changelog
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from GitHub PR titles and descriptions, which represents an attack surface where malicious text could attempt to influence the agent's output.
- Ingestion points: PR titles and descriptions are fetched using
gh pr listandgh pr viewinSKILL.md. - Boundary markers: Absent. There are no explicit instructions for the agent to use delimiters or ignore instructions embedded within the PR data.
- Capability inventory: Subprocess calls to
git,gh,grep, andpython3are used withinSKILL.md. - Sanitization: Absent. The skill relies on agent filtering logic rather than programmatic sanitization of the external text.
- [COMMAND_EXECUTION]: The skill executes several standard shell commands to manage the changelog and verify the output.
- Uses
gitfor tag management, history retrieval, and diffing. - Uses
gh(GitHub CLI) to pull metadata from the project's repository. - Executes
python3 -m http.serverto host a local preview of the site for manual verification via screenshots.
Audit Metadata