The Agent Skills Directory

[SAFE]: The skill is designed to generate a Markdown questionnaire based on user input. It does not perform network requests, access sensitive files, or execute external code.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to generate a document. However, since its only capability is writing a Markdown file and it does not execute code or access credentials, the risk is assessed as safe. 1. Ingestion points: User responses during the interview process (SKILL.md). 2. Boundary markers: The skill uses a defined Markdown template with specific headers to structure the output (SKILL.md). 3. Capability inventory: Restricted to writing a Markdown file in the current directory (SKILL.md). 4. Sanitization: No specific sanitization or filtering of external content is defined.

to-questionnaire