ci(coverage): gate on patch coverage via Codecov, not global threshold

[JSONbored]

Problem

The repo's 97% coverage gate was a global/project-total threshold in vitest.config.ts. A global ratio is non-compositional: each open PR is measured against a moving baseline, so the first PR to merge consumes the buffer above 97% and knocks every other open PR below the gate — even PRs that changed nothing coverage-relevant. The result is constant churn: merge one PR, watch the rest go red, force everyone to re-sync and re-test.

The current numbers prove it — branches and functions sit at exactly 97.00% (13281/13691 and 3566/3676): zero buffer. A single uncovered branch in any merge failed all in-flight PRs.

Fix

Move the 97% requirement to Codecov patch (changed-lines) coverage, which depends only on a PR's own diff and is immune to what else merges. The bar is unchanged for new code — arguably stricter, since every PR must now test its own changes rather than hiding undertested code behind a global buffer.

File	Change
vitest.config.ts	Emit lcov; drop hard 97% global thresholds to a loose 90% backstop (catastrophe net only — actual coverage is ~97–99%)
codecov.yml (new)	patch target 97% (the real gate); project informational (tracked, non-blocking)
.github/workflows/ci.yml	Upload coverage to Codecov (SHA-pinned codecov-action@v5.5.5); reword gate guidance
CONTRIBUTING.md	Document patch-coverage rule; drop the "aim for 98% to absorb CI variance" band-aid

Verified locally

• actionlint clean
• npm run test:coverage passes; coverage/lcov.info emitted (92 files) at the path the upload step reads
• Global coverage 97–99% — comfortably above the new 90% backstop

Follow-up (manual, after merge)

1. Add codecov/patch as a required status check in branch protection.
2. Open PRs adopt the new gate by syncing their branch with main (CI checks out the PR head SHA, so a re-run alone isn't enough) — the same "sync with main" step that previously broke them now fixes them.

Codecov app is installed and CODECOV_TOKEN is set.

[ghost]

Note

Gittensory Gate skipped

PR closed before full evaluation. No late first comment was created.

Signal	Result	Evidence	Action
Gate result	⚠️ Skipped	#763 is no longer open.	No action.

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by Gittensory, a quiet PR intelligence layer for OSS maintainers.

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[ghost]

gittensory · advisory review

Reviewed 4 changed file(s) — two independent AI reviewers.

Changed files: .github/workflows/ci.yml, CONTRIBUTING.md, codecov.yml, vitest.config.ts

Suggested action: ✅ Safe to merge — both reviewers found no blocking issues.

Reviewer A · gpt-oss-120b — recommends ✅ merge
This PR introduces a Codecov patch‑coverage gate, adds a Codecov config, adjusts CI to upload coverage, and lowers the global vitest thresholds to a safe 90% backstop while updating documentation. The changes are straightforward and do not affect runtime code or security boundaries.

Suggestions

• Verify that the CODECOV_TOKEN secret is configured in the repository settings so the upload step succeeds.
• Confirm that the npm run test:coverage step runs before the Codecov upload in the CI workflow (the diff only shows the tail of the job).
• Consider adding a brief comment in vitest.config.ts noting why the thresholds were lowered to 90% for future maintainers.

Worth double-checking

• If the CI job order is incorrect, the lcov.info file might not exist when Codecov runs, causing a failed upload.
• Missing or mis‑named CODECOV_TOKEN secret would break the coverage upload step.

Reviewer B · nemotron-3-120b-a12b — recommends ✅ merge
The PR shifts coverage enforcement from a global vitest threshold (97%) to a patch-based gate via Codecov (97% on changed lines), with vitest thresholds lowered to 90% as a catastrophe backstop. This eliminates cross-PR churn while maintaining quality guards. Changes are well-documented, logically sound, and align with the project's testing conventions.

Worth double-checking

• Ensure CODECOV_TOKEN secret is configured in the repo (required for Codecov action to function).
• Verify that 90% vitest thresholds are sufficiently below actual coverage to avoid false positives on routine runs.
• Confirm Codecov's patch coverage calculation correctly ignores excluded paths (src/env.d.ts, apps/, test/, scripts/).

[codecov]

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

Thanks for integrating Codecov - We've got you covered ☂️