ci: 6 open fix PRs stalled — 3 failing runs on PR #32, fixes ready on PR #33 and #34

[erudenko]

CI Maintenance Audit — 2026-04-30

Automated CI maintenance scan found 3 failed workflow runs and a backlog of 6 competing fix PRs, none merged in 18+ days. Fixes are already implemented and passing CI. This issue documents the state and recommends a merge path.

---

Failed Runs Found

All failures are on PR #32 (fix/ci-workflow-trigger-broadening), head commit 6f38db6, from 2026-04-12:

Run ID	Workflow	Job	Result	Root Cause
24298790072	Test Plugins	test (20)	❌ FAILED	TS6133 unused imports
24298790072	Test Plugins	test (22)	⚠️ CANCELLED	Upstream job failed
24298790072	Test Plugins	marketplace-sync	⏭ SKIPPED	Upstream job failed
24298790076	Test Stats Plugin	test	❌ FAILED	Stale hardcoded dates
24298782547	Test Stats Plugin	test	❌ FAILED	Stale hardcoded dates

---

Root Causes

1. TypeScript TS6133 — unused imports

Three imports left behind after refactors, caught by tsc --noEmit under noUnusedLocals:

• tools/claudeup-core/src/services/conventions-manager.ts — existsSync
• tools/claudeup-core/src/__tests__/unit/doctor.test.ts — vi
• tools/claudeup-core/src/__tests__/integration/conventions-integration.test.ts — removeGitignoreEntries

2. Stale hardcoded test dates

plugins/stats/tests/integration.test.ts used hardcoded "2026-03-26" for session dates. The stats plugin's time-windowed DB queries (getSessionSummary, getTopTools, getDurationTrend) use 7-day and 14-day rolling windows. By April 12 this date was 17 days old — outside all windows — causing assertions on row counts to fail.

3. browser-use version mismatch

plugins/browser-use/plugin.json was bumped to v1.1.2 but .claude-plugin/marketplace.json was left at v1.1.1. The marketplace-sync integration test catches this drift.

4. PR #32 is a partial fix

PR #32 only added the workflow trigger fix (removed hardcoded branches allowlist from test-plugins.yml and test-stats.yml). It did NOT include the code-level fixes above, so CI ran but the tests themselves failed. PR #32 is superseded by PR #33.

---

Current State of All Open Fix PRs

PR	Branch	Base	CI	Notes
#22	fix/ci-typescript-and-stats-dates	fix/ci-type-errors-and-stale-test-date	GitGuardian only	Test CI never ran (created before trigger fix) — superseded
#23	fix/ci-typescript-and-stats-dates-resolved	fix/ci-type-errors-and-stale-test-date	GitGuardian only	Same as #22 — superseded
#24	fix/ci-24065489776-ts-and-dates	fix/ci-type-errors-and-stale-test-date	✅ 7/7 passing	Comprehensive fix — base for #33
#28	fix/ci-24276607217-browser-use-version	fix/ci-24065489776-ts-and-dates	GitGuardian only	Stacked on #24; test CI never ran — superseded
#31	fix/ci-main-browser-use-version	main	GitGuardian only	Targets dist-only main — no test workflows there
#32	fix/ci-workflow-trigger-broadening	fix/ci-type-errors-and-stale-test-date	❌ 3 failures	Partial fix — superseded by #33
#33	fix/ci-consolidated	fix/ci-type-errors-and-stale-test-date	✅ 7/7 passing	All fixes + trigger broadening — ready to merge
#34	fix/ci-all-fixes-final	fix/ci-type-errors-and-stale-test-date	✅ 13/13 passing	All fixes including integration.test.ts dates — ready to merge

---

Recommended Actions

Immediate

1. Merge PR fix(ci): consolidated — TS imports, stale dates, browser-use version, unrestricted triggers #33 or PR fix(ci): complete all-fixes branch — rolling stats dates in integration.test.ts + browser-use 1.1.2 #34 into fix/ci-type-errors-and-stale-test-date — both are fully green. PR fix(ci): complete all-fixes branch — rolling stats dates in integration.test.ts + browser-use 1.1.2 #34 is more thorough (13 checks vs 7) and also covers integration.test.ts dynamic dates. Prefer fix(ci): complete all-fixes branch — rolling stats dates in integration.test.ts + browser-use 1.1.2 #34 if it doesn't conflict.
2. Close PRs fix(ci): remove unused TS imports + fix stale stats test dates #22, fix(ci): resolve merge conflict — unused TS imports + stale stats test dates #23, fix(marketplace): bump browser-use version 1.1.1 → 1.1.2 in marketplace.json #28, fix(ci): remove hardcoded branch filters from workflow triggers #32 as superseded (not_planned). They either have no test CI or have failures that are fully covered by fix(ci): consolidated — TS imports, stale dates, browser-use version, unrestricted triggers #33/fix(ci): complete all-fixes branch — rolling stats dates in integration.test.ts + browser-use 1.1.2 #34.
3. Decide on PR fix(marketplace): bump browser-use version 1.1.1 → 1.1.2 to fix CI #31 — it targets the dist-only main branch with a marketplace.json bump. If the source-code fix (via fix(ci): consolidated — TS imports, stale dates, browser-use version, unrestricted triggers #33/fix(ci): complete all-fixes branch — rolling stats dates in integration.test.ts + browser-use 1.1.2 #34) also updates marketplace.json on the source branch, this PR is only relevant if main is ever re-synced from source.

Longer-term

• Consider renaming or protecting fix/ci-type-errors-and-stale-test-date as the primary development branch, or establish a clear develop/main branching strategy to avoid future confusion.
• The test files should continue using dynamic date constants (TODAY, YESTERDAY, etc.) to prevent stale-date failures from recurring.

---

Generated by automated CI maintenance scan.