[ThomasK33]

Why

#176 moved the repo to pnpm so that Dependabot can update dependencies. On its first run on main, Dependabot's security jobs for @vitest/mocker, brace-expansion, minimatch and postcss all failed with security_update_not_possible. A pin held each package:

• vitest is pinned exactly and pulls @vitest/mocker 4.1.11.
• postcss and brace-expansion are pinned exactly in pnpm.overrides. With an exact override, pnpm update postcss@8.5.32 --lockfile-only exits 0 but leaves 8.5.18 in the lock.
• minimatch 10 and glob 13 come from dependencies whose ranges exclude the fixed majors.

All of these are dev dependencies. None of them ships in the npm package.

What changed

Advisory	Vulnerable	Fix
GHSA-qxg6-89mg-rqv8, GHSA-xp26-3454-mjgp	@vitest/mocker < 4.1.12	vitest 4.1.11 → 4.1.12
GHSA-fxqj-rqcc-2cmp	postcss <= 8.5.22	override floor ^8.5.32
GHSA-q2hr-2g5m-vwhr	brace-expansion < 5.0.12	override floor ^5.0.12
GHSA-3g3m-2fjr-6pw4	glob <= 14.0.1	drop rimraf, the only package that pulled glob
GHSA-g3hv-75wg-m249	minimatch < 11.0.1	new override floor ^11.0.3

1. No more rimraf. The clean script (fixed in fix: make npm run clean remove the build info files #348) now runs node -e with fs.rmSync(..., { recursive: true, force: true }) on the same four paths. The script uses only single quotes inside its double-quoted argument, so it works in sh and in cmd.exe. The clean-script.test.ts test from fix: make npm run clean remove the build info files #348 still passes against it.
2. Overrides are caret floors. Every pnpm.overrides entry is now ^<version> instead of an exact version. The lock keeps the same resolved versions for the entries this PR does not bump. With ^8.5.18, pnpm update postcss@8.5.32 --lockfile-only --no-save -r moved the lock to 8.5.32, while the exact override left it at 8.5.18. This helps Dependabot only for postcss and nanoid, which sit under vite's re-resolved subtree. For minimatch and brace-expansion, Dependabot's command leaves the lock unchanged even with a caret floor, so a future advisory on them still needs a manual bump, as before this PR.
3. minimatch override. release-please's @google-automations/git-file-utils still asks for minimatch ^10.2.4, also in its latest release, and every 10.x is vulnerable. git-file-utils uses only new Minimatch(glob).match(path), which works the same with the CommonJS build of 11.x. An earlier verifier ran findFilesByGlob with 34 patterns and got the same results on 10.2.5 and 11.x.
4. Other lock changes. tslib 2.9.1 → 2.8.1 and @napi-rs/wasm-runtime 1.2.7 → 1.2.6 (main had releases younger than 7 days; both sit only under rolldown's wasm32 binding). tinyglobby 0.2.15 and picomatch 4.0.4 drop out because vitest 4.1.12 uses newer ones. minimatch 11 adds @isaacs/nonrecursive-glob-match, and lru-cache 11 leaves with glob. The production dependency closure is unchanged.
5. No locked release younger than 7 days. Dependabot applies its 7-day cooldown to every locked package (see fix(deps): unblock Dependabot's pnpm resolve (release age and trust checks) #346), so I relocked with --config.minimum-release-age=10080. minimatch resolves to 11.0.3 (11.0.4 is 3 days old), and dropping rimraf removes rimraf@6.1.4 (2 days old).

BEGIN_COMMIT_OVERRIDE
fix(deps): clear the moderate dev-dependency advisories that Dependabot cannot fix
END_COMMIT_OVERRIDE

Validation

• pnpm audit: "No known vulnerabilities found" (main: 6 moderate).
• .agents/skills/verify-agent-tty/scripts/dependabot-resolve.sh (empty-store replay of Dependabot's resolve, from fix(deps): unblock Dependabot's pnpm resolve (release age and trust checks) #346): 3 of 3 resolved.
• pnpm install --frozen-lockfile, then mise run --force of format-check, workflow-lint, lint, audit, typecheck, clean, build, test-unit (146 files, 2137 tests), install-smoke and validate-bundles: exit 0. After clean, dist/ and both build info files are gone, and build writes dist/cli/main.js.
• Integration and e2e run in CI.

---

Generated with xum • Model: anthropic:claude-opus-5-5 • Thinking: high

[ThomasK33]

@codex review

[ThomasK33]

@codex security review

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-06T01:00:00.383889Z	b589f2b	Manual request
🔒 Security Review	✅ Completed 2026-11-05T21:46:14.866083Z	0d38513	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Bravo.

Reviewed commit: 0d38513193

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

🛡️ Codex Security Review

Security review completed. No security issues were found in this pull request.

Reviewed commit: 0d38513193

View security finding report

Only the user who started this review can view the report in Codex.

ℹ️ About Codex security reviews in GitHub

This is an experimental Codex feature. Security reviews are triggered when:

• You comment "@codex security review"
• A regular code review gets triggered (for example, "@codex review" or when a PR is opened), and you’re opted in so security review runs alongside code review

Once complete, Codex will leave suggestions, or a comment if no findings are found.

[ThomasK33]

@codex review

[ThomasK33]

@codex security review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. More of your lovely PRs please.

Reviewed commit: b589f2bc9b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

🛡️ Codex Security Review

Security review completed. No security issues were found in this pull request.

Reviewed commit: b589f2bc9b

View security finding report

Only the user who started this review can view the report in Codex.

ℹ️ About Codex security reviews in GitHub

This is an experimental Codex feature. Security reviews are triggered when:

• You comment "@codex security review"
• A regular code review gets triggered (for example, "@codex review" or when a PR is opened), and you’re opted in so security review runs alongside code review

Once complete, Codex will leave suggestions, or a comment if no findings are found.

[ThomasK33]

Review record for the merge decision.

• Final commit: b589f2b (main 2322619 merged in).
• CI on the final commit: every check passed, including quality-gates, quality-gates-macos and pr-title. One duplicate pr-title run was cancelled by the workflow's concurrency setting.
• Codex: 2 rounds (0d38513, b589f2b). Both reviews found no issues in each round. There are no unresolved threads.
• Independent verifiers: PASS on 0d38513 and PASS on b589f2b. Between them, a replay of Dependabot's resolve on the combined fix(deps): unblock Dependabot's pnpm resolve (release age and trust checks) #346 + fix(deps): clear the moderate dev-dependency advisories that Dependabot cannot fix #347 tree showed that rimraf 6.1.4 and minimatch 11.0.4 were younger than the 7-day cooldown. The head drops rimraf and locks minimatch 11.0.3.
• Clean-context advisor: READY. Its known limit is that Dependabot still cannot move minimatch or brace-expansion past their override floors. That was already true with the exact pins, and the PR body states it.
• Reason for stopping: every gate is clean on the final commit.

---

Generated with xum • Model: anthropic:claude-opus-5-5 • Thinking: high