aia-generation
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- Indirect Prompt Injection Surface: The skill processes user-supplied information about AI systems to generate structured assessments. This ingestion point is a potential vector where instructions could be embedded in the provided data to influence output.
- Ingestion points: The conversational intake process defined in Step 2.
- Boundary markers: The skill uses a structured markdown template and a detailed citation-tagging system (e.g., [settled], [verify]) to differentiate between model knowledge and primary sources.
- Capability inventory: The skill is instructed to read policy files and write output documents to specific matter-slug directories.
- Sanitization: The instructions do not specify explicit sanitization or filtering of the user's input strings.
- No Executable Code: The skill is composed entirely of markdown instructions (SKILL.md) and does not include or call external scripts, binaries, or build processes, which minimizes the attack surface relative to code-based execution.
Audit Metadata