code_review
Analyze code or diffs using two parallel LLM reviewers to identify context-blind errors and improve code quality.
Instructions
Review a code snippet or diff using two LLM reviewers in parallel.
At least one of code or paths must be provided.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| code | No | The code snippet or diff to review. Pass a meaningful amount of code for a useful review. Required unless `paths` is provided (both can also be provided together). | |
| paths | No | File paths to source files to review. Pass either: a list of paths (e.g. ['src/main.py', 'src/utils.py']), a newline-separated string of paths, or a JSON array string. Paths are relative to the current working directory; absolute paths also work. Files are read automatically from disk. Required unless `code` is provided (both can also be provided together). | |
| context | No | Additional guidance for the reviewer (e.g., review goals, known issues, areas to focus on, style conventions, constraints). Max 10,000 characters. |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| result | Yes |