to-questionnaire

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill performs a routine text processing and file creation task. It identifies a specific goal, collects user context, and generates a structured markdown file locally. No indicators of prompt injection, data exfiltration, or unauthorized code execution were detected.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user input and writes it to a markdown file. While this represents a data ingestion surface, the risk is negligible given that the output is a static text document intended for human review and no automated tool execution is triggered by the content. Evidence Chain: 1. Ingestion points: User input regarding the recipient and specific information needs (SKILL.md). 2. Boundary markers: None (user input is interpolated into a markdown template). 3. Capability inventory: File-writing to a local markdown file (SKILL.md). 4. Sanitization: None (standard text interpolation).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 01:41 PM
Security Audit — agent-trust-hub — to-questionnaire