to-questionnaire

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill functions as a surface for indirect prompt injection by taking user input and writing it directly to a markdown file (to-questionnaire-<slug>.md).
  • Ingestion points: User answers concerning the questionnaire recipient and the specific information gaps are ingested in SKILL.md.
  • Boundary markers: The markdown template does not include boundary markers or explicit instructions for downstream agents to ignore embedded text within the questionnaire fields.
  • Capability inventory: The skill directs the agent to perform file-system write operations in the current working directory.
  • Sanitization: The instructions lack any requirement for the agent to sanitize, escape, or validate user input before writing it to the filesystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 02:17 AM
Security Audit — agent-trust-hub — to-questionnaire