CodeScene

Analyzes code as it's typed in the editor and calculates a CodeHealth™ metric at the file level, displaying both current and previous scores with delta values to show degradation or improvement. The metric is computed using proprietary fact-based analysis rules and rendered inline in a real-time monitoring widget that updates continuously during the development session without requiring manual triggers.

Identifies code smells (structural anti-patterns and maintainability issues) within the current file and renders them as inline diagnostic items in the VS Code editor, with actionable improvement guidance provided for each detected smell. Detection runs automatically as code is typed, leveraging CodeScene's proprietary analysis rules to flag issues like high cyclomatic complexity, code duplication, and other maintainability concerns.

Leverages CodeScene's remote AI service (CodeScene ACE) to automatically refactor detected code smells and technical debt directly within the VS Code editor. The system identifies refactoring opportunities based on code health analysis, sends code context to CodeScene's hosted AI backend, and applies transformations back to the editor. Requires explicit organizational consent and activation before AI services become accessible.

CodeScene ACE integrates with multiple LLM providers (OpenAI GPT, Google Gemini, Anthropic Claude) to power code analysis and refactoring capabilities. The extension abstracts away model selection and routing, allowing organizations to choose their preferred LLM provider while maintaining consistent code analysis and refactoring workflows. Model inference is executed on CodeScene's remote backend, not locally in the extension.

Maintains a real-time monitoring widget in VS Code that tracks code health metrics at the file level, displaying current CodeHealth score, previous score, and delta (change) value. The widget updates continuously as code is edited, providing visual feedback on whether recent changes improved or degraded code quality. Historical tracking enables developers to see the trajectory of code health changes within a single editing session.

Implements an organizational-level consent and activation model where CodeScene ACE (AI-powered refactoring) must be explicitly enabled by organization administrators before any developers can access AI services. This governance layer ensures that organizations maintain control over AI service usage, data transmission, and compliance with internal policies. Consent is enforced at the extension level, preventing unauthorized use of AI capabilities.

Analyzes source code across multiple programming languages using language-agnostic code health metrics and code smell detection rules. The extension automatically detects the language of the current file and applies appropriate analysis rules without requiring language-specific configuration. Supports 'most popular languages' but specific language coverage is not documented.

Automatically analyzes code as it's typed in the editor without requiring manual trigger, analysis commands, or explicit save events. The extension runs continuous background analysis on the current file, updating diagnostics and metrics in real-time as developers edit code. This passive analysis approach integrates code quality feedback directly into the natural development workflow without interruption.