release
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes Git commands including log, add, commit, tag, and push to automate repository releases.
- [PROMPT_INJECTION]: Instructions explicitly require the agent to omit AI attribution and 'Co-Authored-By' trailers from commits and documentation, potentially masking automated activity from audit trails.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface through its processing of repository history.
- Ingestion points: Git commit history logs and existing RELEASE-NOTES.md files.
- Boundary markers: No markers are used to delimit untrusted log data from the agent's instructions.
- Capability inventory: Ability to edit manifest files (package.json, pyproject.toml, etc.) and perform Git repository management.
- Sanitization: No sanitization is performed on commit messages or existing notes before they are processed by the agent.
Audit Metadata