Concourse runs coding agents behind gates that prove the work — every change arrives as proof you can check, not a promise.
Every change is gated until it proves itself — the plan, the tests, the evidence laid out in one place. The deterministic checks — tests, coverage, CI, secret-scan — hard-stop anything that fails before it reaches you. So your call on each change is a glance at the proof, not a forensic read of the diff.
When a run is ready to merge, Concourse shows you proof, not a promise: the satisfied acceptance gate, every changed file, the tests behind each criterion. Approve in one click — or open the diff and look closer.
Most tools collapse “the step ran” and “the result is acceptable” into one green check. Concourse keeps them apart. Deterministic facts can hard-stop a run; a model’s judgment can inform and loop it — but never gets the last word on a merge.
Tests, coverage, CI, secret-scan. Deterministic and binary — they pass or block the merge. No model in the loop, no negotiation.
A reviewer model scores the work and explains its reasoning. A low score sends the run back to revise — it never approves a merge on its own.
Every run moves through the same phases — and stops at any one the moment it needs you. Pause, inspect the evidence, send it back a step, resume where it left off. Nothing is a black box.