warden
Docs · overview

How warden works

A high-level look at the loop — enough to know what you're installing, before you read a single config file.

The loop

1

Intent

You write the issue — what should change, and why.

2

Agent

Your chosen coding agent works the task inside an isolated workspace.

Gate

Every change is checked against the scope you declared, and scanned for secrets. Fails closed, not open.

4

Validate

Your own test suite runs against the result — no separate CI to wire up.

5

Draft PR

warden opens a draft pull request. It never approves, never merges — that's always yours.

What you control

Scope

Which files and paths a run is allowed to touch.

Agent

Which coding agent runs the task — claude-code, codex, or your own.

Validation

The commands that must pass before a draft PR opens.

Budget

A cost cap per run — warden stops before it's exceeded.

Where it runs

Runtimea single secure container per run Source visibilityyour repo never leaves your infrastructure Outputdraft pull request only Requiresa repo, an agent, a test command
warden never merges. Draft PRs are the only thing it produces — review and approval stay with your team.

Want to try it on a real repo?

Request early access