// When GitHub Actions fails, fetch failing job logs and assign each failing job to a separate subagent that fixes its slice of the problem in parallel. Use for multi-job CI failures where jobs are independent.
| name | parallel-ci-triage |
| description | When GitHub Actions fails, fetch failing job logs and assign each failing job to a separate subagent that fixes its slice of the problem in parallel. Use for multi-job CI failures where jobs are independent. |
| user-invocable | true |
Speed up fixing broken CI by splitting failing jobs (or independent failure clusters) across parallel subagents. Each subagent owns one vertical slice: logs, root cause, code fix, and local verification for that slice.
gh) installed and authenticated (gh auth login), or use the GitHub web UI / API to copy logs manually.From the repo root:
gh run list --limit 5
gh run view <RUN_ID> --log-failed
Or open the Actions tab, open the failed workflow run, and note which jobs failed (not just which step — group by job name).
If gh is unavailable, download logs from the GitHub UI and paste them into the conversation.
lint, test-node-18, e2e).If two failures share the same root cause in the same file, assign one subagent to fix both.
For each failing job, launch a generalPurpose subagent in a single message so they run concurrently:
Task: Fix CI failure for job "<JOB_NAME>"
Context:
- Workflow run: <RUN_URL or RUN_ID>
- Branch: <branch>
- Relevant log excerpt (failed steps only):
<paste gh run view --job <JOB_ID> --log or the failed section>
Instructions:
1. Infer the root cause from the log (command, stack trace, file:line).
2. Open and edit only what this job requires.
3. Run the same commands locally that failed in CI (or the narrowest equivalent, e.g. `npm run lint`, `pytest tests/foo`, `pnpm test --filter pkg`).
4. Report: what failed, what you changed, and confirmation that the local command passes.
Include the exact failing command and error lines so the subagent does not guess.
Collect each subagent’s changed files. Resolve overlaps manually if two agents touched the same file.
Run the full CI-equivalent locally when possible:
# Example: match your repo
npm run lint && npm test
Commit with a conventional message, push, and re-check the workflow:
gh run watch
Generate or edit images using the OpenAI Image API (gpt-image-2). Use when the user asks to generate, create, draw, render, illustrate, mock up, or edit an image, icon, logo, mockup, illustration, OG image, blog hero, marketing asset, or similar visual. Also use when the user supplies a reference image and asks to modify, restyle, or remix it. Triggers on: "generate an image", "create an image", "make a picture of", "edit this image", "restyle this", "make a mockup of", "draw a", "render a", "illustration of".
When the same multi-step workflow repeats in Cursor (user corrections or agent redos), capture it as a new SKILL.md under .cursor/skills/ so future sessions load it automatically.
After navigating and interacting in Cursor's built-in browser, use browser_network_requests to audit every fetch/XHR for failures, slowness, duplicate calls, and suspicious payloads. Use for API-heavy pages and after backend or client networking changes.
Run four parallel read-only subagents that each review the same diff from a different lens — security, performance, correctness, and readability — then merge findings into one report. Use before merging large or risky PRs.
Execute a user flow step-by-step in Cursor's built-in browser while documenting each action, then emit a Playwright test that replays the same flow using stable selectors derived from the accessibility tree.
Automatically persist useful context — research, decisions, learnings, templates — to workspace files so knowledge survives across conversations.