to-questionnaire
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates as a structured prompt for document generation. It does not perform any suspicious network activity, access sensitive local configuration files, or attempt to bypass security controls.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided input to populate a Markdown template, creating a surface for potential data interpolation. However, since the skill's purpose is document creation and it does not execute the resulting content or pass it to high-privilege tools, this represents a safe and standard implementation of its functionality.\n
- Ingestion points: User conversation regarding the questionnaire's recipient and the specific facts or decisions required (SKILL.md).\n
- Boundary markers: The template uses Markdown headers and blockquotes to organize the document, though it lacks specialized delimiters for user-provided strings.\n
- Capability inventory: The agent is instructed to write the generated questionnaire to a local Markdown file using a slug-based filename.\n
- Sanitization: No explicit sanitization logic is provided in the instructions, relying on the agent's standard text generation capabilities.
Audit Metadata