Browser verification & evidence
Inspect actual browser assertions, screenshots and logs from an isolated application check.
WHEN IT HELPS
Put it to work.
Verify a small frontend behavior against selected application files and see what actually happened.
Before you begin
- An operator-enabled isolated browser runtime.
- Exact selected application files and an approved start command/scenario in an Agent task.
WALK THROUGH IT
See the workflow, step by step.
Synthetic sample content, with simplified screens. No model call, upload, account change or paid task runs here.
Open Browser evidence from Engineering. Read the runtime status; configuration presence is not a guarantee that a run will succeed.
In a supported Agent task, select the exact files and describe the action/assertion. Review files, command and browser steps before approval.
Open the saved artifact to review the screenshot, DOM/log evidence and assertion outcome. A failed assertion should stay failed until a new check proves the fix.
Browser evidence / Runtime
A REQUEST TO START WITH
Use browser_verify with index.html and styles.css. Open /, click #submit, and verify that #result contains Saved. Show the actual evidence.
THE USEFUL RESULT
What you should leave with.
A recorded isolated browser check and private evidence artifacts, when the runtime is available.
Know the boundaries.
- This runtime does not control your personal browser, authenticated external accounts or arbitrary remote staging sites.
- A visual example here is illustrative; it is not a screenshot from a completed live verification.
IF SOMETHING DOESN’T WORK
Runtime disabled or no artifact exists.
Ask an operator to provision the reviewed runtime. Then inspect the real task error; do not report browser verification if the task produced no evidence.
Contact support