release
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
Bashto automate project management tasks throughgit,gh(GitHub CLI), andfj(Forgejo CLI). These commands are standard for release orchestration and include safety checks, such as prompting the user before performing commits or pushes. - [EXTERNAL_DOWNLOADS]: The skill performs network requests to official registries like
pypi.organd GitHub APIs to verify artifact publication status. These requests are restricted to well-known, trusted domains and do not involve downloading or executing untrusted code. - [INDIRECT_PROMPT_INJECTION]: The skill processes information from commit messages and issue trackers. While this data is external, the skill handles it using structured parsing methods and human-in-the-loop validation, mitigating the risk of instructions embedded in metadata being executed without oversight.
- [DATA_EXPOSURE]: Access is limited to project manifest files (e.g.,
pyproject.toml,package.json) required for versioning. No evidence of unauthorized access to sensitive system files or credentials was found.
Audit Metadata