compliance-license-audit

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data from dependency manifests and source code files, providing a potential surface for indirect prompt injection.
  • Ingestion points: Scans various manifest files (package.json, requirements.txt, Cargo.toml, etc.) and repository source code.
  • Boundary markers: Absent; the skill does not instruct the agent to distinguish between file content and instructions during the audit.
  • Capability inventory: Utilizes Read and Grep tools and has file-write permissions for the /docs/compliance/ directory.
  • Sanitization: No evidence of sanitization or escaping of file content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 11:39 AM
Security Audit — agent-trust-hub — compliance-license-audit