Skip to content

Security: AFunLS/self-evolving-agent-patterns

Security

SECURITY.md

Security Policy

⚠️ Important: AI Agent Security

This package creates AI agents that can execute shell commands and modify files. This is by design — agents need real tools to be useful. However:

  • Never run agents with elevated privileges (no sudo, no root)
  • Use in sandboxed environments when possible (containers, VMs)
  • Review the default tools in src/agent_patterns/tools.py before use
  • The self-evolution feature (self_evolving=True) allows the agent to modify its own prompt file — review changes periodically

Reporting Vulnerabilities

If you discover a security issue, please email hello@tutuoai.com instead of opening a public issue.

We'll respond within 48 hours and work with you to address the issue.

Supported Versions

Version Supported
0.1.x ✅ Current

There aren’t any published security advisories