severability-general

Pass

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: SAFE
Full Analysis
  • [Indirect Prompt Injection] (SAFE): The skill processes user-provided contracts, which represents a surface for indirect prompt injection.
  • Ingestion points: User uploads or pastes contracts as described in SKILL.md.
  • Boundary markers: Absent; the agent does not use specific delimiters to isolate user text from its own instructions.
  • Capability inventory: No subprocess calls, network operations, or file system write capabilities found in any scripts.
  • Sanitization: Absent.
  • Conclusion: While the surface exists, there are no dangerous capabilities for an attacker to exploit via injected text.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 21, 2026, 11:46 AM
Security Audit — agent-trust-hub — severability-general