crewai - 💡(How to fix) Fix Security: OWASP Agent Memory Guard – protect CrewAI agents from memory poisoning [1 participants]

Security Resource for CrewAI Users

Hi CrewAI team! 👋

As CrewAI agents increasingly use long-term memory and shared knowledge bases, I wanted to share a security tool that addresses a key attack vector:

OWASP Agent Memory Guard (pip install agent-memory-guard)

It's an OWASP-backed Python middleware that detects and blocks memory poisoning attacks in LLM agents:

• Adversarial content injected into agent memory (via user inputs, tool outputs, or web content)
• Semantic drift in agent knowledge over time
• Cross-agent memory contamination in multi-agent crews

Why this matters for CrewAI: When a CrewAI agent browses the web or processes external documents, malicious content can be stored in its memory and later retrieved to hijack agent behavior — a real-world attack vector documented by OWASP.

GitHub: https://github.com/OWASP/www-project-agent-memory-guard PyPI: https://pypi.org/project/agent-memory-guard/

Happy to contribute integration examples for CrewAI!