AI Code Reviews Changelog

​

v4.205.0

​

Added

• New AI Code Review Features:
  • /matter review-max - Advanced AI Code Review with tool-use enabled analysis
  • /matter fix - Automated Code Fix Generation for review comments
• Enhanced documentation for new AI-powered code review capabilities
• Confidence scoring system for review comments
• Automated fix PR generation with separate branch creation

​

Changed

• Updated command list documentation with detailed feature descriptions
• Enhanced code review workflow with deeper analysis capabilities
• Improved developer experience with automated fix application

​

Benefits

• Time saved: 15-30 minutes per PR on fix application
• Review depth: 3x more comprehensive with tool-use
• Accuracy: Confidence scoring reduces false positives
• Developer satisfaction: Zero manual copy-pasting

​

Formatting & Config

• Enhanced Readability: Improved AI-generated PR review formatting for a better developer experience.
• Optimized Configuration: Fine-tuned AI model settings specifically for code review workflows.

​

Reliability & Reasoning

• Robust Fallbacks: Intelligent model fallback mechanisms for uninterrupted reviews.
• “Thinking” Models: Added support for reasoning tool calls before review generation.

​

Fix in Cursor Button

​

Faster Review Cycles

​

Hallicination Evals

​

Login via Microsoft

​

Enable PR Review Skip Alerts

​

PR Review Priority Settings

​

Gitlab V16 Support

​

Major AI Code Review Engine Upgrade

​

Dependabot PR Reviews Skip

​

Migration to Axon-Code Model

​

Codebase Learnings File Patterns Update

​

Codebase Learnings in GA

​

Mermaid Diagram Bugfix

​

Auto-Resolve Conversations

​

MATTER.md Rules

​

Example

# CODE_REVIEW_RULES

## 1. Code Format & Style
   - Use camelCase for variables and functions
   - Use PascalCase for constructors and classes
   - Use const for immutable values, let for variables
   - Prefer arrow functions for callbacks and short functions

## 2. Modern JavaScript Patterns
   - Use destructuring for object/array extraction
   - Prefer template literals over string concatenation
   - Use optional chaining (?.) for safe property access
   - Use nullish coalescing (??) for default values

## 3. Async Operations
   - Prefer async/await over .then() chains
   - Always handle promise rejections with try/catch
   - Use Promise.all() for parallel async operations
   - Avoid mixing async/await with .then()

## 4. Error Handling
   - Validate function parameters at entry
   - Use meaningful error messages
   - Handle edge cases (null, undefined, empty arrays)
   - Log errors with sufficient context

## 5. Performance & Memory
   - Avoid creating functions inside render loops
   - Use array methods (map, filter, reduce) appropriately
   - Clean up event listeners and timers
   - Minimize DOM manipulation in loops

​

Low Verbosity Summary

​

Bugfix - AI Summary generation

​

Review Engine v6

• Improved Accuracy: Enhanced understanding of code context and structure.
• Faster Processing: Optimized algorithms for quicker review generation.
• Better Language Support: Expanded support for more programming languages and frameworks.
• Enhanced Security Checks: More robust detection of security vulnerabilities and code smells.

​

Improvements

• Stronger code reviews with our updated review engine
• Updates to on-prem enterprise with better scaling and lite-reviews for cost savings
• Bug fixes in summary generations and mermaid diagrams
• Tanka poem in Summary: Similar to Haiku but expressing deeper emotions or nature themes.

​

Slack Alerts 🔔

​

Auto Resolve Comments ✅

​

Similarity Search 🔍

​

Integrations

​

Gitlab Setup is GA

​

Atlassian Bitbucket Setup is GA

​

Developer Workflow Integration

• Keyboard Shortcut Customization: Personalize your MatterAI interaction experience.

​

Code Analysis Upgrades

• Cross-Repository Analysis: Identify patterns and issues across multiple repositories.
• Legacy Code Support: Improved analysis for older codebases and deprecated languages.

​

Review Engine Improvements

• End Line Character Fixes: Intelligent handling of line ending inconsistencies in GitHub.
• Committable Code Generation: Enhanced code suggestions that are ready to commit without manual fixes.
• Reduced Verbosity: More concise code and text outputs without sacrificing clarity.

​

LLM Cost Metrics

• Token Usage: Tools to monitor and optimize token consumption.
• Model Performance: Side-by-side comparison of different LLM providers.

​

Enterprise AI Integration

• Vertex AI Support: Native integration with Google’s Vertex AI platform for enhanced model capabilities.
• Multi-Model Orchestration: Intelligent routing between different AI providers based on task requirements.

​

Enterprise Security Enhancements

• Role-Based Access Control: Granular permission management for teams of all sizes.

​

New Features & Improvements

• ✅ Similarity Search (GA): Now generally available with AST grep functionality (no AI required).
• 🔍 Enhanced Code Reviews: Deeper analysis, better accuracy, and improved string literals handling.
• 📝 Code Suggestion Formatting: Fixed edge-case issues.
• 🤖 Vertex API Support: Added Service Account authentication.
• ✨ Gemini Models: Updated to latest stable versions.

​

Fixes & Tweaks

• 🔧 GitHub New-Line Character Parsing: Fixed edge cases.
• 📊 Updated PR Review & Summary Formats: Improved presentation and clarity.
• 📈 Added LLM Cost/Usage Tracking Metrics: Better visibility into AI consumption.
• 🐛 Command Pattern Recognition: Bug fixes for improved command detection.

​

AI Summaries v2

• Advanced Detail Summary: Comprehensive PR summaries with deeper context understanding
• Structured Information: Automatically splits summaries into logical sections for better readability
• Security Visualization: Tables for package vulnerabilities, CVEs, and detailed explanations

​

AI Code Review Engine V2

• Precise Code Suggestions: Deterministic commit positioning ensures feedback exactly where you need it
• Duplicate Prevention: Smarter review intelligence avoids re-analyzing unchanged code sections
• Cost Optimization: Auto-model swapping selects the best AI for each use-case, balancing performance and efficiency

​

Command Improvements

• /matter review: Reviews only new commits for faster and more efficient reviews
• /matter review-full: Triggers a comprehensive re-review of the entire codebase when needed

​

Agentic Chat

• Code-Aware Conversations: Parses PR diffs, file context, and repo structure for relevant answers
• Technical Explanations: Explains implementation decisions with precise technical details
• Impact Analysis: Identifies potential side effects and integration points of code changes

​

Knowledge Retention

• Codebase Familiarity: Develops understanding of project-specific patterns and conventions

​

Continuous Improvement

• Adaptive Reviews: Refines review approach based on team interactions and preferences
• Team Alignment: Becomes increasingly aligned with team-specific needs and standards

​

AI Memories

• Learning & Adaptation: Observes developer interactions to capture team preferences and coding styles
• Pattern Recognition: Identifies recurring patterns in code reviews and discussions for better context
• Contextual Understanding: Builds deeper understanding of your codebase and project architecture