Reproducible product bugs
Suitable when the scope and expected result can be made explicit before implementation.
Bug Reproduction + Fix
Turn “it breaks sometimes” into a reproducible fix.
We start by proving the failure or establishing the closest reliable reproduction, freeze the expected behavior, fix the root cause inside the approved boundary, add regression coverage, and review the change separately.
Good fit
Bounded work with a visible finish line.
Error-handling defects
Suitable when the scope and expected result can be made explicit before implementation.
Incorrect state transitions
Suitable when the scope and expected result can be made explicit before implementation.
API / UI regressions
Suitable when the scope and expected result can be made explicit before implementation.
Read-only preflight
What we verify first.
Preflight gathers evidence. It does not authorize implementation.
Observed vs expected behavior
Captured as evidence for qualification and the Sprint Brief before implementation begins.
Logs or reproduction steps that are safe to share
Captured as evidence for qualification and the Sprint Brief before implementation begins.
Affected issue / module
Captured as evidence for qualification and the Sprint Brief before implementation begins.
Existing test and build path
Captured as evidence for qualification and the Sprint Brief before implementation begins.
Lane use
Implementation stays inside the approved brief.
The lane works from the approved reproduction and acceptance criteria. If the bug reveals a materially larger redesign, it stops for guidance instead of expanding the job.
Definition of done
Evidence before handoff.
The original failure no longer reproduces under the agreed conditions, regression evidence exists, and the broader suite is checked.
Client controls merge. Deploy remains off by default.