proof-of-work
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process data from external, untrusted sources such as task descriptions and pull request bodies.
- Ingestion points: Task/issue descriptions and PR body content (SKILL.md).
- Boundary markers: None specified.
- Capability inventory: No tools or scripts are defined or invoked by this skill.
- Sanitization: No sanitization or filtering logic is present.
- [NO_CODE]: The skill consists entirely of instructional markdown and documentation for the agent's workflow. It does not contain scripts, binaries, or automated tool definitions.
- [SAFE]: No sensitive data exposure, command execution, or remote code patterns were detected. The instructions follow standard software development practices for change verification.
Audit Metadata