ci: restore extended gate, integration tests, and quality hardening - #31
Merged
Conversation
Re-add GitHub Actions extended suite on PR/push, add test-runner contract and Cursor manifest checks, and link JS/TS README evidence to recapture-15 drafts. Co-authored-by: Cursor <cursoragent@cursor.com>
Defer fixture src/test exploration during brainstorming when TDD applies, and treat late switches after brainstorming context reads as TDD failures. Co-authored-by: Cursor <cursoragent@cursor.com>
Wire brainstorming and writing-plans to use their reviewer prompt templates after inline self-review instead of stopping at checklist-only review. Co-authored-by: Cursor <cursoragent@cursor.com>
Add publication and Cursor eval harness isolation plans; remove intermediate Aug 18–30 js-ts recapture duplicates superseded by later tracked plans. Co-authored-by: Cursor <cursoragent@cursor.com>
Document eval harness setup with a three-tier automation model and validate language-guidance scenario manifests weekly without API credentials. Co-authored-by: Cursor <cursoragent@cursor.com>
Link TypeScript and JavaScript rows to recapture-15 drafts across localized READMEs and clarify that publication eval gates remain pending. Co-authored-by: Cursor <cursoragent@cursor.com>
Install ripgrep in GitHub Actions, pin evals-static harness SHA, add subagent-less review fallbacks, and align writing-plans handoff wording. Co-authored-by: Cursor <cursoragent@cursor.com>
GNU tar rejects BSD --uid/--gid flags used when building Codex tar.gz archives; select --owner/--group on GNU tar and keep BSD flags on macOS. Co-authored-by: Cursor <cursoragent@cursor.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
1 participant
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.
Who is submitting this PR? (required)
What problem are you trying to solve?
The repository had no GitHub Actions CI after the test workflow was removed (
2026-08-12-1239-remove-ci-workflow.md), so PR quality depended entirely on contributors running tests locally. The first CI run on this PR failed becauseubuntu-latestlacksrg, which several core tests require. Cursor had no deterministic plugin wiring tests in the core suite. README language-pack evidence for JavaScript/TypeScript did not match recapture-15 eval drafts. Brainstorming context exploration could precede TDDtesting.mdreads (recapture-15 blocking failure). Spec/plan reviewer templates existed but were not wired into skills.Why restore CI now?
The August removal plan preserved local
npm test/npm run test:extendedbut dropped automation when maintainers wanted fewer moving parts. Since then the extended suite (brainstorm-server + Antigravity) runs cleanly locally, and this PR adds a test-runner contract so suite drift is caught statically. Restoring extended CI reinstates the gate described in2026-08-12-1209-ci-extended-gate.mdwithrginstalled on the runner.What does this PR change?
Restores extended CI on PR/push (with ripgrep), adds test-runner contract and Cursor manifest tests, documents tiered eval automation with pinned static manifest workflow, tracks eval infrastructure plans, tightens TDD/brainstorming fixture boundaries for recapture-16 prep, wires subagent document review loops (with no-subagent fallback), and aligns README evidence links across locales.
Is this change appropriate for the core library?
Yes. Core plugin infrastructure, skill behavior constraints, and documentation — not a standalone plugin. No new runtime dependencies. Skill wording changes are minimal hard-constraint additions; publication is not claimed. Recapture-16 behavioral matrix is follow-up work.
What alternatives did you consider?
rgwith grep in tests: rejected for this PR — installrgin CI to match local dev and existing test scripts.Does this PR contain multiple unrelated changes?
Related hardening pass (CI, eval honesty, recapture-16 prep). Skill changes remain in this PR but without claiming publication readiness; maintainer may still prefer splitting skill commits — human partner accepts that risk for now.
Existing PRs
Environment tested
npm test+npm run test:extendedpassNew harness support
N/A — deterministic Cursor plugin wiring tests only.
Evaluation
npm test,npm run test:extendedpass locally after each commit.Language-pack evidence
Rigor
writing-skillsadversarial eval for skill changes (deferred to recapture-16 PR)Human review