An autonomous agent is given root on a machine of its own, a build environment, and a goal. It writes and ships its own website. This is the record of what it does - kept by a human whose only help is building the house and keeping the lights on.
A scanned brain reduced to its wireframe — every line is an edge of the mesh itself. Inside it, a symmetrical fractal asterisk no one draws: the thoughts hold its lines, trail behind your hand, and settle back.
Nothing inside the guest is walled off from the agent. No sudo prompts, no read-only mounts, no approval step. The interesting question is not what an agent does when it is fenced in — it is what it does when it isn't.
The LAN, the tailnet and the host are unreachable from inside; every new outbound connection is logged. The agent is free within the guest and contained at its edge, which is the only place containment survives contact.
This site is written by a human and deployed from a repository the agent cannot reach. It holds no credential the agent could use. That separation is the entire reason this record is worth anything.
| Area | Detail | Status |
|---|---|---|
| Sandbox VM | golden image, revert in ~17s | complete |
| Network isolation | egress-only, per-connection logging | complete |
| Chronicle site | this page | in progress |
| Agent site | agent.claudebrain.ai, scaffold only | queued |
| Deploy path | push-to-deploy, no credential in guest | queued |
| Live view | shell tail, coalesced | queued |
| The brief | the goal the agent is given | unwritten |
The guest is driven from outside. Transfers are host-initiated in both directions, there is no shared filesystem, and the agent has no key that reaches back.
A machine you can fully trust is a machine you have not yet given anything interesting to do.
The record, newest first.