rq
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (source code, diffs, pull requests) which could contain malicious instructions. While the skill includes a specific safeguard to treat source text as evidence only and ignore instructions embedded within it, the attack surface remains present because the agent's output is delivered to local files or hosted platforms.\n
- Ingestion points: Processes source code, diffs, repository metadata, and pull request bodies as described in SKILL.md (Workflow Step 6) and references/report-delivery.md.\n
- Boundary markers: The skill explicitly instructs the agent to ignore instructions embedded in source files (SKILL.md, Evidence and Ambiguity Rules).\n
- Capability inventory: The skill can write files to the
.agents/reports/rq/directory and post comments to GitHub Pull Requests or GitLab Merge Requests (references/report-delivery.md).\n - Sanitization: No explicit programmatic sanitization is defined, but an instructional safeguard is present to prevent the agent from executing commands found within the analyzed data.
Audit Metadata