pr-creation
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- COMMAND_EXECUTION (LOW): The skill uses standard development tools for branch management and API interaction. Evidence: SKILL.md utilizes
git fetch,git rebase,git push, andgh pr createcommands. - PROMPT_INJECTION (MEDIUM): Vulnerable to indirect prompt injection where malicious content in source code, commit messages, or issue descriptions could influence the documentation generated for pull requests.
- Ingestion points: Implementation details, commit history, and verification results are used to populate the PR body.
- Boundary markers: None are specified in the templates to separate agent instructions from processed data.
- Capability inventory: Includes the ability to push to remote branches and create Pull Requests via the GitHub CLI.
- Sanitization: No sanitization or validation of the ingested repository content is performed before it is sent to the GitHub API.
Audit Metadata