advise-legal-work-skills

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns detected. The skill functions solely as a set of instructions for a 'Legal Work Skill Advisor' and does not contain scripts, binaries, or automated shell commands.
  • [SAFE]: The skill incorporates domain-specific safety measures, explicitly instructing the agent to require human legal review for conclusions and to avoid presenting AI-generated output as final legal advice.
  • [SAFE]: Privacy and confidentiality are addressed through instructions to protect privileged material and minimize exposure of confidential data during the design process.
  • [SAFE]: The workflow emphasizes a 'read-only inspection first' policy, which ensures the agent asks for explicit user permission before making edits, running commands, or publishing content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 07:27 PM
Security Audit — agent-trust-hub — advise-legal-work-skills