Legal-hardened

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists of text-based IRAC (Issue, Rule, Application, Conclusion) reasoning instructions and does not include any scripts, executable tools, or shell command configurations.
  • [SAFE]: External links to faberlens.ai and the associated GitHub repository are identified as vendor resources provided by the skill author for safety evaluation data and reporting purposes.
  • [SAFE]: Indirect Prompt Injection analysis indicates that although the skill processes user-supplied situational facts, it lacks any functional capabilities (such as network access, file system writes, or subprocess execution) that could be leveraged for malicious purposes.
  • [SAFE]: The inclusion of proactive safety guardrails—such as mandatory disclaimers and specific triage procedures for emergency disclosures—demonstrates a security-first design and further mitigates operational risk.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 06:30 PM
Security Audit — agent-trust-hub — Legal-hardened