to-questionnaire
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates entirely within the agent's standard conversational context and instructions to write local Markdown files. It contains no commands for network access, remote execution, or sensitive data harvesting.
- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill ingests user input to define the recipient's role, expertise, and the required information gaps (SKILL.md).
- Boundary markers: The instructions provide a specific
<questionnaire-template>structure to wrap the user-provided content. - Capability inventory: The skill utilizes the agent's capability to write content to a local file named
to-questionnaire-<slug>.md. - Sanitization: No explicit content sanitization is defined, but the risk is minimal as the output is a static Markdown document intended for human review.
Audit Metadata