Skip to content

v1.0.3 - Claude Code as Orchestrator with Multi-LLM Support

Latest
Latest

Choose a tag to compare

View all tags
@mivertowski mivertowski released this 22 Aug 19:53
· 3 commits to main since this release
v1.0.3
b37021c

🎉 Major Release: v1.0.3

✨ Highlights

  • Claude Code as Orchestrator: Use Claude Code itself as the orchestrator LLM with automatic permission bypass
  • Multi-LLM Support: Now supports 7+ LLM providers for maximum flexibility
  • Interactive GitHub Integration: Enhanced CLI with priority-based issue selection
  • Local LLM Support: Run models locally with CUDA/CPU acceleration

🚀 New Features

Claude Code Integration

  • 🤖 Use Claude Code as the orchestrator LLM with --llm claude-code
  • 🔓 Automatic permission bypass for fully automated execution
  • 📦 Works with Claude Code subscriptions - no API key required
  • ⚡ Optimized for Claude Opus 4.1 (default model)

Multi-LLM Provider Support

Now supporting 7 different LLM backends:

  • Claude Code: Official SDK with automation support
  • OpenRouter: Access to 100+ models
  • Ollama: Easy local model management
  • Local CUDA: GPU-accelerated inference
  • Local CPU: CPU-only inference
  • llama.cpp: Optimized C++ inference
  • VLLM: High-performance serving

Enhanced GitHub Integration

  • 📋 Interactive Issue Selection: Browse and select issues with priority indicators (p0-critical, p1-high, p2-medium, p3-low)
  • 🔄 Automated Processing Mode: Continuously process issues by priority with configurable polling
  • 💬 Progress Comments: Real-time updates posted to GitHub issues
  • 🔧 Smart PR Creation: Automatically adds "Fixes #X" for issue auto-closing

Performance & Compatibility

  • 📊 Token Optimization: 50-80% reduction in API costs
  • 🔒 Corporate Proxy Support: Works behind Zscaler and other proxies
  • 🎯 Claude Opus 4.1: Latest and most capable model as default
  • 📈 Better Error Recovery: Exponential backoff and retry strategies

📝 API Changes

New LLM Provider Interface

// Unified interface for all providers
import { LLMProviderFactory } from 'cco-cli';

const client = await LLMProviderFactory.createProvider(
  'claude-code', // or 'openrouter', 'ollama', etc.
  { permissionMode: 'bypassPermissions' },
  logger
);

GitHub Orchestrator Enhancements

// Interactive issue selection
const issues = await orchestrator.selectIssuesInteractively();

// Automated processing
await orchestrator.processIssuesAutomatically({
  pollInterval: 300,
  labels: ['bug', 'enhancement'],
  createPR: true
});

🛠️ Usage Examples

Using Claude Code as Orchestrator

# Process issue with Claude Code
cco github --issue 123 --llm claude-code

# Interactive mode with Claude Code
cco github --interactive --llm claude-code

Local LLM Support

# Using Ollama
cco github --issue 123 --llm ollama --local-model codellama:13b

# Using CUDA acceleration
cco github --issue 123 --llm local-cuda --local-model meta-llama/Llama-2-13b-chat-hf

Automated Issue Processing

# Process all issues automatically
cco github --auto --poll-interval 30 --create-pr

# Process specific labels
cco github --auto --labels "bug,enhancement" --max-issues 5

🐛 Bug Fixes

  • Fixed Git remote URL parsing for HTTPS repositories with .git suffix
  • Fixed environment variable loading in npm scripts
  • Fixed Claude Code permission prompts in automated mode
  • Fixed SSL certificate validation in corporate environments

📦 Installation

npm install -g cco-cli@1.0.3

📚 Documentation

  • Comprehensive API documentation added
  • Updated README with all features
  • New SDK integration guide
  • Enhanced troubleshooting guide

🙏 Acknowledgments

Thanks to all contributors and users who provided feedback for this release!

Full Changelog: v1.0.2...v1.0.3
NPM Package: https://www.npmjs.com/package/cco-cli

Assets 2
Loading