Browser agent keeps going after it already called done

In headless mode with a custom structured-output controller, the run never ends after the task is complete.

Difficulty

Moderate

Time

~15 min

Steps

Fixed in

—

Last checked

2026-10-15

Unverified: we could not confirm from the sources that a fix shipped. Source: issue closed as completed, no linked fix. Last checked 2026-10-15.

Symptom

In headless mode with a custom structured-output controller, the run never ends after the task is complete.

Evidence what you'll see

Started in version 0.1.40, earlier versions finished
A valid done action with structured output appears in the log

Cause

A change to last-step handling in 0.1.40 broke detecting the terminal state when a custom output model was used.

Fix steps

0 of 3 done

Step 1Try 0.1.45 or later, as the maintainers advised. We could not confirm a linked fix, so test it on your version.
Step 2If it still happens, cap max_steps and check after each step for a done result.
Step 3Open a fresh issue with a minimal repro if the latest version still loops.

Your check marks stay in this browser only.

Verify the fix worked

The run exits right after the done action, and the structured result is returned.

Did this fix it?

Sources and credit

Summarized in our own words from these public reports. Credit goes to their authors.

browser-use issue #951 ↗

Provided as is, without warranty. Test in a non-production environment and back up first. Security-related steps: verify against the official docs. Terms · Request removal or attribution change