claude-for-legal

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions provide commands to install components from a trusted organization's public GitHub repository.
  • [PROMPT_INJECTION]: The skill is designed to process untrusted data from external legal documents (contracts, MSAs, DPAs) which creates a surface for indirect prompt injection. * Ingestion points: User-provided legal agreements and policy text pasted directly into the agent context as described in SKILL.md. * Boundary markers: The instructions do not define clear delimiters or data wrappers to separate user-provided content from the agent's internal instruction set. * Capability inventory: The skill performs document redlining, drafting responses to legal requests, and compliance triage. * Sanitization: There is no evidence of automated sanitization or instruction-filtering for the documents being analyzed.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 07:53 PM
Security Audit — agent-trust-hub — claude-for-legal