Skip to main content

v2.1.0

New Plan Development Mode​

Plan Mode​

Serious programming, following the "Plan Creation → Step-by-Step Implementation" workflow. Suitable for progressive development and task management, ensuring orderly project advancement.

Architecture and Performance Optimization​

Dynamic Model Loading​

Convert claude-code models to dynamic loading mechanism using the getClaudeCodeModels() function for on-demand loading, optimizing memory usage and startup performance.

Terminal Simplification​

Simplify terminal integration for editors like VSCode and JetBrains, improving developer experience.

Directory Scanning Optimization​

Optimize directory scanning performance by caching gitignore check results, reducing redundant computations.

Tool System Enhancements​

Model-Specific Tool Customization​

Add excludedTools and includedTools configurations to support customizing available tool sets for different models, enhancing model adaptability.

Opt-in Tool Support​

Add customTools functionality, allowing users to enable specific tools on-demand, providing more precise tool control.

Text Replacement Tools​

Add search_and_replace tool for batch text replacement operations, and search_replace native tool for single precise replacement.

Important Change: Remove line_count parameter from write_to_file tool, simplifying the tool interface.

Native Tool Ecosystem Expansion​

Multi-Provider Support​

Enable native tool support for DeepSeek, Doubao, Requesty, and multiple other AI providers, expanding ecosystem coverage.

Mainstream Platform Integration​

Add native tool support for OpenAI-compatible providers, Vertex Gemini, Grok, Bedrock, and other platforms, enhancing interoperability.

Stability and Error Handling​

Race Condition Fix​

Fix race condition issues in the new_task tool for native protocol, improving stability in concurrent scenarios.

Exception Handling Enhancement​

Handle malformed native tool calls to prevent system hanging, enhancing robustness.

Model Fetching Optimization​

Fix Vercel AI Gateway model fetching issues to ensure model service availability.

Improvement: Enhance error message readability and optimize documentation links for faster problem resolution.

User Experience Improvements​

Display Mode Management​

Add ModeSwitch component, allowing users to flexibly switch between different display modes in chat.

Auto-Approval Mechanism Optimization​

Improve auto-approval timer visibility in follow-up suggestions, providing more intuitive time feedback. Optimize interaction logic so that auto-approval timeout is automatically cancelled when users start typing, preventing accidental operations.

Internationalization Improvements​

Update i18n features, improving skipHint text clarity in multiple languages, enhancing multilingual user experience.

Testing and Development Tools​

Terminal Test Updates​

Update terminal test mocks and shell path handling logic, ensuring accuracy and consistency of test environments.