validate-idea
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill package consists exclusively of markdown instructions and reference quotes. It contains no functional code or executable scripts.\n- [NO_CODE]: No scripts (Python, Node.js, Shell) are included in the skill package. Its functionality is delivered through natural language instructions for the agent.\n- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze user-provided documents, creating a surface for potential indirect prompt injection.\n
- Ingestion points: User-attached documents such as PRDs, pitches, survey notes, and repositories as specified in Step 1 of
SKILL.md.\n - Boundary markers: Absent.\n
- Capability inventory: None. No subprocess calls, network operations, or file-write capabilities are defined in the skill content or its reference files.\n
- Sanitization: Absent.
Audit Metadata