fix(ci): unbreak eslint and the multi-semantic-release git fixtures

[prisis]

Two independent breakages found while chasing the red pipelines on #403/#404/#405. Both are on main today, neither comes from those branches.

1. lint:eslint fails on a clean checkout (commit 2)

pnpm run lint:eslint reports nine errors on plain main, so the Lint (eslint, types, attw) job is red on every pull request:

• six eslint-disable comments no longer match a reported problem — vitest/require-mock-type-parameters in rc.test.ts, rc-unmocked.test.ts and semantic-release-clean-package-json/__tests__/index.test.ts, @typescript-eslint/naming-convention in bin/cli.test.ts. With reportUnusedDisableDirectives those are errors. Deleted.
• release-notes-preset-compat.test.ts calls generateNotes from @semantic-release/release-notes-generator, which ships no type declarations, so the call and its result are typed as error and trip no-unsafe-call/no-unsafe-assignment. Declared the module next to the other untyped semantic-release modules in src/types/semantic-release.d.ts.
• the same file's JSDoc names the package unquoted, which reads as an inline tag (jsdoc/escape-inline-tags). Quoted.

After this, pnpm run lint:eslint and pnpm run lint:types both pass repo-wide.

2. The git fixtures inherit the developer's git setup (commit 1)

25 of multi-semantic-release's integration tests fail locally on a machine configured the way most are today. Three separate causes, all in __tests__/helpers/git.ts:

GIT_DIR and friends. Git exports GIT_DIR, GIT_INDEX_FILE, GIT_OBJECT_DIRECTORY and GIT_WORK_TREE to its hooks, and they take precedence over cwd — for the helper's own calls and for the semantic-release runs the tests spawn. Run the suite from a hook, which the repo's own lint-staged config does (vitest related on a changed test file), and the fixtures operate on the repository being committed to.

This is not theoretical: it happened to me while preparing this branch. The pre-commit hook ran these tests, a fixture's git add/git commit went to the real index, and the commit that came out deleted 389 files — the whole repository — because the tree it committed was a fixture's. I reset it; nothing was pushed. The helper now drops those variables at import.

Verified both ways, with GIT_DIR/GIT_INDEX_FILE/GIT_WORK_TREE pointed at a scratch repo:

before: 2 failed (2)   after: 2 passed (2)   scratch index: unchanged

and the commit of this very change went through the real pre-commit hook without disturbing the index.

The remote's default branch. gitInitRemote runs git init --bare, so the remote's HEAD follows the developer's init.defaultBranch. For anyone not on master that HEAD names a branch which is never pushed, and semantic-release dies with git fetch --tags → "couldn't find remote ref HEAD". Now created with --initial-branch=master, matching what gitInit makes.

Tag signing. tag.gpgsign = true in a global config turns the plain git tag <name> <sha> semantic-release runs into a signed tag, which needs a message it does not pass: "fatal: no tag message?". The helper already disabled commit.gpgsign; it now disables tag.gpgsign too.

With all three, multi-semantic-release.test.ts and bin/cli.test.ts go from 25 failures to 29/29 passing locally.

Not fixed here

• bin/cli.test.ts runs right at the 5 s per-test timeout on a loaded machine and times out intermittently; it passes on its own. Worth a longer testTimeout, but that is a separate call.
• prettier --check fails on three files in semantic-release-pnpm that nothing here touches. CI does not run prettier, so it is not blocking.

🤖 Generated with Claude Code

https://claude.ai/code/session_01MqVjBDe8SkgMK7LpDpgjJ7

[coderabbitai]

Warning

Review limit reached

@prisis, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 40 minutes

Limit details: You’ve used the included review currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

How can I continue?

Wait for the limit to reset, then comment @coderabbitai review or push new commits to the PR.

An organization admin can change what happens after included review limits in Billing.

How do review limits work?

CodeRabbit enforces per-developer PR review limits within each organization.

For paid Pro and Pro+ reviews, CodeRabbit uses a developer's included PR review attempts over the past 7 days to set the current hourly allowance. At typical activity levels, the full plan allowance applies. Higher sustained activity can lower the allowance until earlier attempts leave the 7-day window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: f148bda5-e2b0-4101-a672-6fb927ff6b4b

📥 Commits

Reviewing files that changed from the base of the PR and between aad7d28 and b60faae.

📒 Files selected for processing (7)

• packages/multi-semantic-release/__tests__/bin/cli.test.ts
• packages/multi-semantic-release/__tests__/helpers/git.ts
• packages/multi-semantic-release/__tests__/release-notes-preset-compat.test.ts
• packages/multi-semantic-release/src/types/semantic-release.d.ts
• packages/rc/__tests__/rc-unmocked.test.ts
• packages/rc/__tests__/rc.test.ts
• packages/semantic-release-clean-package-json/__tests__/index.test.ts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Thank you for following the naming conventions! 🙏

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.48%. Comparing base (71c4ed4) to head (b60faae).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #407      +/-   ##
==========================================
+ Coverage   86.67%   87.48%   +0.81%     
==========================================
  Files          47       26      -21     
  Lines        1426      983     -443     
  Branches      387      279     -108     
==========================================
- Hits         1236      860     -376     
+ Misses        178      113      -65     
+ Partials       12       10       -2     

Flag	Coverage Δ
multi-semantic-release	87.70% <ø> (ø)
rc	95.45% <ø> (ø)
semantic-release-clean-package-json	76.19% <ø> (ø)
semantic-release-pnpm	?
see 22 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.