Couldn't run: the verdict for checks that could not finish, with the step that failed.

01Definition

A verdict that refuses to guess.

Couldn't run is one of Verify's four verdicts, alongside accept, needs review and reject. It means the checks could not be completed, for example because the project would not install, the build failed for reasons unrelated to the change, or the test runner could not be detected. The comment names the step that failed.

It is a statement about the verification, not about the change.

• Install or build failed
• Runner not detected
• The failing step is named

02Why it matters

Silence is not a pass.

Many automated checks fail open: if they cannot run, nothing is reported and the pull request looks fine. That is the worst outcome, because a reviewer reads the absence of a warning as evidence.

A verifier that says “I could not check this, and here is where I stopped” gives the reviewer the truth and a place to start.

• Failing open hides problems
• Absence of a warning is not evidence
• Name where it stopped

03In Verify

Built.

Couldn't run is a built verdict. Every Verify comment also lists the individual checks that did not run, so even an accept shows exactly what it rests on. A check that did not run is never reported as passed.

Verify is not installable yet.