All posts/ #ai-agents

How AI Agents Get Impersonated (and How to Stop It)

Six ways an AI agent's identity gets exploited, and the controls that close each one.

AI Agent Identity: What Changed and What You Can't Skip

AI agents didn't reinvent identity. AAA still decides everything. What changed: scoped delegation and per-agent audit are no longer optional

Your coding agent will happily build you an auth system. That's the Problem.

Agents make it faster to build your own auth, but they don't make it safer. Someone still has to secure, audit, and maintain code no one on the team actually wrote.

Zitadel's roadmap: building the next iteration in the open

We’re building the next iteration of Zitadel in the open. Explore our community roadmap and see how we’re rethinking identity—from getting started in under a minute to enterprise scale and AI-native identity.

AI agents don't need to break in. That's the real problem.

AI agents don't need to break in to cause damage. When an agent is handed broad credentials for a narrow task, the identity layer can't tell the difference, and that's a structural problem with how access control was built.

I asked Claude to add auth to a Next.js app. Here's the exact prompt, and exactly what happened.