make-release
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from git commit logs and repository files to generate release documentation.\n
- Ingestion points: Git commit history, pyproject.toml, package.json, and CHANGELOG.md.\n
- Boundary markers: The skill lacks explicit prompt delimiters but includes a verification step to ensure changelog headings match the intended version.\n
- Capability inventory: File writing, subprocess execution (uv, git, gh, npm), and network operations for publishing.\n
- Sanitization: The skill mitigates this risk by requiring an explicit 'Release plan and approval' phase where all proposed changes and changelog entries are presented to the user for review before any commands are executed.\n- [DYNAMIC_EXECUTION]: The skill can generate and write GitHub Action workflow files for automated publishing.\n
- Evidence: Instructions to 'Draft one dedicated publish workflow' for PyPI publishing.\n
- Observation: The generated workflow content is subject to the same mandatory user approval process as all other release mutations.\n- [CREDENTIALS_UNSAFE]: The skill manages publishing credentials for PyPI and GitHub.\n
- Evidence: References to PYPI_API_TOKEN and UV_PUBLISH_TOKEN.\n
- Observation: The skill follows security best practices by explicitly forbidding the agent from printing tokens in chat or committing them to the repository, instead favoring secure secret management tools.
Audit Metadata