address-pr-feedback
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses established and widely-used command-line tools such as git, gh, jq, and gt for local repository management and GitHub interaction.
- [SAFE]: All network activity is restricted to GitHub's official API via the gh CLI, which is consistent with the skill's purpose and follows security best practices.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it fetches and processes external pull request comments and review data. 1. Ingestion points: SKILL.md fetches comments and reviews via the gh api. 2. Boundary markers: Absent. 3. Capability inventory: The agent can edit files, rewrite commits, and push branches using git and gh. 4. Sanitization: Absent. This surface is evaluated as safe because the ingestion of external feedback is the core functionality of the skill and the operations are performed using standard developer tooling.
Audit Metadata