leaving-pr-comments
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and respond to content within GitHub Pull Requests, which includes untrusted data from external contributors (PR descriptions, code comments, and code changes). This represents a vulnerability surface for indirect prompt injection where malicious instructions embedded in a PR could attempt to influence the agent's output or actions.
- Ingestion points: The agent reads the contents of GitHub Pull Requests (SKILL.md) to generate comments.
- Boundary markers: Absent. The instructions do not specify the use of delimiters (like triple backticks or XML tags) or provide warnings to ignore instructions found within the PR content.
- Capability inventory: The agent has the capability to write and post comments to GitHub and resolve review threads.
- Sanitization: The skill includes an instruction to manually check comments for secrets and private details before posting, but lacks technical sanitization or escaping mechanisms for the ingested data.
Audit Metadata