Open-source memory poisoning defense for AI agents — detect, sanitize, quarantine, and block unsafe memory before it influences an agent.
Replies
Best
Maker
📌
Hi HN, I built Memgar, an open-source memory layer for AI agents.
The goal is to let agents persist, retrieve, and version useful context across runs instead of starting from scratch every session. It is intended for local/dev workflows where memory needs to be inspectable and controllable.
I would appreciate feedback on the design, API, and whether the Git-like approach to agent memory makes sense.
Yours comments and contributions would make me happy.
Replies