Agent Memory Guard is a runtime defense layer that protects AI agent memory from injection, poisoning, and leakage attacks. It scans memory operations in real-time with 7 detection layers covering OWASP ASI threat categories. Ships with a CLI scanner, REST API server, ML-based detection, and native integrations for LangChain, CrewAI, and LlamaIndex. Zero external dependencies. 59 microsecond median latency.
