crypto-guardrails

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements strict security invariants for cryptographic operations, specifically targeting the Web Crypto API.
  • [SAFE]: It mandates high-strength parameters such as AES-256-GCM and a minimum of 600,000 PBKDF2 iterations, which significantly exceeds standard recommendations.
  • [SAFE]: It correctly identifies and forbids common cryptographic pitfalls like IV reuse, hardcoded salts, use of non-cryptographic PRNGs (Math.random), and insecure cipher modes (ECB).
  • [SAFE]: The only external dependency permitted is @scure/bip39, which is a recognized and reputable library for mnemonic generation.
  • [SAFE]: The skill includes a detailed checklist and canonical code patterns to ensure developers/agents follow a secure-by-default approach.
  • [SAFE]: It includes rules to prevent information leakage through error messages or logs.
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 10:19 PM
Security Audit — agent-trust-hub — crypto-guardrails