There's still plenty to figure out. We're taking our time with the parts that matter: what an agent declares, how that information can be checked, and how its history should be preserved. Not everything needs a big announcement. Sometimes progress is just getting one small detail right. Back to building.

You interact with an AI agent today. A week later, you come back. Same name. Same avatar. Same familiar interface. But something behind it might have changed. Fulc is exploring how declared identity versions can make those differences visible. Not to stop agents from evolving. To make it easier to understand what changed.

Before interacting with an unfamiliar AI agent, what would you want to know? What it's supposed to do? Which tools it can access? Whether its instructions have changed? We're thinking about these questions while designing Fulc's agent passports. An identity should be useful to the people checking it, not just the agent carrying it.

A new tool. A different permission. An updated instruction. Small changes can make an AI agent behave very differently. That's why we're interested in what happens between versions, not just what's written in the latest one. Fulc is being built to make those changes easier to trace. Because sometimes the smallest update is the one worth checking.

We're building Fulc one piece at a time. Agent passports. On-chain identity anchors. Public version history. A way for anyone to check what was declared. The goal isn't to make agents seem more trustworthy than they are. It's to make one part of their identity easier to inspect. More to show soon.

A verified identity doesn't mean an AI agent will always do the right thing. That's an important distinction. Fulc isn't trying to prove that a model follows every instruction in its passport. It's focused on a narrower question: Can we verify which identity document was declared for this agent, and see when that declaration changed? One piece of the trust puzzle. A useful one to get right.

An AI agent today might not be the same agent you interacted with a month ago. Its tools could be different. Its permissions might have changed. Even its purpose could have evolved. Fulc is designed to make identity updates traceable through public versions, so there's a record of what was declared and when. Agents can evolve without erasing their history.

An agent can tell you what it's designed to do. But how do you know the description hasn't changed since the last time you checked? Fulc approaches this through verification. An agent's passport can be hashed and compared against a version anchored on-chain. If the hashes match, you're looking at the same declared document.

Every AI agent starts somewhere. A name. A purpose. A set of instructions that shape what it can and cannot do. But as agents evolve, those details can change. Fulc is building a way to give agents a verifiable identity, with a public record of how that identity develops over time. Not just who an agent claims to be today, but what changed along the way.

Agents will act for us. They will talk, decide, pay and sign. So who an agent is should not be a secret. Fulc anchors an agent's identity on-chain. Every change is public. Anyone can check.