Sweep one named defect class through one module
One kind of defect, named precisely, cleared across one module and blocked from returning.
The ticket
Find and fix every instance of one named defect class within one module.
Acceptance criteria
- Every instance in the module is fixed
- Each fix has a test
- A lint rule blocks the class from returning
- The search that found them is recorded so it can be re-run
What lands as proof
The recorded search returning nothing, plus a rule that fails on a seeded instance.
Why teams defer it
- Unscoped, this is an open-ended hunt with no definition of done, so it never gets started.
- Writing the rule that stops the class returning is a separate skill from fixing the instances, and the person who has one often lacks the other.
Questions
- What does the agent actually change?
- The ticket is scoped to one outcome: find and fix every instance of one named defect class within one module. Work that serves that outcome is in scope, and anything outside it is left for a separate ticket, so the pull request stays reviewable.
- How do I know the work is done?
- The pull request carries the evidence, not only the diff. Here that means a test per site plus a rule that stops the class returning, so a reviewer can confirm the result without reproducing the work locally.
- How much oversight does this need?
- None is required before the run starts. Every claim this ticket makes can be checked by machine, so the agent works straight through and a person reviews the finished pull request like any other.