policy-diff

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements strong defensive prompting requirements, such as the 'No silent supplement' rule, which prevents the agent from fabricating data or using external sources without explicit user consent.
  • [SAFE]: Data access is restricted to a specific configuration directory (~/.claude/plugins/config/...) used for storing policy indexes and matter-specific workspaces, minimizing exposure of sensitive system files.
  • [SAFE]: The skill requires mandatory source tagging for all output (e.g., [user provided], [web search
  • verify]), which helps the user identify potential fabrication or low-confidence information.
  • [SAFE]: No obfuscation, dynamic code execution, or persistence mechanisms were detected in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 03:58 PM
Security Audit — agent-trust-hub — policy-diff