test-design
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill instructions follow a logical workflow for documentation generation.\n- [PROMPT_INJECTION]: While the skill ingests external data from task files, it does not possess dangerous capabilities that could be exploited via indirect prompt injection. The risk is minimal as it only produces static documentation. 1. Ingestion points: task_file and risk_profile_file. 2. Boundary markers: None. 3. Capability inventory: Local file read/write only. 4. Sanitization: None.\n- [EXTERNAL_DOWNLOADS]: CI/CD templates reference trusted services like GitHub Actions and Codecov, which are standard in software development.
Audit Metadata