Find the bugs you didn’t write Scenarios for

Give Deep Scan an outcome to investigate. Testers work through the feature in parallel, keep going after the first bug, and return each potential bug with the evidence to reproduce it.

Read the docs

Beyond your Scenarios

Adversarial testers look for problems in a feature, including cases nobody wrote down.

Evidence you can act on

Each potential bug comes with expected and actual behavior, reproduction steps, and a recording.

Confirmed on rerun

Rerun a finding in a fresh browser to see whether it repeats before anyone spends time on it.

1

Describe an outcome

Give Deep Scan an observable goal and say what the QA account may change and what to preserve. Start from a finished discovery’s mapped flows, or describe the area yourself.

2

Testers split the work

Coordinators hand independent charters to testers that work in parallel, each in its own browser. A tester that finds a bug records it and keeps investigating what it hasn’t covered.

3

Review the evidence

Each potential bug has what was expected, what happened, the page, and the steps to reproduce it, with a recording of the tester’s browser.

P2

TL2 directory count contradicts its empty unfiltered Members listing

/g/trust_level_2

Expected
The directory’s member count should reconcile with the unfiltered Members listing.
Actual
Directory showed 2 for trust_level_2, but Members said “There are no members in this group” with blank search before and after reload.

Reproduction steps

  1. 1Authenticate with the granted password handles and confirm Admin/Welcome back.
  2. 2More > Groups; note trust_level_2 count 2.
  3. 3Open Trust level 2.
  4. 4Clear username or email search and reload; observe the empty Members listing.
4

Rerun in a fresh browser

Check whether a reported issue repeats before you act on it, then turn the important cases into regression Scenarios.

5

See what was covered

A completed scan shows what was investigated and what was blocked. It doesn’t guarantee that every bug was found, so gaps are listed first.

What you get back

The Deep Scan report

Every potential bug comes back with what was expected, what happened, the steps to reproduce it and a recording, beside an honest account of what the testers covered and what blocked them.

What you can do with it

  • Rerun in a fresh browser

    A new browser replays the reproduction steps. The bug is marked Seen again on rerun, or Not seen on rerun, before anyone spends time on it.

  • Send it to your coding agent

    Open Claude, Codex or Cursor with the bug, the expected and actual behavior, the reproduction steps and a link back to the evidence.

  • Keep it as a regression Scenario

    Write the important cases as Scenarios so they run again on every pull request once the fix lands.

  • Watch the moment it happened

    Jump the recording to the bug’s evidence and see the page exactly as the tester saw it.