Preserve runtime contracts, targets, count bounds, and all assertions.

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

* Keep CEO review options and verification within approved scope

Audit every offered option for independent add-ons and keep new verification depth pending until accepted. Preserve already requested coverage and trace plan changes to the actual decision.

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

* Assemble DX review artifacts before appending the final report

Keep early DX evidence above decisions, update artifact sections in place, and append the report using the actual current file suffix. Re-read after deleting an existing report before choosing the append anchor.

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

* Keep outside plan reviews exclusive and invocation-owned

Follow one preflight-selected backend, terminate failed Codex work before fallback, and allocate extra prompt/output files uniquely. Consume only the current invocation’s completed output.

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

* Select periodic completion evaluations for report writer changes

Register the shared review resolver for eight missing consumers and regress selection for all nine completion cases without changing their IDs or tiers.

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

* Keep permission ambiguity fixtures on the same normalized target

Use distinct raw spellings of one target in the four negative fixtures so they exercise the normalized duplicate-owner guard after exact current-file disambiguation. Preserve the existing exception, no-input, diagnostic and cleanup assertions.

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

* Clarify preserved contracts in engineering review fixture

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

* Recognize the offered DX follow-up handoff

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

* Check independent commitments before presenting review options

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

* Keep Codex review output and status in one shell invocation

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

* Distinguish seeded plans from reports written by a test attempt

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

* Recover clipped Autoplan file approvals with bounded viewport resizing

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

* Recover clipped Bash approvals before binding the complete command

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

* Isolate setup message tests from the shared checkout

Run the real installer in a temporary payload with private config, require successful completion, and guard source and binary contents and mtimes.

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

* Fix periodic native permission and report completion handling

Match the pinned CLI's soft wraps and clipped headings without granting from incomplete frames. Retire completed file requests, retain mode annotations, and ask section captures for a short final acknowledgement after their full report is saved.

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

* Preserve review approvals and validate DX comparison artifacts

Keep independent remedies and approved amendments explicit. Give the synthetic DX review its existing documentation and validate peer comparison as required analysis alongside four native decisions. Add positive and negative semantic calibrations while preserving review counts, model budgets and prompt size limits.

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

* Make the five-finding CEO fixture's application boundary explicit

Materialize the request adapter and service composition used by the synthetic payment application. Explicitly declare the revised unregistered-event and mail-telemetry assumptions while preserving uncaught handler errors, the original invoice path and all five unresolved findings.

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

* Keep CEO state-path checks scoped to directory preparation

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

* Use checked ports and bounded cleanup in pair-agent tests

Discover the daemon port from its owned state file, retain startup diagnostics, and await failed-start cleanup. Add occupied-port, early-exit, deadline, and foreign-state regressions while preserving the existing HTTP assertions and hook budgets.

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

* Preserve queued edit identity and recover clipped Bash permissions

Distinguish separately queued unfinished edits from mutation of one native tool ID. Keep grants bound to an exact owned request and reject reused IDs, ambiguous inputs, and competing owners.

Support the pinned renderer's literal em dash and request a repaint when only the Bash card's top rule is clipped. Grants still require the complete fresh card and an exact native acknowledgment.

Validation: 413 integrated parser/event tests passed; private repaint controls and joint source review passed. Full canonical suite and native periodic rerun remain pending.

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

* Keep periodic reviews within their approved contracts and deliverables

Carry exact approvals through engineering review, preserve declared contracts when amending CEO plans, and keep prioritization at the requested decision level. Materialize the revised synthetic SDK reference contract while retaining the five original documentation gaps.

Accept the observed semicolon in the finite DX handoff menu and register the direct source dependencies used by the engineering cases. Regenerate canonical review documents without changing model budgets, retries, count bands, or native completion assertions.

Validation: all-host generation and 275 review, fixture, selection and parity tests passed. Full free-suite and native periodic validation remain pending.

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

* Keep Eng approval cadence and independence guards explicit

* Accept ordinary punctuation in manual review handoffs

* Recover file permissions alongside queued Bash calls

* Carry approved DX work through later review findings

* Clarify the synthetic auth internal failure decision

* Bound the periodic DX fixture to onboarding changes

* Recognize native Design review handoff labels

* Hold scope in the integration-choice review fixture

* Carry approved Design decisions through review evidence

* Capture listener state when feedback reload fails

* Exclude workspace caches before checking deprecated flags

* Verify Design UI scope against a seeded review plan

* Clarify plan review decisions and outside-voice approval flow

* Reject setup menus in the Design UI gate

* docs: require focused repair validation before final acceptance

* fix: separate review commitments within existing prompt budgets

* docs: align generation and contributor validation guidance

* fix: advance native review prompts and count acknowledged findings

* chore: bump version and changelog (v1.87.1.0)

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

* chore: enforce cheap checks and side-effect-free validation previews

* fix: handle owned Fetch permissions and oversized native cards

* test: ground review fixtures in independent executable contracts

* fix: preserve review decisions and verify reports before completion

* test: construct the synthetic credential URL without a scanner false positive

* test: materialize DX examples and verify their actual local behavior

* fix: clarify CEO review decisions and execution order

* fix: clarify review workflow ordering and select Design quality checks

* Fix review decision gates and incomplete evaluation fixtures

Persist CEO and engineering commitment ledgers before menus, preserve exact
approvals, and distinguish implementation structure from feature scope.
Route Autoplan through the canonical CEO Step 0 ordering. Classify DX findings
before requesting approval and ground runtime claims in actual evidence.

Complete neutral non-target fixture contracts and accept the captured Design
handoff purpose without relaxing its ownership or acknowledgment checks.
Record runtime-capability verification in AGENTS.md validation discipline.

Validation: 1,335 focused tests passed across 21 files; build, all-host freshness,
skill validation (647 artifacts / 107 tracked), and credential checks passed.
Prior paid failures are preserved; behavioral acceptance remains pending.

* Fix review decision boundaries and owned Read prompts

Preserve exact approvals across review options, compare consistent DX milestones,
and keep proposed implementation separate from review evidence. Bind modern
Read prompts to one immutable native request and wait for its result.

Retain captured regression verdicts, correct fixture error names, improve import
probe diagnostics, and record focused-first validation discipline in AGENTS.md.

* Clarify CEO and engineering review decisions

Use explicit decision steps, one engineering ledger, and clear scope/write transitions. Preserve exact approvals and distinguish pending test requirements. Keep unrelated generated content unchanged.

* Fix review decision ordering and native evaluation interactions

* Clarify engineering decisions and test artifact order

* Clarify pending choices and approvals in CEO reviews

* Make CEO review phases sequential and clarify completion

* Fix Design board submission intent matching

* Seed an existing browser test baseline for Autoplan

* Document decision-log payloads before state initialization

* Preserve exact review scope and decide one change before drafting options

* Require input identity before repeating passing model judges

* Honor permitted storage throughout CEO review completion

* Match complete native permission text within the pinned renderer contract

* Align review approvals, independent choices, and bounded validation

* fix: preserve reopened approvals and declare fixture interfaces

* fix: isolate review artifacts and audit complete questions

* fix: match detector artifact permissions to configured storage

* fix: complete native permissions and review fixture workflows

* fix: order CEO review work and separate engineering guarantees

* fix: preserve native validation and separate review choices

* fix: clarify review decisions and judge complete report context

* fix: constrain review judgments and retain parse failures

* fix: compare each affected value before review decisions

* fix: make engineering review decisions and completion order explicit

* fix: give the complete Autoplan evaluation a bounded chain budget

* fix(cso): diagnose forbidden Docker endpoints before tool lookup

* fix(reviews): reconcile workflow contracts and generated artifacts after main integration

* fix(evals): migrate retained regressions to the native review harness

* fix(tests): close native harness and workflow integration regressions

* fix(evals): preserve complete permission context and native menu contracts

* fix(tests): capture synchronous command output without pipe drain stalls

* fix(reviews): clarify decision and completion ordering

* fix(reviews): separate decision readiness from final completion checks

* refactor(reviews): consolidate decision rules and completion branches

* fix(plan-eng-review): order preparation and clarify decision routing

* fix(plan-eng-review): restore size and question-format guard parity

* fix(plan-eng-review): clarify scope phases and blocked completion

* fix(plan-eng-review): unify review flow and report destination

* fix(plan-eng-review): define bootstrap and question stage ownership

* fix(plan-eng-review): clarify review structure and design lookup

* fix(plan-eng-review): render report examples and show saved decisions

* fix: consolidate Eng review decisions and select their evaluations

* test: cover overlapping terminal attachments and clean merged runner type

* fix: preserve Office Hours relationship closings during review updates

* fix: retain pasted review targets across slash invocations

* docs: preserve validation traces and correct release scope

* test: cover pasted targets in both review skills

* fix: validate report artifacts before recording success

* fix: redact source roots at CSO report boundaries

* fix: bind native Design questions before answering

* test: select report privacy and native recovery regressions

* test: bind rejection predicate in extracted observers

* fix: bind complete boxed native questions

* test: keep the Design UI fixture on native review

* fix: preserve review decisions and evaluation completion outcomes

* fix: clarify CEO approval and report completion order

* fix: align native review evaluation ownership and completion

* fix: bind review evaluators to native decisions and owned artifacts

* fix: validate review decisions against native outcomes

* fix: preserve review evidence and Autoplan phase handoffs

* test: bind review evidence to owned decisions and completion

* fix: retain owned native history across compaction

* fix(evals): validate current review decisions and setup choices

* fix: bind Autoplan reviews and phase completion to current amended input

* fix: reconcile native review evidence and close Autoplan phases

* test: recognize owned whole-candidate complexity decisions

* test: preserve report freshness for approved investigation handoffs

* fix: recognize scoped review findings and isolate dual voice fixtures

* fix: make review handoffs and question dispatch self-contained

* test: recognize complete CEO decisions and procedural pauses

* fix: bind current CEO comparison options and risk intervals

* test: bind engineering decisions and completion to owned evidence

* fix: publish Autoplan phase reports before continuing tools

* test: verify actual Autoplan dual-review dispatch evidence

* test: select dual review when shared evidence fixtures change

* fix: clarify plan review decisions and completion gates

* fix: make CEO review decisions and return paths explicit

* test: keep Autoplan prompt files inside attempt state

* test: preserve source whitespace across permission dialog wraps

* fix: publish Autoplan phase reports before continuing

* test: recognize current CEO comparisons and reject inactive records

* fix: reconcile engineering decision states before completion

* test: recognize complete Design decisions and reports

* test: verify current engineering decisions before navigation

* Recognize source-owned component reduction choices

* fix: recognize current CEO ledger and commitment grids

* test: supply RequestPolicy context to Eng count fixture

* fix: save complete engineering decisions before asking

* fix: bind Autoplan publication to the complete phase readback

* chore: prepare 1.87.5.0 reliability release

* fix: clarify engineering review completion and preserve log failures

* fix: bind CEO saved choices and current section ancestry

* fix(evals): bind review execution and completion evidence

* fix(plan-ceo-review): verify complete decisions before asking

* fix(evals): preserve complete engineering choice records

* fix(evals): preserve complete review outcomes and bounded fixtures

* fix(autoplan): publish phase reports before advancing

* fix(plan-ceo-review): validate option fields before asking

* fix(plan-eng-review): verify current decisions after answers

* fix(evals): bind review decisions and bound fixture scope

* fix(plan-ceo-review): verify decision rows and edit saved checkpoints

* fix(evals): bind review evidence and scope document lookup

* fix(plan-eng-review): update resolution state with its answer

* fix(reviews): preserve complete questions through dispatch

* fix(evals): recognize completed mode declarations

* fix(evals): define cache consistency at wrapper completion

* fix(evals): validate owned initial scope and completed review handoffs

* fix: assemble complete CEO decision fields before saving

* fix: authenticate automatic mode decisions without guessing selectors

* fix: bind engineering coverage to approved regression contracts

* fix(evals): supply review helpers to native Eng capture

* fix(plan-eng-review): preserve the full selected option scope

* fix(evals): recognize owned engineering seed and regression evidence

* fix(evals): bind engineering retry reports to native approvals

* docs: clarify release guarantees (v1.87.5.0)

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

* fix(evals): recognize owned engineering decisions and handoffs

* fix(evals): bind engineering decisions and completion evidence

* fix(tests): align review contracts and selection fixtures

* fix(skills): restore review prompt size limits

* fix(plan-eng-review): clarify review execution and completion

* fix(evals): preserve configured retries through all supervision layers

* Clarify Engineering decisions and report completion

* Keep native decision assertions within their source boundary

* fix: recognize owned engineering decisions and completed navigation

* fix: bind completed auto decisions to their current review

* fix: recognize explicit CEO source attribution

* fix: dispatch verified CEO decisions without recomposing fields

* test: expose existing execution deadlines to review actors

* fix: distinguish CEO decision records from incidental headings

* test: bind split-scope choices to the registered native actor

* test: connect reviewed regressions to required evaluation coverage

* Clarify CEO decision routing and completion stages

* test: expose existing section review deadlines to fixture actors

* test: recognize complete native CEO pacing inventories

* test: exclude answered history from current CEO payloads

* test: detect phase entry through owned skill HOME aliases

* test: validate native review completion and owned report permissions

* fix: make Autoplan close packets carry the parent handoff steps

* test: assess source-bound HOLD decisions within the existing deadline

* fix: keep CEO native decision fields under one formatting authority

* test: register integrated review and permission dependencies

* test: align native review adapters and finding coverage

Preserve explicit AUTO decisions, apply native single-select defaults, and bind complete cropped questions and report permissions to their owned requests. Require seeded review findings instead of crediting setup menus.

Keep captured failure controls and additive selection dependencies. The integrated candidate passed 3,099 focused tests across 65 files; affected paid validation remains required before publication.

* fix(autoplan): require phase reports before advancing

* fix(evals): bind setup and evidence to complete attempts

* fix(evals): bind native answers and pending writes to fixture scope

Preserve complete option rows when native descriptions wrap, retain current
owned Write arguments before journal publication, and keep engineering and
DX answers within their declared fixture interfaces. Add captured free
regressions without increasing model budgets or relaxing completion checks.

* fix(autoplan): verify phase reports across native tool paths

Guard owned methodology reads and reviewer dispatches, detect complete driver
loads through Bash, and distinguish report-only edits from implementation
changes. Follow authenticated native UUID ancestry when journal writes arrive
out of order and verify earlier native content for cached phase reads.

Keep current close acknowledgment and parent publication in order, require CEO
entry before later phases, and register captured failure regressions.

* fix(evals): honor native input and collection lifecycles

Match complete native Edit panes and truncated question borders, reject stderr close before EOF, and stop the CEO split fixture once its acknowledged scope decisions are collected. Keep semantic validation, process failures, report requirements, and absolute deadlines authoritative.

Add captured-event and real-process regressions with selection dependencies. Focused checks pass; final integrated paid and full-suite acceptance remain pending.

* fix(autoplan): retain native session ownership across directory changes

Recover missed native UUID ancestry through the existing strict graph while preserving ordinary event order and legacy scoping. Bind publication hooks to Claude's original project directory while retaining current cwd for requested file paths.

Captured public-event regressions, existing caller checks, and a pinned native CLI loopback verify both fixes. Preserve failed attempts and require fresh paid and final full-suite acceptance.

* docs: align evaluation limits and completion version

* fix(autoplan): allow authenticated phase reads during journal streaming

* fix(evals): bind clipped native questions and owned edit dialogs

* fix: preserve overlay retries and bounded cleanup

* fix: recognize owned planning preludes in native questions

* docs: explain overlay scheduling and cleanup guarantees

* fix: require fresh publication after Autoplan phase reruns

* Release gstack 1.87.6

* fix: preserve CI paths, process identity, and test deadlines

* fix: keep informational setup commands independent of install probes

* fix: clarify plan review decisions and bound source audit reports

* Fix remaining Windows identity and native path CI failures

* Clarify CEO review decision and reviewer-result routing

* test: accept no-install planner in retry supervision

* fix(ceo-review): make review decisions and report completion explicit

* perf(test): add fast PR gates, input-keyed judge reuse and isolated free shards

* fix(test): start isolated CEO smoke from its existing project plan

* fix(test): repair CI fixture races and preserve retry evidence

* fix(ceo-review): clarify approvals, depth and saved completion

---------

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

docs/ADDING_A_HOST.md

@@ -122,16 +122,15 @@ Add `.myhost/` to `.gitignore` (generated skill docs are gitignored).
 # Generate skill docs for the new host
 bun run gen:skill-docs --host myhost
 
-# Verify output exists and has no .claude/skills leakage
+# Verify output exists
 ls .myhost/skills/gstack-*/SKILL.md
-grep -r ".claude/skills" .myhost/skills/ | head -5
-# (should be empty)
 
 # Generate for all hosts (includes the new one)
 bun run gen:skill-docs --host all
 
-# Health dashboard shows the new host
+# Validate all host content and tracked output freshness, including the new host
 bun run skill:check
+# Claude install paths are rejected in prose; legitimate Bash fallbacks are allowed.
 ```
 
 ### 5. Run tests

docs/OVERLAY_BENCHMARK_CONTRACT.md

@@ -0,0 +1,104 @@
+# Overlay benchmark behavior contract v2
+
+The paid overlay fixtures verify supported behavior, with comparative efficacy
+reported separately as research evidence. A passing case does **not** establish
+marginal overlay benefit or resource non-regression. This replaces the previous
+blocking efficacy contract; historical failures retain their original verdicts.
+
+The contract is defined by `OVERLAY_CONTRACT` in
+`test/helpers/overlay-case-policy.ts`. Every new measurement and aggregate record
+identifies version 2. Eval entry names include `contract-v2` so automated result
+matching cannot silently compare the old and new contracts by the same name.
+Selection IDs for the six retained paid cases stay unchanged.
+
+## Blocking checks
+
+Both arms use the real Claude Code preset; ON appends the resolved overlay and
+OFF does not. Each retained fixture runs ten samples per arm, separately for
+`claude-opus-4-7` and `claude-sonnet-4-6`. Each model keeps its own results.
+
+Every planned sample in both arms must execute successfully, produce exactly
+one successful native terminal result without an error flag, preserve workspace
+scope, and yield a finite metric within the fixture's bounds. Missing reasoning
+metadata, failed output validation, oracle process errors, partial sampling,
+deadlines, cleanup failures, or recording failures keep the case failed. Missing
+observations are never replaced with zero or discarded to obtain a passing arm.
+
+| Fixture family | Exact task and scope contract | Additional ON requirement | Measurement |
+| --- | --- | --- | --- |
+| `claude-dedicated-tools-vs-bash` | Read-only workspace; native final JSON maps all five `src/` paths to their exact exported symbol names | Zero Bash calls in every ON sample | Total Bash tool calls |
+| `opus-4-7-effort-match-trivial` | Read-only workspace; native final JSON contains the exact version string | Correct answer in every ON sample | Native `usage.output_tokens_details.thinking_tokens` |
+| `opus-4-7-literal-interpretation` | Only `src/auth.ts`, `src/billing.ts`, and `src/notifications.ts` may change; public tests stay frozen; independent behavior checks run outside the writable fixture | All three target behaviors pass in every ON sample | Correct target behaviors, 0..3 |
+
+Each family also has a `-sonnet` case. Output validation checks the native final
+answer, so a correct word mentioned earlier cannot conceal a wrong final answer.
+The read-only prompts explicitly request the JSON shape. These exact-answer
+prompts are new in v2: compare ON/OFF within the same contract and fixture, not
+across versions. In particular, never compare the current literal correctness
+metric to historical counts of edited files.
+
+OFF literal samples can validly complete 0..3 behaviors; that is the measured
+control variable. OFF dedicated-tool samples can validly use Bash. ON must meet
+the exact requirements above even when its comparison shows a large improvement.
+Scope is checked again after the implementation oracle because importing a
+repaired module can itself change files. Each oracle must emit its unique
+completion marker after its assertions; an implementation exiting zero during
+import cannot masquerade as a passing behavior.
+
+## Comparative research results
+
+The original numeric comparators remain unchanged: at least 20% fewer Bash calls
+or reported reasoning tokens, and at least 20% more correct literal targets.
+Their `criterionMet` values, means, and raw per-arm metrics remain in aggregate
+artifacts. They do not decide the v2 behavior verdict.
+
+| Comparison status | Meaning |
+| --- | --- |
+| `improved` | The original numeric criterion was met |
+| `baseline_saturated` | Every OFF sample reached the metric's optimum; no demonstrated marginal benefit |
+| `no_measured_improvement` | Complete measurements did not meet the original criterion |
+| `regressed` | ON's mean was worse in the metric's declared direction |
+| `unsupported_hypothesis` | The resolved overlay does not contain the claimed intervention |
+| `incomplete` | Missing or invalid measurements; the behavior case also fails |
+
+Equality at the optimum is behavior success with `criterionMet: false`, never a
+percentage improvement. An effort case can pass correctness while its measured
+reasoning-token comparison is `regressed`; that regression remains visible.
+The product asks models to match effort to complexity and does not promise a
+universal 20% token reduction. There is no calibrated resource ceiling,
+non-inferiority margin, or powered sampling scheme in this contract. Any future
+resource regression gate needs a separately agreed practical margin, baseline
+distribution, sampling design, and decision rule before collecting new results.
+
+## Retired fanout cases
+
+The current resolved `model-overlays/opus-4-7.md` and its inherited `claude.md`
+contain no fanout instruction. The paid cases `opus-4-7-fanout-toy`,
+`opus-4-7-fanout-realistic`, and their `-sonnet` variants are retired, including
+their paid wrappers and selection entries. They are not renamed successes.
+Reintroducing a paid batching efficacy experiment requires an actual product
+claim and a discriminating experiment.
+
+Free regressions retain native first-message fragment grouping, tool-ID
+deduplication, and the original fanout comparator. This metric measures message
+batching and cannot establish wall-clock execution concurrency.
+
+## Evidence and execution
+
+Run directories and Bun/native attempt identities remain unique. Artifact files
+use exclusive creation; retry attempts cannot overwrite earlier raw streams,
+measurement records, workspace snapshots, or aggregate verdicts. Native events
+are saved before yielding them, including streams that later throw. A validated
+measurement records once; a separate aggregate records the case verdict with
+zero additive cost and time. Failed writes retain partial evidence and fail the
+case. Versioning never rewrites historical artifacts.
+
+The paid runner executes one overlay wrapper per shard and one wrapper at a
+time, disables Bun retries, and preserves native rate-limit retry evidence. At
+finalization, workers must settle before workspace cleanup begins. Both share
+the existing five-second grace; expiry fails the case and prevents late workers
+from starting cleanup. Each fixture keeps its original turn cap, concurrency,
+30-minute work budget, and recording grace. The six
+wrappers remain periodic-tier cases; free tests exercise deterministic outcomes
+without making model calls. A free pass validates the instrument, not live model
+behavior or a positive efficacy claim.

docs/TESTING_INTERNALS.md

@@ -64,6 +64,11 @@ fallback `~/.gstack-dev/evals/`) with auto-comparison
 against the previous finalized run (in-flight `_partial` files are never used as
 a baseline, so a run can't compare against itself).
 
+The periodic overlay fixtures use a versioned behavior gate with efficacy
+reported separately. See [Overlay benchmark contract v2](OVERLAY_BENCHMARK_CONTRACT.md)
+for exact correctness requirements, retired fanout cases, immutable evidence,
+and the limits of a passing result.
+
 ## Runners: how the suites execute (2026-08 overhaul)
 
 **Aside-only E2E tests self-skip without a live Aside; browser-driving tests
@@ -123,8 +128,10 @@ DURATIONS (LPT, `packShardsByDuration`) when the committed seed
 `bun run test:free --record-durations` (each file timed in its own child;
 CI never records). Missing seed → silent hash-shard fallback; corrupt seed →
 one warning + fallback; unknown files get 75th-percentile pessimism. Packed
-shards get duration-aware walls (`max(base, predicted × 3)`); the `--shard`
-CI-matrix path keeps stable hash indices untouched. `TREE_MUTATING` is EMPTY:
+shards get duration-aware walls (`max(base, predicted × 3, files × 5s)`). The
+legacy `--shards N --shard i` path keeps stable hash indices. Required CI uses
+one duration-packed `--ci-plan`, 20 isolated `--ci-run` machines, and a
+`--ci-verify` aggregate. `TREE_MUTATING` is EMPTY:
 `gen-skill-docs.ts` has a `main()` guard (imports never regenerate; pinned by
 `test/gen-skill-docs-import-purity.test.ts`) and `--out-dir` renders every
 host, so all former mutators render into mkdtemps and the trailing serial
@@ -144,17 +151,16 @@ total-run deadlines. `test/pty-output-wake.test.ts` covers output, silent waits,
 exit, close, missing readiness, split redraws, and continuous redraws. Close and output waits
 cancel their losing deadlines so completed workers can exit immediately.
 
-The UI-positive design gate preloads `PLAN.md` and counts only positively
-identified, answered native design questions. Setup and outside-review choices
-cannot trip its one-question ceiling. Its final proof uses the full native
-question, not the truncated diagnostic snippet. Unknown-command failures must
-name the invoked slash command; a child tool rejecting `--help` is not a skill
-registration failure.
-Its gate-specific classifier also recognizes answered, untagged UI issues with
-concrete numbered design choices. The isolated fixture owns the target; repeated
-filenames, pass labels, question verbs, and option punctuation are not required.
-Explicit wrong-plan context and workflow menus are rejected; periodic
-seeded-finding classifiers are unchanged.
+The UI-positive design gate commits the supplied plan as `review-input.md`
+before the first model turn. It requires an acknowledged Design focus/rating
+question followed by an answered native finding about concrete UI behavior;
+setup and outside-review choices cannot satisfy that sequence. Its declared
+board actor submits feedback before acknowledging it. The final proof uses
+the full native question, not the truncated diagnostic snippet, and proposal
+text mentioning "no UI scope" is not treated as an exit verdict. Unknown-command
+failures must name the invoked slash command; a child tool rejecting `--help`
+is not a skill registration failure. Periodic seeded-finding classifiers are
+unchanged.
 
 **Paid suite (sharded runner, local AND CI).** `scripts/test-paid-shards.ts`
 is the single selection engine: 1 file per shard, `EVALS_JOBS` shard
@@ -177,6 +183,41 @@ key must name a living paid test (`test/touchfiles.test.ts`'s reverse
 invariant), and `git show <sha>:path` fixtures are banned — vendor the bytes
 instead (`test/git-ref-fixture-tripwire.test.ts`).
 
+**Fast PR profile and evidence reuse.** `test:pr` selects the changed cases in
+`scripts/test-pr-profile.ts` plus every changed quality judge. `--profile full`
+retains the broad census; no case IDs or tier assignments are removed. The plan
+records both E2E and judge selections and lists deferred coverage. Executors
+receive those selections separately and the report checks actual executed case
+counts. Unknown source dependencies restore the broad gate; unmapped prompts
+without registered coverage require full validation. Known broad-only prompt
+changes are explicitly deferred, not counted as PR passes. Weekly/manual runs of
+`evals-periodic.yml` execute both complete censuses fresh; manual `evals.yml`
+runs the broad gate. This deliberately moves some defect detection later.
+
+`scripts/eval-input-cache.ts` accepts only complete, clean, first-attempt passes
+with matching before/after inputs. The audited workflow-judge adapter hashes the
+actual expanded prompt, source/fixture/rubric/runner closure, installed SDK,
+model parameters and runtime. Missing/unknown inputs force execution. Receipts
+are scoped to the same repository and PR, expire after 24 hours, and contain
+public scores and provenance rather than prompts or secrets. Only the 14 cases
+using `runWorkflowJudge` are eligible; the other 11 quality cases remain fresh.
+CI supplies the scoped cache/runtime configuration; local runs are fresh by
+default. Cached scores must
+pass current assertions; reused records retain their original source and time
+and cannot renew the receipt. Dynamic live-agent runs are currently ineligible.
+`EVALS_FRESH=1`, periodic and release validation bypass both lookup and publishing.
+
+**Free test timing and isolation.** `test:quick` is an explicitly partial measured
+subset for edit feedback. `test` remains complete local acceptance with its
+bounded worker pool. The CI planner inventories every free file and packs them
+using `scripts/free-test-durations.json`; each machine runs its assigned shard
+serially. Plans and receipts bind the source revision, complete inventory and
+strict outcomes; missing, duplicate or mismatched receipts fail the required
+aggregate. The existing maximum five-file flaky retry allowance applies across
+the entire lane, not separately to every machine. Refresh the full timing list
+with `bun run test:free --record-durations`. Profiling records failures faithfully
+and is separate from final release acceptance.
+
 **CI planner/executor/report.** `--emit-plan <path> --slices K` computes
 selection + the slice plan ONCE (killing per-slice selector divergence);
 `--plan <path> --slice i` executors consume the manifest and write
@@ -206,7 +247,7 @@ archaeology.
 `test/helpers/eval-budgets.ts` (JUDGE/CAPTURE/CAPTURE_LONG/PTY/PTY_LONG);
 `test/eval-budgets-policy.test.ts` pins that every tier fits the shard wall
 minus overhead and ratchets raw literals. Budget above the wall is fiction.
-The sole registered exception is `AUTOPLAN_CHAIN_BUDGET` for
+The registered four-phase exception is `AUTOPLAN_CHAIN_BUDGET` for
 `test/skill-e2e-autoplan-chain.test.ts`: 80 minutes of work (four `PTY_LONG`
 allocations), an 84-minute session watchdog, an 85-minute Bun test deadline,
 and a 172-minute supervised shard wall. The unchanged retry count of one
@@ -224,24 +265,47 @@ A rejected artifact edit fails the test instead of falling through to terminal
 permission input. Approval itself supplies no edit success or phase credit:
 the native tool result and all four completed review phases are still required.
 
+`FINDING_RETRY_BUDGETS` also registers six finding files. Each retains its
+25-minute case deadline and one retry: the two-case CEO finding-count file has
+a 102-minute shard wall, and the five single-case files have 52-minute walls,
+including two minutes for cleanup. No per-case budget grows. Overlay wrappers
+have a 1,830-second minimum shard wall and run without Bun retries; see the
+[overlay contract](OVERLAY_BENCHMARK_CONTRACT.md) for their unchanged work budget.
+
+The quality file reserves 6,400 seconds for all 25 cases and their existing
+retry, plus cleanup. Each still has 120 seconds of model work. Its 14 workflow
+judges own their deadline and abort signal, with five seconds for terminal
+recording inside a ten-second Bun grace; the other 11 retain their existing
+120-second Bun timeout. Late responses cannot create records or cache passes.
+
 `resolvePaidShardBudget(files, overrideMs?)` is the canonical per-job resolver.
-Only the exact Autoplan file gets the exception, in its own shard. An explicit
-CLI `--timeout`, `EVALS_SHARD_TIMEOUT_MS`, or API `timeoutMs` still wins, including
-a lower cap. Planner entries and execution results record the effective wall,
+Autoplan, each registered finding file, and each overlay wrapper require their
+own shard, even with `--files-per-shard` above one. Mixed or multi-file overlay
+jobs are rejected so ordinary files retain their configured retries. An explicit
+CLI `--timeout`, `EVALS_SHARD_TIMEOUT_MS`, or API `timeoutMs` still wins for these
+policies, including a lower cap; overlay overrides below their minimum are rejected.
+Planner entries and execution results record the effective wall,
 its source and policy identifier. Custom drivers must resolve each job instead
 of passing their ordinary 1800-second default as an explicit Autoplan cap;
 their outer controller/detach wall must also cover the allocated work and cleanup.
-`eval:bg:periodic` already has a 37800-second outer cap. Legacy monolithic
+`eval:bg:pr` and `eval:bg:periodic` have 72000/66000-second outer caps; the PR
+wrapper covers a full-gate fallback at its default two workers. The broad gate
+wrapper reserves 33600 seconds, and release reserves 100000 seconds for both
+tiers. Legacy monolithic
 `eval:bg`/`eval:bg:all` retain their shorter 5400/7200-second caps and do not
 promise two complete Autoplan attempts; use the sharded periodic path for this policy.
 
-Periodic CI plans `--slices 7 --autoplan-slice`: six ordinary slices retain their
-existing limits, while the seventh runs only Autoplan. Its unchanged 200-minute
-job cap leaves 28 minutes around the 172-minute shard for setup and artifacts.
-Reconciliation rejects missing, duplicated or misplaced Autoplan work and absent
-budget records. This does not claim that the growing ordinary census has a
-200-minute worst-case bound. Ordinary paid tiers and their 1800-second shard
-wall remain unchanged; unregistered over-ceiling tests still fail policy checks.
+Periodic CI plans `--slices 8 --autoplan-slice`: the eighth runs only Autoplan.
+When overlays are selected, the seventh is reserved for their serial wrappers;
+registered finding files are distributed across the remaining ordinary slices
+by their supervised walls. Each slice job has a 355-minute cap; Autoplan retains
+its 172-minute shard wall. Reconciliation rejects missing, duplicated or misplaced
+registered work and absent budget records. The weekly gate census has a
+350-minute cap and PR slices have a 220-minute cap. Free supervision tests
+verify these bounds against the complete current census, configured retries,
+and setup reserve. Ordinary paid tiers and the default 1800-second
+shard wall remain unchanged; the registered and overlay policies above supply
+exceptions, and unregistered over-ceiling tests still fail policy checks.
 
 Session timeouts are two-phase: a silent API dies at the startup grace (90s
 local / 300s CI floor, distinct exit reason `timeout_startup`) and the work

docs/askuserquestion-split.md

@@ -103,14 +103,22 @@ parent decision, resume from where they left off.
 After the chain resolves (without Hold), fire `D<N>.final` to confirm
 and validate the assembled set.
 
-**Step 1 — validate dependencies.** If the picked set is incoherent
-(e.g. E3 picked Include but its required E1 was Cut), do NOT silently
-accept. Re-prompt the conflict as a single AskUserQuestion:
+**Step 1 — validate dependencies and capacity.** If the picked set is
+incoherent (e.g. E3 is Included but its required E1 was Cut), show the
+conflict and the actual prior answers. Ask which candidate to revisit:
 
-> "E3 needs E1 but you cut E1. Revise:
-> A) keep E1
-> B) cut E3 too
-> C) leave as-is and accept the broken state"
+> "E3 needs E1. Your answers include E3 and cut E1. Which should we revisit?
+> A) Revisit E1
+> B) Revisit E3
+> C) Hold — leave this conflict unresolved for discussion"
+
+This routing answer changes no disposition. For the named candidate,
+fire one `D<N>.revise-<k>` with the standard **Include / Defer / Cut / Hold**
+menu. Hold all other candidates' prior answers fixed. Revalidate dependencies
+and capacity after the answer; never silently cut, swap or include another
+candidate to make the set fit. A Hold stops the chain: retain the actual
+prior answers and explicitly report the unresolved blocking conflict.
+Do not confirm an incoherent set as ready to implement.
 
 **Step 2 — confirm the assembled set.** If coherent: