Work the crash tracker by user impact, one issue per ticket
The crash affecting the most users, taken one issue at a time rather than as an inbox.
The ticket
Fix the crash affecting the most users, with a test that reproduces it.
Acceptance criteria
- The tracker issue is linked and its user count recorded
- A test reproduces the crash
- The fix handles the cause, not only the reported stack frame
- The issue is resolved in the tracker after the fix ships
What lands as proof
The tracker issue, the reproducing test, and the affected-user count that justified picking this one.
Why teams defer it
- The tracker is a firehose, so it gets checked after incidents rather than as routine work.
- Turning a stack trace into a reproduction is most of the effort and none of the visible progress.
Questions
- What does the agent actually change?
- The ticket is scoped to one outcome: fix the crash affecting the most users, with a test that reproduces it. 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 stack trace linked, reproduction test, fix, 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.