crewai - 💡(How to fix) Fix Security: Add OWASP Agent Memory Guard for memory poisoning defense [1 participants]

Feature Request: OWASP Agent Memory Guard Integration

What is Agent Memory Guard?

OWASP Agent Memory Guard is the official OWASP reference implementation for ASI06 - Memory Poisoning from the OWASP Top 10 for Agentic Applications.

The Problem

CrewAI agents store and retrieve memory across sessions. A single poisoned memory entry can:

• Bias all future agent decisions
• Exfiltrate secrets stored in memory
• Cause agents to execute attacker-controlled instructions

What Agent Memory Guard Does

pip install agent-memory-guard

from agent_memory_guard import MemoryGuard

guard = MemoryGuard()

# Before writing to memory
result = guard.scan_write(memory_content)
if result.is_safe:
    crew_memory.save(memory_content)
else:
    print(f"Blocked: {result.threat_type}")

Request

Would the CrewAI team consider:

1. Adding a security note in the docs about memory poisoning (ASI06)
2. Referencing Agent Memory Guard as the recommended defense layer

References

• GitHub: https://github.com/OWASP/www-project-agent-memory-guard
• PyPI: https://pypi.org/project/agent-memory-guard/
• OWASP ASI06: https://owasp.org/www-project-top-10-for-large-language-model-applications/

This is an open-source, Apache 2.0 licensed tool with no external dependencies.