fix-findings

To cluster review findings and apply fixes interactively, the skill groups verified findings into clusters and guides you through per-cluster decision prompts, validating changes before applying them. It operates within non-destructive safe paths with interactive gating.

Yes, you can run automated review fixes in the background. The skill supports a background fixer workflow that executes safe fixes asynchronously, reducing turnaround time when interactive gating is not required for specific verified findings.

Clustering in code review resolution groups verified findings logically so they can be addressed together. The skill uses clustering logic to batch related issues, enabling per-cluster decision prompts and guided interactive fixes for more efficient resolution.

Yes, the interactive fix workflow requires confirmation before applying changes. The skill uses interactive gating and per-cluster decision prompts, ensuring it only acts on verified findings after explicit approval and within non-destructive safe paths.

The primary limitation is that automation is gated by interactive confirmation and restricted to non-destructive safe paths. Background execution is only applied to safe fixes asynchronously, meaning destructive or unverified findings cannot be automatically resolved without guided approval.

To start fixing verified findings from a self-review, the skill gathers findings from context, applies clustering logic to group them, and steps you through interactive fixes. You validate changes through per-cluster decision prompts before they are applied.