Problem
The current Code Review Copilot provides useful static analysis using regex-based checks (such as syntax validation, security patterns, performance warnings, and best-practice suggestions).
While this is helpful, it cannot explain why an issue matters or provide contextual feedback like an AI-powered code review assistant.
Proposed Enhancement
Enhance the existing Code Review Copilot by integrating AI-generated code review while preserving the current regex-based analysis.
The goal is not to replace the existing implementation, but to complement it.
Expected Experience
When a user submits code:
-
The existing static analysis should continue to detect syntax, security, and best-practice issues.
-
The AI should provide:
-
A short overall review of the code.
-
Readability and maintainability feedback.
-
Suggestions for improvements.
-
Potential bugs or edge cases.
-
Performance recommendations (where applicable).
-
Security observations (where applicable).
-
A revised/refactored version of the code (when appropriate).
-
A brief summary explaining the suggested changes.
This would make the Code Review Copilot feel more like a real AI assistant instead of only a rule-based analyzer.
Questions for the Maintainers
Before starting implementation, it would be helpful to clarify a few things:
-
Should the AI review use Sarvam AI, since the project already integrates it for the "Draft with AI" feature, or would another provider be preferred?
-
Should AI review always run, or should it only run after the existing regex-based checks pass?
-
Should the AI-generated review be displayed alongside the current analysis, or should it replace any existing sections in the UI?
-
Are there any expected limits on input size, supported programming languages, or API usage that contributors should keep in mind?
-
Is there any preferred response format (plain text vs. structured JSON) for the AI output?
If this enhancement aligns with the project roadmap, I'd be happy to work on it.
Problem
The current Code Review Copilot provides useful static analysis using regex-based checks (such as syntax validation, security patterns, performance warnings, and best-practice suggestions).
While this is helpful, it cannot explain why an issue matters or provide contextual feedback like an AI-powered code review assistant.
Proposed Enhancement
Enhance the existing Code Review Copilot by integrating AI-generated code review while preserving the current regex-based analysis.
The goal is not to replace the existing implementation, but to complement it.
Expected Experience
When a user submits code:
The existing static analysis should continue to detect syntax, security, and best-practice issues.
The AI should provide:
A short overall review of the code.
Readability and maintainability feedback.
Suggestions for improvements.
Potential bugs or edge cases.
Performance recommendations (where applicable).
Security observations (where applicable).
A revised/refactored version of the code (when appropriate).
A brief summary explaining the suggested changes.
This would make the Code Review Copilot feel more like a real AI assistant instead of only a rule-based analyzer.
Questions for the Maintainers
Before starting implementation, it would be helpful to clarify a few things:
Should the AI review use Sarvam AI, since the project already integrates it for the "Draft with AI" feature, or would another provider be preferred?
Should AI review always run, or should it only run after the existing regex-based checks pass?
Should the AI-generated review be displayed alongside the current analysis, or should it replace any existing sections in the UI?
Are there any expected limits on input size, supported programming languages, or API usage that contributors should keep in mind?
Is there any preferred response format (plain text vs. structured JSON) for the AI output?
If this enhancement aligns with the project roadmap, I'd be happy to work on it.