GitHub Actions Rerun Waste Audit

Quantify wasted GitHub Actions minutes caused by reruns so flaky workflows can be fixed with data.

Security Scan

The skill is internally consistent with its stated purpose: it analyzes local GitHub Actions run JSON files to estimate wasted minutes from reruns, requires only bash/python3, and does not ask for credentials or perform network exfiltration — but note a minor inconsistency (the SKILL.md shows using the `gh` CLI while the declared required binaries omit it) and the script will read any files matched by RUN_GLOB so you should avoid pointing it at sensitive paths.

Purpose & Capability

Name/description match the included script and fixtures: the tool parses run JSON exports and computes rerun waste. Minor mismatch: SKILL.md examples show using the `gh` CLI to collect run JSON, but the declared required binaries list only 'bash' and 'python3' (omitting 'gh'). Including 'gh' in the requirements would be more accurate.

Instruction Scope

Runtime instructions and the script stay within scope: they read run JSON files (default artifacts/github-actions/*.json), apply filters, compute aggregates, and emit text/JSON. The script does not access external endpoints or other credentials. One operational note: RUN_GLOB is user-controlled — the script will read any files matching that glob, so do not point it at sensitive system files or directories.

Install Mechanism

No install spec (instruction-only plus bundled script). No downloads or archive extraction; nothing is written to disk beyond reading existing JSON files. This is low-risk from an install perspective.

Credentials

The skill requests no secrets or external credentials. All environment inputs are configuration knobs (glob, thresholds, regex filters). The number of env inputs is reasonable for this task and none are named like PASSWORD/TOKEN/KEY.

Persistence & Privilege

always is false and the skill does not attempt to modify agent/system configuration or other skills. It runs ad hoc and does not request permanent elevated presence.

Assessment

This skill appears to be what it says: it analyzes local GitHub Actions run JSON files and does not ask for credentials or make network calls. Before installing or running: (1) if you plan to collect runs with the GitHub CLI, add 'gh' to the required binaries or ensure it's available; (2) avoid setting RUN_GLOB to match sensitive system paths — the script will read any files you point it at; (3) test using the included fixtures RUN_GLOB to confirm output format; (4) review the script if you intend to run it in automated CI (note FAIL_ON_CRITICAL will cause a nonzero exit); and (5) if you need stricter guarantees, run it in a sandbox or review the full script output to ensure it meets your privacy/policy requirements.

Use this skill to measure CI waste caused by reruns and repeated attempts in GitHub Actions.

What this skill does

• Reads one or more run JSON exports (gh run view --json ...)
• Groups attempts by repository + workflow + branch + commit + job name
• Estimates rerun waste from all attempts except the latest attempt in each group
• Flags severity using waste-minute thresholds
• Emits text or JSON output for triage dashboards and CI quality reviews

Inputs

Optional:

• RUN_GLOB (default: artifacts/github-actions/*.json)
• TOP_N (default: 20)
• OUTPUT_FORMAT (text or json, default: text)
• WARN_MINUTES (default: 10)
• CRITICAL_MINUTES (default: 30)
• FAIL_ON_CRITICAL (0 or 1, default: 0)
• WORKFLOW_MATCH, WORKFLOW_EXCLUDE (regex, optional)
• JOB_MATCH, JOB_EXCLUDE (regex, optional)
• REPO_MATCH, REPO_EXCLUDE (regex, optional)
• BRANCH_MATCH, BRANCH_EXCLUDE (regex, optional)

Collect run JSON

gh run view <run-id> --json databaseId,workflowName,headBranch,headSha,runAttempt,url,repository,jobs \
  > artifacts/github-actions/run-<run-id>.json

If runAttempt is missing, this skill also checks run_attempt and job-level attempt.

Run

Text report:

RUN_GLOB='artifacts/github-actions/*.json' \
WARN_MINUTES=8 \
CRITICAL_MINUTES=20 \
bash skills/github-actions-rerun-waste-audit/scripts/rerun-waste-audit.sh

JSON output + fail gate:

RUN_GLOB='artifacts/github-actions/*.json' \
OUTPUT_FORMAT=json \
FAIL_ON_CRITICAL=1 \
bash skills/github-actions-rerun-waste-audit/scripts/rerun-waste-audit.sh

Repo/workflow filter:

RUN_GLOB='artifacts/github-actions/*.json' \
REPO_MATCH='^flowcreatebot/' \
WORKFLOW_MATCH='(CI|Build)' \
bash skills/github-actions-rerun-waste-audit/scripts/rerun-waste-audit.sh

Run with bundled fixtures:

RUN_GLOB='skills/github-actions-rerun-waste-audit/fixtures/*.json' \
bash skills/github-actions-rerun-waste-audit/scripts/rerun-waste-audit.sh

Output contract

• Exit 0 in reporting mode
• Exit 1 when FAIL_ON_CRITICAL=1 and at least one critical group exists
• Text output includes summary, thresholds, and top rerun-waste groups
• JSON output includes summary, ranked groups, and critical_groups