Crew diff review
crew_review_diffReviews working-tree changes for bugs, security, concurrency, and data-integrity issues, separating evidence-checked findings from unverified ones before you declare work complete.
Instructions
Have a scout review the current working-tree changes (staged, unstaged and new files) for bugs, security, concurrency and data-integrity problems. Findings are split into evidence-checked (the quoted code really exists) and unverified. Run it after making changes and before saying you're done.
Tool call template
This tool declares no required parameters. Add optional parameters from the input schema as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {},
"name": "crew_review_diff"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| base | No | Git ref to diff against (default HEAD). Use e.g. 'main' for a whole branch. | |
| focus | No | What to look at hardest, e.g. 'transaction boundaries' | |
| force | No | Review even if the diff is too large for the scout (it is truncated, and may time out). Default false. | |
| paths | No | Limit the review to these paths | |
| workspace | No | Absolute path of the repository. Defaults to the directory Claude Code launched the server in. | |
| acknowledged | No | Findings from an earlier run you already fixed, rebutted or accepted, so they aren't raised again. Start each with the finding's headline file:line, then a few words, e.g. 'src/jobs/sync.ts:110 fan-out not transactional: deliberate'. Matching findings within a few lines are collapsed to one line. | |
| includeUntracked | No | Include new untracked files (default true). AI-tool folders like .codex/ and .claude/ are always skipped. |