design-ai-department
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is purely instructional, providing templates for AI department design and governance. It does not contain executable code, network calls, or requests for sensitive file access.
- [INDIRECT_PROMPT_INJECTION]: The skill includes an ingestion surface for untrusted workflow data. 1. Ingestion points: User-provided workflow and system descriptions in SKILL.md. 2. Boundary markers: The skill references a packet contract for governing constraints but lacks explicit delimiters for user input. 3. Capability inventory: No executable capabilities, network operations, or file-writing tools are defined; the skill is limited to generating design text. 4. Sanitization: The skill explicitly requires defining adversarial tests for prompt injection and sensitive input handling.
Audit Metadata