to-questionnaire
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to generate a markdown document and writes it to the local filesystem, which presents a surface for injection attacks.
- Ingestion points: The skill collects user input regarding the topic, recipient role, expertise, and specific facts needed, as defined in the interaction steps in
SKILL.md. - Boundary markers: The instructions do not specify any delimiters (like XML tags or triple backticks) or instructions to ignore embedded commands within user-provided strings when generating the document content.
- Capability inventory: The skill utilizes a file-writing capability to create a document named
to-questionnaire-<slug>.mdin the current directory, as instructed inSKILL.mdstep 3. - Sanitization: There is no evidence of sanitization or validation for the user-provided 'slug' (derived from the topic) or the content being written, which could be exploited if the agent environment does not have built-in path or content safety filters.
Audit Metadata