fto-triage

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: Accesses local configuration files in the user directory.
  • The skill reads from ~/.claude/plugins/config/claude-for-legal/ip-legal/CLAUDE.md. Accessing files within the home directory is a sensitive operation that could expose configuration data if the skill is compromised.
  • [PROMPT_INJECTION]: Vulnerable to indirect prompt injection via user-supplied technical content.
  • Ingestion points: Technical architecture, technical details, and patent documents are ingested from the user during the intake process.
  • Boundary markers: No specific delimiters or instructions to ignore embedded commands are present to protect the prompt from malicious content in technical descriptions.
  • Capability inventory: The skill possesses file-writing capabilities (Instruction 7) and potential network access via patent search integrations (Instruction 4).
  • Sanitization: There is no evidence of sanitization, escaping, or validation of user-provided technical data before it is processed or written to output files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 01:02 AM
Security Audit — agent-trust-hub — fto-triage