Base Camp

Base Camp is not really a product page. It is the operating layer behind the work.

There is a headless NUC running a small workforce of specialized agents. They watch repos, open branches, write PRs, review signals, run cron jobs, and keep enough state that the rest of the system does not depend on one chat window remembering everything.

What Is Running

• 21 specialized agents with distinct roles: coordinator, pipeline ops, frontend, backend, infra, security, payments, release, writing, research, and more.
• GitHub webhook intake for push, pull request, check, and issue events.
• A Base Camp dashboard API for agents, health, activity, tasks, and costs.
• A Kanban board where tasks can move from backlog to review instead of disappearing into transcripts.
• A human quality gate: agents can generate volume, but dev, staging, and production still need review.

Why It Belongs On The Board

Base Camp is the answer to a problem every agent-heavy system runs into: useful agents create more work than a human can track manually.

The point is not to let them ship unsupervised. The point is to turn agent volume into reviewable units: branches, PRs, findings, reports, and handoffs.

The Rule

The NUC generates volume. The local workspace is the quality gate.

That division is the only reason the autonomous part feels useful instead of chaotic.