Questions, answered.

No pitch — just straight answers to what team.management is, who it’s for, and why it works the way it does.

Who is this for?

People who want to know what's actually happening in their projects. If you're serious about what you build, you don't take the AI's word for it — you want to see the work.

Is this only for coding?

No. It's built into Claude Code, so it lives in your terminal — but what it manages is your process, not your code. Anything you'd hand an AI in steps — research, writing, analysis, ops — runs on the same rails, with the same enforcement and the same record.

Why does this exist?

Because AI is a convincing liar. It says "done" with total confidence, even when it cut the corner. This gives you control over what your AI is allowed to do — and a record of what it actually did.

Why do I need it?

Maybe you don't — if you read every line your AI writes and never ship on trust. Most people don't; they trust it a little more each week. This is for the moment the work matters and you can't actually see what the AI did. It gives you that back — so "done" means done, and you can prove it.

How do I know the AI really did the work?

It can't just tell you. The steps are enforced while it works, and you get a record you can check afterward. You're not trusting a summary the AI wrote about itself.

Why more than one AI model?

One model has one point of view. Different models are trained on different data, by different teams — so they notice different things and disagree in useful ways. It's the same reason you'd want a second opinion from someone who trained somewhere else. You decide where it counts — investigating a problem, reviewing the code — and team.management can put Claude, Codex, and Antigravity on it side by side. More opinions, fewer blind spots.

Can I use my own model, or a local one?

Yes. Claude Code can run a local model or another provider's, and team.management doesn't care which. The rails are in the protocol that runs around the model — the steps, the gates, the reviews — not in the model itself. Swap the model and watch them still fire.

Will it make my AI write better code?

No. It won't rescue a vague task or a missing test. It keeps the process honest — getting the spec right is still your job.

Is it free?

Yes. Open source under the MIT license — no account, no API key, no paid tier.

What do I get, exactly?

Your process on rails. The AI can't skip the steps you set, and you're left with proof of what happened — not a promise.