Blog

Field notes on sovereign AI, agent teams and the RiseLab products — what we shipped, how it works and what it changes for government contractors, sales and operations teams.

All RiseLab Platform

Abstract illustration of a layered filing archive with one highlighted folder being inspected by a magnifying glass, representing a poisoned entry in AI agent memory

RiseLab Platform

AI Agent Memory Security: Stop Poisoned Memory From Persisting

Once agents remember things between sessions, a single poisoned entry can misdirect them for weeks. Here is how to approach AI agent memory security, and which RiseLab controls help around the edges.

October 2, 2026