The Agent Skills Directory

[COMMAND_EXECUTION]: The exec function in scripts/analyze.ts and scripts/release.ts executes shell commands by passing string-interpolated arguments to sh -c. This pattern is susceptible to command injection if variables like the version string (provided via CLI arguments) or repository metadata (derived from git remotes) contain shell metacharacters such as semicolons or backticks.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests untrusted data from the environment. Ingestion point: git log output in scripts/analyze.ts. Boundary markers: None. Capability inventory: git push, git tag, gh release create, and file writing in scripts/release.ts. Sanitization: The scripts use regular expressions for parsing conventional commits but do not sanitize the resulting content before interpolating it into shell commands or writing it to the changelog.
[EXTERNAL_DOWNLOADS]: The skill legitimately interacts with external services to perform release tasks. It utilizes git for remote repository synchronization and the GitHub CLI (gh) for CI status monitoring and creating official releases on GitHub. these operations target well-known developer services.

release