github-actions-stuck-run-audit

What This Skill Does

The github-actions-stuck-run-audit skill is a specialized diagnostic utility designed to monitor and evaluate the health of GitHub Actions workflows. In high-velocity development environments, it is common for workflows to become trapped in non-terminal states like 'queued', 'in_progress', or 'waiting', effectively creating a bottleneck that delays deployment pipelines and consumes concurrency limits. This skill ingests JSON exports of your GitHub Actions runs and applies sophisticated logic to identify these stale processes. It aggregates data by repository, workflow, or specific branch, allowing teams to visualize exactly where their CI/CD health is degrading. By using configurable thresholds—including stuck-age, total stuck-run volume, and stuck-rate—the skill calculates a severity score for each group. It acts as a gatekeeper for your automated processes, providing both readable text summaries and machine-readable JSON outputs that can be integrated into broader automated monitoring pipelines.

Installation

To install this skill, use the OpenClaw command-line interface within your local repository:

clawhub install openclaw/skills/skills/daniellummis/github-actions-stuck-run-audit

Ensure that you have access to the source repository at openclaw/skills. Once installed, you will need to point the tool at your exported GitHub Actions run data, typically generated via the gh run view command provided in the skill documentation.

Use Cases

This skill is essential for DevOps engineers and SRE teams managing multiple high-traffic repositories. Common use cases include:

• Automated CI Gatekeeping: Failing builds or deployments if the queue for critical pipelines reaches a 'critical' severity status.
• Bottleneck Analysis: Identifying specific branches or workflow configurations that consistently fail to transition to completion, often pointing to resource contention or runner infrastructure issues.
• Maintenance Audits: Running weekly reports to identify under-performing workflows that require optimization or cleanup.

Example Prompts

1. "Audit the GitHub actions runs in the artifacts folder, grouping by repository-workflow, and output the result as a text report."
2. "Run a stuck-run audit with a 60-minute critical threshold and fail the process if any critical workflows are detected."
3. "Analyze the action runs in the fixture directory using a simulated timestamp of March 8th, 2026, to verify our alerting logic."

Tips & Limitations

• Data Consistency: The accuracy of this skill is entirely dependent on the quality of the JSON exports collected via gh run. Ensure your collector script captures all necessary fields (e.g., status, createdAt, updatedAt).
• Threshold Tuning: The defaults (45 minutes for warn, 120 for critical) are starting points. Adjust these in your own CI environment based on typical job durations—short-lived unit tests will need much stricter timeouts than long-running integration suites.
• Performance: While the audit process is lightweight, parsing thousands of JSON run files may take a few seconds. If audit volume is massive, consider using the TOP_N filter to focus only on the most severe offenders.