automation-audit-ops

To audit GitHub Actions workflows, you perform a read-only inventory that traces run logs and configuration files to classify each automation as live, broken, or redundant. This generates a keep, merge, or cut report for project stability.

The best way to inventory repository automation surfaces is to enumerate GitHub Actions, configuration hooks, and build scripts, then evaluate their operational status against concrete proof like run logs. This identifies stale workflows before changes.

Evidence-based state classification works by tracing concrete proof from execution logs and command outputs to evaluate the health of CI/CD components. Automations are classified as live, broken, or unverified based on this traced evidence.

Yes, a read-only audit can identify stale Composer scripts and MCP servers by cataloging these configuration surfaces and evaluating their health. It checks operational status through command outputs to classify them as live or redundant.

No, this automation audit does not require write access because it performs a comprehensive, read-only evaluation of your repository configuration files and logs. It simply traces evidence to provide actionable recommendations.

GitHub Actions workflows are classified as unverified during an audit when there is no concrete proof of their operational status in run logs or command outputs. This classification indicates a lack of evidence for live or broken states.