The instance row locator used Locator.GetByText scoped to the panel, which
bypasses the [data-test='flow-instance-row'] contract InstancePicker rows
are documented to carry. Match through the data-test attribute instead,
with an exact-text regex since the test only knows the instance name (not
its id, unlike Debug View's tests).
Claude-Session: https://claude.ai/code/session_01NcMG6xXdsJbmD2wYAgkK29
Flow Debugger's script-entry instance selection and Debug View's instance
selection both go through the shared InstancePicker combobox now instead of
a flat <select>. Update the tests to type into the {prefix}-input, wait for
the matching {prefix}-row, and click it; keep the <select>-polling helper
for the site select, which is unchanged.
Claude-Session: https://claude.ai/code/session_01NcMG6xXdsJbmD2wYAgkK29
An ephemeral inbound method (created and deleted through the CLI, with a never-taken
Route.To branch so prediction resolves a route target) is run live in a real browser through
the fixture's /auth/login; the reload test drives the OnAfterRenderAsync localStorage restore
and then proves the page is still interactive. Both fail on the pre-fix image at the
'Predicted' wait (the circuit deadlock) and pass on the fixed one.
Claude-Session: https://claude.ai/code/session_01NcMG6xXdsJbmD2wYAgkK29