ln-62-release-publisher
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, specifically git commit messages, repository manifests, and documentation. This presents a potential surface for indirect prompt injection.
- Ingestion points: Reads git commit history, commit bodies, repository manifests, and migration notes (SKILL.md).
- Boundary markers: The skill explicitly requires a human-in-the-loop approval step ("Wait for explicit user approval of that exact proposal before committing version changes, tagging, or publishing") which serves as a critical security boundary.
- Capability inventory: The skill uses git commands, GitHub CLI/API, and repository-native "release gates" (testing/validation scripts).
- Sanitization: It mandates the use of temporary files for release notes to prevent shell re-interpretation of Markdown, quotes, and code blocks during publication.
- [CREDENTIALS_UNSAFE]: The skill contains explicit instructions to handle credentials securely, directing the agent to use the host credential store and avoid echoing secrets.
- [COMMAND_EXECUTION]: The skill uses repository-native gates and standard tools like
gitandgh(GitHub CLI) for its primary function. These are appropriate for a release management tool.
Audit Metadata