open-pr
Pass
Audited by Gen Agent Trust Hub on Mar 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
bashtool to executegit status,git push, andgh pr create. These operations are used as intended to manage branches and initiate the Pull Request process. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface (Category 8) because it populates PR fields with unvalidated data from the local repository and conversation history.
- Ingestion points: Data is pulled from
git diff,git log, and the conversation context. - Boundary markers: There are no specific delimiters or instructions provided to the agent to treat this gathered content as untrusted.
- Capability inventory: The skill has access to
bashfor shell commands andReadfor file access. - Sanitization: There is no explicit sanitization of the generated title or body before interpolation into the
ghcommand. However, the use of the--webflag ensures that the PR is not submitted automatically, requiring a human to review and approve the content in a browser.
Audit Metadata