prompt-section-design
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [PROMPT_INJECTION] (SAFE): No malicious patterns or behavioral override instructions were detected. The skill provides a framework for document structure rather than controlling agent logic.
- [COMMAND_EXECUTION] (SAFE): The skill utilizes read-only tools (Read, Grep, Glob) intended for accessing reference documentation. No arbitrary code execution or shell access is present.
- [DATA_EXFILTRATION] (SAFE): No network operations (curl, wget, fetch) or access to sensitive paths (e.g., .ssh, .aws) were found.
- [INDIRECT_PROMPT_INJECTION] (SAFE): While the skill defines how to ingest user-provided prompt descriptions (Category 8 surface), the current capability set is restricted to read-only operations and display, resulting in negligible risk.
Audit Metadata