Repository navigation
Conversation
…arity Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Evals: passedSuite: copilot-smoke Lint: validated (no model execution) - Evidence Run: passed - 1 trial - Evidence
Quality & token summary
|
There was a problem hiding this comment.
🔵 Needs a closer look
The Vally 0.20.0 upgrade affects coupled adapter/CLI contracts for suite selection, grading thresholds, and JSONL evidence parsing, requiring human review of model-free native acceptance.
No open issues 🎉
1 file not reviewed
pnpm-lock.yaml: Generated file
What changed in this PR
This maintenance pass updates dependencies and CI caching for the thin Vally GitHub Action adapter, while keeping its development tooling aligned with Node 24.
Changes:
- Add lockfile-keyed pnpm store caching to CI.
- Upgrade the Vally test fixture to 0.20.0, refresh tooling and transitive dependencies, and synchronize smoke-test pins and documentation.
- Derive the native test fixture version from its installed manifest and refresh bundled dependency notices.
| File | Description |
|---|---|
tests/native.test.ts |
Read the fixture’s Vally version from its installed package manifest. |
README.md |
Update the Vally pin example and development fixture documentation. |
pnpm-lock.yaml |
Lock updated direct and transitive dependencies. |
package.json |
Bump Vally CLI, Node 24 types, and oxlint. |
dist/THIRD-PARTY-NOTICES.txt |
Refresh bundled dependency version notices. |
dist/bundled-dependencies.json |
Update bundled dependency inventory versions. |
.github/workflows/ci.yml |
Add pnpm caching and update native smoke tests to Vally 0.20.0. |
Review effort: Balanced (auto)
wbreza
left a comment
There was a problem hiding this comment.
Approved at c17d061. No actionable findings or outstanding fixes were identified in github/evals-action#27.
The existing model-free tests cover suite selection, weighted scoring, native result parsing, and failed or skipped runs. CI for this head passed those tests, the packaged smoke checks, and the bundle rebuild comparison. I did not run PR scripts locally; generated bundle consistency was verified through that CI comparison.
Generated by 🌳Breza Bot
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
There was a problem hiding this comment.
Approval remains at 577d3a5 for github/evals-action#27. No substantiated code findings remain from this review.
I assessed the new cache-output concern in its existing thread. The proposed attack requires control over code this job already executes, so I could not establish an additional security risk. That thread remains open for maintainer review; I did not resolve or dismiss it.
CI for this head passed the model-free native contract tests, packaged smoke checks and bundle verification. I inspected the results without running PR scripts locally.
Generated by 🌳Breza Bot
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
wbreza
left a comment
There was a problem hiding this comment.
Approved at c6e32ac for github/evals-action#27. No outstanding actionable findings remain.
The cache-output formatting request is addressed in the code: the path is now written using a random delimiter and quoted printf arguments. CI for this head confirms the cache receives the expected pnpm directory and restores it successfully; all listed checks pass. I inspected the code and saved CI evidence without running PR scripts locally.
Generated by 🌳Breza Bot
There was a problem hiding this comment.
🔵 Needs a closer look
The pass combines a Vally CLI/runtime upgrade, transitive and generated bundle changes, and new hosted-runner caching, whose Linux compatibility and cross-workflow cache behavior require validation on actual GitHub runners.
No open issues 🎉
1 file not reviewed
pnpm-lock.yaml: Generated file
Review effort: Lite (auto)
Summary
Consolidated
repo-maintainpass across governance, workflow caching, vetted dependency updates (with >= 7-day release cooldown and MIT/permissive license verification), code/configuration checks, test coverage gates, and strict Vally 0.21.0 skill/eval lint checks.Phase Breakdown
94d4c07pnpmstore caching (actions/cache@caa296126883cff596d87d8935842f9db880ef25) in.github/workflows/ci.ymlwhile preservingnode24/.node-version(24.21.0) parity andactions/checkout@v7.c17d061,577d3a5@microsoft/vally-cli(0.19.0->0.21.0) with mechanical companion syncs (tests/native.test.tsdynamic manifest version lookup,.github/workflows/ci.ymlsmoke pins,README.md),oxlint(1.87.0->1.90.0),@types/node(24.19.0->24.19.2, keeping Node 24 alignment),fast-uri(3.1.8->3.1.9), andfast-xml-parser(5.11.2->5.12.0), and rebuilddist/.96.82%line /91.08%branch coverage (180/180Vitest tests +6/6packaged tests passing; diff coverage100%).vally 0.21.0strict skill + eval lint passing acrossexamples/copilot-smokeandtests/fixtures;0broken local links and100%action.ymlinput/output parity.Repo Health Score:
89/100 (B)->97/100 (A+)(+8)actions/setup-node&actions/cache; lockfile-keyedpnpmstore cache added96.82%lines,91.08%branches,100%diff coverage,0untested critical entrypoints@microsoft/vally-cli@0.21.0; >= 7-day cooldown and permissive SPDX checks passed0secret/CodeQL findings; repository protection settings enabled0broken links;README.mdinputs/outputs matchaction.ymlvally lint --strictclean onexamples/copilot-smoke(skill + eval) andtests/fixtures>24hwithout commentsOpen Dependabot PR Disposition (Left Open Until Merge)
@microsoft/vally-cli0.19.0->0.20.0(superseded by0.21.0, including mechanical companion version syncs)oxlint1.87.0->1.90.0(exact package //directory / version match)fast-uri3.1.8->3.1.9(exact package //directory / version match)@types/node24.19.0->26.6.6(semver-major Node 26 definitions skipped to preservenode24Action runtime alignment; updated@types/nodeto24.19.2instead)Verification
pnpm install --frozen-lockfile --ignore-scripts(zero private registry URLs,tarball:keys, orsha1-hashes inpnpm-lock.yaml)pnpm check(tsc --noEmit,oxlint --deny-warnings src tests,vitest run --coverage,node tooling/build.mjs --check,node --test tooling/packaged.test.mjs)pnpm exec vally lint --strict examples/copilot-smoke --eval-spec=examples/copilot-smokepnpm exec vally lint --strict --eval-spec=tests/fixtures