mistral-security-basics
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill promotes secure secret management by providing implementation examples for environment variables and Google Cloud Secret Manager, discouraging the practice of hardcoding API keys.
- [SAFE]: Implements defensive measures against prompt injection by providing functions to sanitize user input for common malicious patterns and enforcing length constraints.
- [SAFE]: Includes output sanitization logic using regular expressions to redact personal identifiable information (PII) like SSNs and email addresses, as well as potential script tags from model responses.
- [SAFE]: Provides a security audit utility that allows the agent to check local files for hardcoded credentials, aiding in proactive security posture management.
Audit Metadata