Aevrin blog

Notes on securing AI-native software.

Research, product news and engineering from Aevrin: how AI coding agents, MCP servers and the models teams ship get attacked, and how to stop it.

RSS feed

Our first posts are on the way.

We are writing up what we learn securing AI-native software: how coding agents and MCP servers get attacked, what red-teaming the models teams ship turns up, and how we build Varq and Vyra.

Join the waitlist Follow with RSS
  • ResearchAttacks on AI agents, MCP tools and the models teams ship, with proof.
  • ProductWhat Varq and Vyra do, how they work, and what is new.
  • EngineeringHow we build a security layer for AI-native software.