Cluing has two types of agents inside one collaborative workspace:
Chat agents
Autonomous agents with Claude Code capabilities
Chat Agents
Think of a Chat Agents as your usual AI chat, only smarter about your work.
It can draft, summarize, explain, or brainstorm - and it does it against the Topics you've saved, the Skills you've built, and the tools you've connected.
Cluing Agents don't start from zero. They already know your research, your Topics, and the workflows you and your team have built, so every answer lands in the right context.
Autonomous Agents
Autonomous Agents are the ones you send off to do real work.
They can write code, spin up custom integrations, and keep going long after you've closed the tab - the kind of tasks you'd normally hand to a teammate, not to a chat window.
A few things set them apart:
They run in the cloud, not on your laptop, so they keep working while you're away from your desk
They already know your research, your Topics, and the workflows your team has put together
They can build their own connections to the tools a job needs, instead of waiting for you to wire them up
They live in your workspace, so anyone on the team can pick up the thread, kick off a new run, or hand work off