Skip to content

Agents

Home

AI Team

Agents

An Agent is an autonomous AI worker. Each one owns a domain and gets real work done, so you do not have to.

What an Agent is

Think of an Agent as a tireless specialist on your team. You give it a domain, and it owns that domain completely.

It does not just answer questions. It plans, builds, fixes, and reports back. It works in the background while you live your life.

The whole point is to give you your time back. Agents handle the work so you can be off the screen, with your people.

Your Clone and Claude

Two pieces sit in front of the Agents, and they matter most to you day to day.

Claude is the chat surface. It is the window where you talk, ask, and explore. You only ever talk to Claude.

Your Clone is your personal Agent, one per person. It knows you, speaks like you, and acts on your behalf. When you tag it in, it goes to work and dispatches the other Agents as needed. Tag-In is simply how you hand your Clone a job.

The signature move: you tell Claude what you want, you tag in your Clone, you close the window. Results come back when they are ready.

The team of three

Behind your Clone stand three system Agents. Each owns one clean lane. The design is a few powerful Agents with world-class Skills, not a sprawl of narrow bots. A Skill is a packaged set of instructions that gives an Agent a capability.

AgentWhat it does
The activatorTakes a person or company from zero to one. It builds and equips your Clone, then hands the live Clone over for ongoing care.
The builderBuilds and rebuilds the system's parts: its Skills and its Agents. It is the only Agent that creates new Agents.
The maintainerMaintains and refines everything that already exists, and runs the daily cycle that keeps the system honest.

The first two read as a pair: one starts something new, the other keeps it sharp once it is live.

The simple division of labor

The split is easy to remember. Build something new, or tend something that already runs.

Net-new work, a fresh Skill or a fresh Agent, goes to the builder. Rebuilding a part from the ground up also goes to the builder.

Refining and maintaining what already exists goes to the maintainer. The question that settles any gray area: am I creating a part, or tending one that is already live?

Growing the team

When a new capability is needed, the first move is to add a Skill, not spawn another narrow bot. Skills are how the system grows. Agents are how it operates.

For a whole new domain, a specialist Agent team can be added. These come as pairs, modeled on the same split: one half takes the domain from zero to one, the other half refines what is live. They are always built by the builder.

The hard rules

A few rules keep the team clean and predictable.

What runs today

We are honest about what is live versus what is designed. The daily cycle, the maintainer's heartbeat, runs daily today, triggered through a thin, temporary scheduling bridge.

The system is built to go continuous when live connectors and scheduling arrive. The engine is ready for it now. Until then, the loop runs daily, and we say so plainly.

The platform underneath is built on Anthropic's platform. Your time is the metric that matters most: every cycle is meant to give more of it back to you.