OWASP Agent Memory Guard

Runtime middleware that detects and blocks memory-poisoning attacks against AI agents' persistent memory.

Pluginfreeglobal

OWASP Agent Memory Guard is a free, open-source runtime middleware that screens every read and write to an AI agent's persistent memory, detecting prompt injection, PII leakage, and tampering before it can survive a context reset, with drop-in support for LangChain, OpenAI Agents SDK, AutoGen, mem0, and CrewAI. It's an officially recognized OWASP Incubator Project used by Microsoft's agentic AI security research. It's aimed at teams building persistent-memory AI agents who need runtime protection against memory-poisoning attacks.

Website

Full match profile

Behind the summary, Matchbox keeps a richer profile of OWASP Agent Memory Guard - the signals our matcher actually reads to decide when to surface it. It stays private; claim the listing to see and control it.

Problem & pain-point mapping
Who we surface it to (audience fit)
What it's a strong alternative to
Trust & credibility signals

Claim OWASP Agent Memory Guard to unlock →

Try Matchbox with your own problem

Describe what is not working - we’ll show you whether OWASP Agent Memory Guard (or something else) actually fits.

Try a problem