Skip to main content
Glama

Codex Bridge

by eLyiN
CHANGELOG.md8.27 kB
# Changelog All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). ## [1.2.1] - 2025-09-04 ### Fixed - **Error Handling Consistency**: Fixed invalid format parameter error handling to always return JSON - Previously returned plain text for invalid formats, now returns structured JSON error - Applied fix to both `consult_codex` and `consult_codex_with_stdin` functions - Ensures consistent API responses across all error conditions ### Added - **Comprehensive Testing Suite**: New `tests/` directory with organized test structure - `tests/test_mcp_tools.py` - Complete test suite for all MCP tools functionality - `tests/README.md` - Testing documentation and guidelines - `tests/TEST_RESULTS.md` - Detailed test execution results and performance metrics - All tests passing with 100% functionality verification ### Changed - **Code Quality**: Removed unused imports and resolved linting warnings - **Project Structure**: Organized testing files in dedicated `tests/` directory - **Documentation**: Enhanced testing documentation and contribution guidelines ### Tested - ✅ All three MCP tools (`consult_codex`, `consult_codex_with_stdin`, `consult_codex_batch`) - ✅ All output formats (text, JSON, code) with proper validation - ✅ CODEX_SKIP_GIT_CHECK environment variable functionality - ✅ Error handling for invalid parameters and edge cases - ✅ JSON response consistency and structure - ✅ Code block extraction and processing ## [1.2.0] - 2025-01-03 ### Added - **Git Repository Check Configuration**: New `CODEX_SKIP_GIT_CHECK` environment variable - Allows running Codex Bridge in directories that are not Git repositories - Defaults to false for security; opt-in with "true", "1", or "yes" - Fixes "Not inside a trusted directory" errors when working outside Git repos - **Enhanced Documentation**: Updated README.md and CLAUDE.md with new configuration options - **Security Guidelines**: Added warnings about using git check bypass only in trusted directories ### Changed - **Command Construction**: All three MCP tools now conditionally add `--skip-git-repo-check` flag - **Error Handling**: Improved error messages and troubleshooting guidance ## [1.1.0] - 2025-01-28 ### Changed - **Default Format**: Changed default output format from "text" to "json" for structured responses - **Default Timeout**: Increased default timeout from 60 to 90 seconds for better reliability - **Documentation**: Updated function docstrings with timeout recommendations (60-120 seconds) - **User Experience**: Improved out-of-the-box configuration to prevent hanging issues ### Added - Enhanced timeout guidance in function documentation - Better default configuration for production use ### Fixed - Reduced likelihood of timeout-related hanging during complex queries - Improved structured output consistency with JSON default ## [1.0.1] - 2025-01-28 ### Fixed - **Process Management**: Resolved infinite codex process spawning issue - **CLI Integration**: Improved subprocess handling for better stability ### Changed - **CI/CD**: Updated CI status badge URL to reflect new repository location - **Build Process**: Removed deprecated activation scripts and configuration files ## [1.0.0] - 2025-01-28 ### Added #### Core Features - **MCP Server Implementation**: Complete Model Context Protocol server for bridging Claude Code to Codex AI - **Two MCP Tools**: - `consult_codex`: Direct CLI bridge for simple queries with optional model selection - `consult_codex_with_files`: CLI bridge with file attachment support for detailed analysis - **CLI Integration**: Direct integration with OpenAI's official Codex CLI - **Model Support**: Support for Codex models #### Architecture & Design - **Stateless Operation**: No session management, caching, or complex state - **Fixed Timeout**: 60-second maximum execution time for all operations - **Simple Error Handling**: Clear error messages with fail-fast approach - **Minimal Dependencies**: Only requires `mcp>=1.0.0` and external Codex CLI #### Deployment & Installation - **uvx Support**: Production deployment via uvx with built wheel - **Development Mode**: Traditional pip installation for development - **Startup Scripts**: - `start_server_uvx.sh` for production - `start_server_dev.sh` for development - **Package Configuration**: Complete pyproject.toml with metadata and entry points #### Documentation - **Comprehensive README**: Installation, usage examples, troubleshooting - **Contributing Guidelines**: Development setup, code style, PR process - **Security Policy**: Vulnerability reporting, best practices, update process - **License**: MIT License for open source use #### Developer Experience - **Claude Code Integration**: Seamless MCP protocol integration - **Error Diagnostics**: Clear error messages for common issues - **Development Tools**: Easy local testing and development setup ### Technical Details #### Dependencies - **Required**: `mcp>=1.0.0` - **External**: Codex CLI (npm package `@openai/codex-cli`) - **Python**: Compatible with Python 3.9+ #### File Structure ``` codex-bridge/ ├── src/ │ ├── __init__.py # Package entry point and version │ ├── __main__.py # Module execution entry point │ └── mcp_server.py # Core MCP server implementation ├── start_server_uvx.sh # Production startup script ├── start_server_dev.sh # Development startup script ├── pyproject.toml # Python package configuration └── [documentation files] # README, CONTRIBUTING, SECURITY, etc. ``` #### Configuration - **Default Model**: Codex model for optimal performance - **Timeout**: 60 seconds for all CLI operations - **Working Directory**: Configurable per request - **File Encoding**: UTF-8 with error handling ### Security Considerations - **Input Validation**: Basic validation for file paths and queries - **Process Isolation**: Subprocess execution for CLI calls - **No Network Exposure**: All network requests handled by Codex CLI - **Minimal Attack Surface**: Simple, stateless architecture ### Performance Characteristics - **Startup Time**: Near-instant MCP server startup - **Memory Usage**: Minimal memory footprint - **Execution Time**: Limited by 60-second timeout - **Scalability**: Stateless design allows multiple concurrent requests --- ## Release Notes ### Version 1.0.0 Highlights This initial release establishes Codex Bridge as a production-ready MCP server with a focus on: 1. **Simplicity**: Straightforward architecture that's easy to understand and maintain 2. **Reliability**: Robust error handling and fixed timeout protection 3. **Integration**: Seamless Claude Code integration via MCP protocol 4. **Performance**: Direct CLI integration for optimal speed and cost efficiency 5. **Community**: Complete open source documentation and contribution guidelines ### Migration Notes This is the initial public release, so no migration is required. Future versions will include migration guidance here. ### Breaking Changes None in this initial release. Future breaking changes will be clearly documented here. ### Deprecation Warnings None in this initial release. Future deprecations will be announced here with migration timelines. --- ## Contributors Special thanks to all contributors who made this release possible: - **Core Development**: Project architecture and implementation - **Documentation**: Comprehensive documentation and examples - **Testing**: Manual testing across different environments - **Community**: Feedback and suggestions for improvement ## Links - **Repository**: [https://github.com/shelakh/codex-bridge](https://github.com/shelakh/codex-bridge) - **Issues**: [https://github.com/shelakh/codex-bridge/issues](https://github.com/shelakh/codex-bridge/issues) - **MCP Protocol**: [https://modelcontextprotocol.io/](https://modelcontextprotocol.io/) - **Codex CLI**: [https://www.npmjs.com/package/@openai/codex-cli](https://www.npmjs.com/package/@openai/codex-cli)

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/eLyiN/codex-bridge'

If you have feedback or need assistance with the MCP directory API, please join our Discord server