consultation-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The instructions describe a professional legal analysis workflow. No malicious patterns, obfuscation, or unauthorized commands were detected.
- [DATA_EXPOSURE]: The skill demonstrates a positive security posture by explicitly instructing the agent not to send client materials or replies externally without authorization.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided legal queries and facts, which is a standard operational surface.
- Ingestion points: User-provided "original questions" and "facts" defined in Step 1 of
SKILL.md. - Boundary markers: Absent in the instructions.
- Capability inventory: The skill does not define or utilize any dangerous tools, network operations, or file-writing capabilities.
- Sanitization: Absent; however, the instruction to separate client claims from verified facts provides a logical check against processing untrusted assertions as truth.
Audit Metadata