Repository navigation
Stop the optional Stop hook checking for a test after the last edit (GST-2811) - #4428
Merged
Merged
Conversation
…GST-2811) Models already run a test after their last edit when /verify asks (Opus 5.5 96%, GPT-6.1 Sol 89%), and the hook's check added no measurable gain. Remove the check_after_last_edit contract item and its shell-write parser, keep the rule as /verify prose, and stop /verify telling models never to edit CHANGELOG. Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
|
😎 This pull request was merged. |
…o the pre-merge check accepts its removal Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
…em stops being checked Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This was referenced Nov 17, 2026
Merged
capy-ai Bot
added a commit
that referenced
this pull request
Nov 17, 2026
- /autoplan's approval page no longer hides a reviewer's product objection behind "keep your direction". (#4431) - Every skill that loads the shared preamble, on every host, now says to ask a question once and to check a file before handing it over. (#4383) - The Captain manuals now name the `Failed or skipped:` and `Also changed:` lines that READY and lane reports carry. (#4432) - gstack's shell locks hold again on Ubuntu 26.04, where `mkdir` reports success when another process won the race (GST-2247). (#4262) - `test/gstack-lock-race.test.ts` runs the old pattern and the helper against a PATH shim that behaves like uutils 0.8.0 `mkdir`, so it runs on any runner image: the old pattern lets two of four racers win, the helper admits exactly one in each of 60 races and never overlaps across take, hold and release loops, and parallel `gstack-config` writers keep every key. (#4262) - Windows Setup E2E now runs when `bin/gstack-lock.sh` changes. (#4433) - Claude Code's skill list now says when to use each gstack skill, so Opus loads /review and /plan-eng-review when a request fits. (#4387) - A pull request's merge head can no longer keep its `/ship` receipt without a full review. (#4365) - `/ship` now counts review rounds per pull request, and the head that merges always gets a full review. (#4360) - /ship's documentation audit no longer accepts a passing result when an earlier line from the same audit said "blocked". (#4434) - /ship's documentation audit no longer spends an attempt when the audit child writes text after its result. (#4285) - The README lists the two new guides, and the install doc counts its hooks right. (#4403) - docs/releases.md now says the stable release cut treats a failed commit status as red and a pending one as still going, as it has since #4257. (#4436) - /ship's `finish --docs-requested` now accepts a docs fix when the audit named the file with a line number. (#4437) - Document the `.gstack/approval-classes.json` schema, `gstack-approval-classes` CLI commands, and owner approval citation format in `docs/ship-policy.md#approval-classes`. (#4355) - `/eval-plan` now runs four result checks instead of leaving them to judgment. (#4384) - The /review coverage-audit eval no longer fails a run that really read the source file when zsh stops a command line at an `echo ====` separator. (#4364) - Sixteen in-repo documentation links reach their section again. (#4394) - gstack agents now open every status report with what failed, what they didn't verify and what else they changed. (#4273) - Two new short guides. (#4380) - A skill you wrote whose contract still lists `check_after_last_edit` keeps the rest of its contract. (#4439) - The ship policy guide now explains how to merge any pull request with `gstack-ship-queue land`, even in a repository with one maintainer. (#4377) - /ship's test step runs a large unit suite on Linux. (#4426) - Every skill that loads the shared preamble, on every host, now says never to print a whole environment, git config or process table. (#4382) - Pull requests from /ship now open with at most three plain lines, and every merge lands as one commit titled with the PR title. (#4438) - `gstack-pr-wait`, `gstack-ci-gate` and the release cut no longer call a commit green while a check on it failed. (#4257) - `gstack-runaway` now says what a thread's cold and empty wakes cost, not just how many there were (GST-2770). (#4385) - The optional Stop hook no longer checks for a test run after the last edit, and /verify asks for it in plain words instead. (#4428) - /ship's test step says when it ran out of time, and you can set its budget. (#4429) - The remote test command no longer puts every file on one command line. (#4429) - The stable release no longer waits on the nightly Windows tests (GST-2815). (#4400) - Ubicloud fan-outs can no longer overshoot the vCPU limit, and owners no longer delete each other's VMs. (#4430) - hono is pinned to 4.13.16, which fixes five advisories the weekly OSV scan found in 4.13.13 (gstack#4406). (#4408) - `package.json` overrides (#4408) - `test/dependency-security.test.ts`: the lockfile floor check now covers `hono >= 4.13.16`. (#4408) - The nightly Windows test run finishes again. (#4409) - A short write-up checklist in ETHOS.md. (#4381) Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The optional Stop hook (
./setup --skill-contracts) no longer stops a run for missing a test after the last edit; /verify asks for that test in its own words instead. This keeps gstack to the steps that measurably help a model finish real work.What changed
/verify now asks for the full check after the last edit and no longer forbids a changelog entry the repo expects.
Before/after as text: /verify now asks for the full check after the last edit and no longer forbids a changelog entry the repo expects
Why
Garry approved cutting the hook's "run a test after your last edit" check because it showed no measured gain. With /verify loaded, Opus 5.5 ran that test in 96% of runs and GPT-6.1 Sol in 89% without the hook, and the hook added between -0.5 and +2 points, with every interval crossing zero (gstack-evals#356). The same eval found Sol, told to use /verify, wrote the changelog entry an upstream fix carried in 0 of 58 tasks against 32 of 58 plain, so /verify stops saying it never edits CHANGELOG.
Blast radius
The change touches the contract checker, the "Done means" renderer, the routing doc and /verify. No shipped skill declares the removed item, because a repo-wide search at the base found it only in the checker, the renderer and the doc example. A custom skill that still lists it now reads as an invalid contract. A smoke run showed the hook letting the stop through with one line naming the file (script ran).
Summary
People who turned on skill contracts got a block when a model skipped a test after its last edit, but models already run that test when /verify asks. The checker loses that item kind and the shell-write parser only it used. /verify keeps the rule as prose and now treats a changelog entry the repo expects as part of the change. The pre-merge check maps the deleted parser to the contract checker's test so it accepts the removal.
Test Coverage
All changed code paths have test coverage: each of the 4 changed runtime files has a changed test that names it. The contract tests now assert that the removed kind is rejected as an unknown kind.
Pre-Landing Review
No issues found.
Gate edits: 21 listed, 21 read, 0 findings. The two RH-1 hunks remove tests of the deleted check and parser, whose subject code this diff deletes under Garry's recorded approval.
Outside review: openai/gpt-6.1-sol, completed, clean.
Review notes
Below the blocking floor (advisory, informational, test-gap or wording items); recorded here and not fixed in this PR.
lib/skill-contract.ts:25A custom contract that still lists the removed check_after_last_edit kind is invalid as a whole, so the optional Stop hook stops checking that skill's other items too, not only the removed one. The changelog entry now says so.verify/SKILL.md.tmpl:31The new changelog sentence could lead a model to edit CHANGELOG.md directly in a repo that expects a fragment; the sentence already scopes it to what the repo expects.docs/skill-routing-and-contracts.md:79The reader did not check the cited figures against gstack-evals#356; they match that PR's description (96.0%, 89.0%, -0.5 to +2.0 points)..gstack/pregate.json:19The dependencies entry for the deleted lib/shell-writes.ts stays after the deletion lands, as the QM review script's entry did (QM edition no longer ships the review skill (GST-2306) #4279).Exploratory QA
Functional smoke on the
gstack-skill-contractCLI and the Stop hook, 3 probes within the 300 s budget, all passing on the current tree:checkon a contract skill that ended without its report: exit 1, onlydebug-reportmissing.checkon a custom skill that listscheck_after_last_edit: exit 5,unknown kind "check_after_last_edit"naming the SKILL.md.SKILL_CONTRACT_STATE_ERRORline naming the file.Design Review
No frontend files changed — design review skipped.
Eval Results
No paid eval ran. The cut rests on gstack-evals#356; the changed /verify prose is covered by the free parity, prompt-size and freshness checks.
TODOS
No TODO items completed in this PR.
Documentation
Status: current — no documentation edits were needed for removing the Stop hook's check_after_last_edit item.
Scope: base 18652cc..head e769afd on remove-test-after-edit-stop-hook-check (14 committed files plus 1 unstaged changelog fragment), with a repo-wide search for the removed kind, shell-writes parser and example item ids.
Documentation health:
Coverage debt: none. Diagram drift: none.
Test plan
bun run test(free suite): carried from CI run 44062410556 at this tree, 1672 test files, 0 failuresbun teston the contract, Stop hook, descriptions, route, parity, module-size and budget suites; typecheck and the test typecheck ratchet;gen:skill-docsfreshness on both hostsAgent notes
Refs GST-2811. Thread GSTA-574 for Captain GStack Minder. gstack: ok 1.99.5.0 for=ship.
Approvals
Approved: yes (record r-68785efeae4d). The full approval record stays private.
Plan: exempt: audited-list (GST-2811: a measured cut Garry approved; the issue carries the diagnosis)
Plan check: exempt audited-list
REVIEW_LOOP: passes=1 blocking_found=0 deferred=0 noted=0 fixed=0 scope=frozen dispatches=1
Gate edits: 21 listed, 21 read, 0 findings