The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes git and gh commands via a Node.js script to gather repository metadata. It includes validation for the --repo argument to prevent arbitrary command injection. The helper script is invoked via Node.js to perform local diagnostics without external dependencies.
[INDIRECT_PROMPT_INJECTION]: The skill ingests commit messages and pull request data from the local environment and GitHub. While this provides a surface for indirect injection, the skill uses strict delimiters and is functionally limited to read-only reporting.
Ingestion points: Commits subjects, bodies, and PR labels/titles processed in scripts/release-status.mjs.
Boundary markers: The script uses specific control characters (\u001E and \u001F) to parse git log output reliably.
Capability inventory: The skill uses git, gh, and node to read information; it contains no write or exfiltration capabilities.
Sanitization: Employs regular expressions to filter and classify Conventional Commit subjects.

release-status