policy-monitor
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFE
Full Analysis
- Local File Access: The skill interacts with configuration files located in
~/.claude/plugins/config/claude-for-legal/privacy-legal/CLAUDE.md. This access is used to store state information, such as the last sweep date and regulatory footprint, which is necessary for the skill's monitoring functionality. - Indirect Prompt Injection Surface: The skill scans external output files such as Privacy Impact Assessments (PIAs), Data Processing Agreements (DPAs), and triage results. Because these files originate from external processes, they could theoretically contain adversarial content designed to influence the agent's output.
- Ingestion points: Files located in the user-defined outputs folder path (SKILL.md).
- Boundary markers: Not explicitly defined; the skill relies on the structure of the input documents (e.g., PIAs, DPAs).
- Capability inventory: Reading local configuration files and authoritative policy documents to generate drafted legal text updates.
- Sanitization: No specific sanitization or filtering of input text is mentioned beyond the standard processing capabilities of the model.
Audit Metadata