skill-manager 技能生命周期

To check if locally stored GitHub skills are outdated, run the check workflow to compare local github_hash metadata against the remote HEAD. This automated detection generates a JSON status report classifying each skill as Current, Outdated, or Error.

The best way to automate GitHub skill updates is using a dedicated workflow that scans local frontmatter metadata, detects version staleness via git-ls-remote, and drives a structured upgrade cycle. This ensures safe, guided refactors with backups for multiple skills.

Git-ls-remote works for detecting updates by fetching the remote HEAD hash of a GitHub repository and comparing it to the github_hash stored in the local skill's frontmatter. If the hashes differ, the skill is flagged as Outdated in the generated JSON report.

Yes, you can audit and manage a large inventory of local skills automatically. The skill-management workflow scans environments with many skills, extracts github_url metadata, and provides inventory features to list, check, and delete skills as needed.

Yes, you need YAML frontmatter metadata to use the skill update workflow. The scanner relies on specific frontmatter fields like github_url, github_hash, and version to identify skills and perform accurate comparisons against the remote repository.

The generated JSON report includes a status summary classifying each scanned skill into Current, Outdated, or Error states. It uses the extracted frontmatter and remote hash comparisons to guide the subsequent update and refactor cycle.