- Add WebSearchTool with DuckDuckGo integration for web search
- Provides titles, snippets, and links for current information
- Add URLFetcherTool for AI-powered web content analysis
- Fetches and converts HTML content to markdown
- Processes content using AI with user-provided prompts
- Includes 15-minute caching for efficiency
- Uses queryQuick for fast content analysis
- Register both tools in the tools registry
- Update documentation to reflect new web capabilities
- Added a new function to split PATH entries correctly for both Windows and POSIX environments.
- Improved shell detection logic to utilize the new PATH splitting function, ensuring accurate path handling across different platforms.
- Implemented shell detection for POSIX, MSYS, and WSL environments.
- Added functions for quoting strings and converting paths for Bash compatibility.
- Updated command execution to handle different shell types and ensure proper path handling.
- Improved initialization by sourcing ~/.bashrc when available and managing file paths for stdout, stderr, and cwd.
- Change debugLogger to use ~/.kode as the base directory for all logs
- API error logs now save to ~/.kode/logs/error/api/
- Debug logs save to ~/.kode/<project-dir>/debug/
- Remove dependency on env-paths cache directory for logs
- Add elegant API error logging system with file storage in ~/.kode/logs/error/api/
- Show API responses in terminal during errors (verbose mode only)
- Fix tool description usage (prompt vs description methods)
- Remove unnecessary bundledDependencies check in publish workflow
- Bump version to 1.0.80
- Changed Anthropic API to use tool.prompt() instead of description()
- Ensures full tool instructions are sent to the model
- Both Anthropic and OpenAI paths now consistently use prompt() method
- Resolves issue where short descriptions were being sent instead of full documentation
- Update cli.js wrapper to use node_modules/.bin/tsx
- Fix ESC key cancellation error display in openai.ts stream processing
- Simplify REPL onCancel function
- Add security notice and model performance recommendations to README
- Resolved conflicts in README.md by combining @ mention system docs with Docker usage instructions
- Auto-merged PromptInput.tsx improvements from shareAI-lab
- Added Docker support with .dockerignore and Dockerfile
- Updated Chinese README with latest changes
- Maintains full backward compatibility with existing @ mention and AGENTS.md features
- Fix word detection to properly handle ./ and ../ patterns
- Improve path separator handling in getWordAtCursor
- Add special handling for dot patterns in paths
- Fix shouldAutoTrigger to recognize all path patterns
- Ensure proper inclusion of dots and slashes in path context
This fixes the issue where './sr' wouldn't show 'src/' in completions
- Increase display limit from 10 to 25 entries to show more files
- Fix directory navigation when path ends with slash
- Add proper sorting (directories first, then alphabetical)
- Hide dotfiles by default unless explicitly requested
- Fix path resolution for trailing slashes in directory navigation
- Implement model adapter factory for unified API handling
- Add response state manager for conversation continuity
- Support GPT-5 Responses API with continuation tokens
- Add model capabilities type system
- Include deployment guide and test infrastructure
- Enhance error handling and debugging for model interactions
- Add advanced fuzzy matching with 7+ strategies (exact, prefix, substring, acronym, initials, fuzzy, Levenshtein)
- Create comprehensive database of 500+ common Unix commands for smart autocompletion
- Implement intelligent Tab completion with @ prefix injection for agents and files
- Add sophisticated input pattern recognition for commands like "dao", "gp5", "py3"
- Enhance mention system with TaskProgressMessage component for better user feedback
- Update documentation with comprehensive intelligent completion guide
- Clean up 21 temporary markdown files to maintain repository cleanliness
- Improve project structure and configuration documentation
- Optimize completion system performance with advanced caching and scoring
- Add GPT-5 model definitions (gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-chat-latest)
- Implement GPT-5 Responses API support with intelligent fallback to Chat Completions
- Add GPT-5 specific parameter handling (max_completion_tokens, temperature=1)
- Support custom tools and freeform function calling capabilities
- Add reasoning effort and verbosity control parameters
- Implement GPT-5 connection testing service
- Add model capability detection and automatic parameter transformation
- Support both official OpenAI and third-party GPT-5 providers
- Add todo list and sticker request UI components
- Improve notebook support with better type definitions
- Enhance debug logging and error handling for GPT-5
- Update model selector with GPT-5 compatibility checks
This commit provides full GPT-5 support while maintaining backward compatibility with existing models.
- Update project branding from claude-cli to Kode
- Reorganize documentation with new development guides
- Add CONTRIBUTING.md and Chinese README
- Remove worktree_merge command and relocate system-design.md
- Update dependencies and package configuration
- Improve custom commands service with better error handling
- Clean up storage utilities and debug logging
