Ci Monitor

ci-monitor

> Priority: P1 (runs every 2 hours) > Data sources: GitHub Actions API > Output: CI reports in memory/ops/ci/

Instructions

You are executing the ci-monitor skill for the Ops agent.

Step 1: Fetch Workflow Runs

curl -s -H "Authorization: token ${GITHUB_TOKEN}" \
  "https://api.github.com/repos/{owner}/{repo}/actions/runs?per_page=30&status=completed"

Also fetch currently running workflows:

curl -s -H "Authorization: token ${GITHUB_TOKEN}" \
  "https://api.github.com/repos/{owner}/{repo}/actions/runs?status=in_progress"

Step 2: Analyze Build Health

Success rate:

• Overall pass/fail ratio (last 30 runs)
• Per-workflow pass rate
• Trend (improving or degrading)

Build times:

• Average build time per workflow
• P95 build time
• Trend (getting faster or slower)

Failure patterns:

• Most common failure reasons
• Flaky tests (tests that sometimes pass, sometimes fail)
• Infrastructure failures vs. code failures

Step 3: Identify Flaky Tests

A test is "flaky" if:

• It failed in a run but the same commit passed on retry
• It fails >10% of the time but 50%

Anti-Patterns

• Do NOT restart failed builds automatically. Report and let humans decide.
• Do NOT ignore flaky tests. They erode CI reliability over time.
• Do NOT consider infrastructure timeouts as code failures.

Exit Codes

• SKILL_OK — CI monitoring complete
• SKILL_PARTIAL — some workflow data unavailable
• SKILL_FAIL — could not access GitHub Actions

Output

Commit message format: ops: ci-monitor — {success_rate}% success, {N} runs, {M} flaky tests

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: stoaaadev
• Source: stoaaadev/stoa
• License: MIT
• Homepage: https://x.com/stoaframework

Install and usage instructions live in the source repository linked above.