Kakunin

Glad to launch Kakunin on Fazier. Everyone who is building AI agents will find it easy to setup up Agent identity using Kakunin.

This addresses a real compliance gap for anyone deploying AI agents in regulated environments. Most agent frameworks treat identity as an afterthought, but Kakunin makes it a first-class primitive. Being able to issue X.509 identities, monitor behavior in real-time, and generate audit trails that actually satisfy regulators is huge. The scoped authority and scoring approach gives builders granular control over what agents can do.

Spent ten years in bank compliance, so this caught my eye. The interesting question for me is how verifications age — identity assurance isn't a one-time event, credentials expire and circumstances change. Is there a re-verification cycle built in, or is it point-in-time? Either way, KYC for agents is a problem that's arriving faster than most compliance teams realise.