CCG is a workflow engine that orchestrates collaboration between multiple large language models through a command-line interface.
The tool solves the problem of coordinating specialized AI models for complex tasks. When a user provides a goal via the `/ccg:go` command, the engine automatically analyzes the intent, selects an appropriate strategy, and orchestrates execution across Claude, Codex, Gemini, and other models. Claude remains the primary orchestrator while delegating specialized work to other models through a Go binary bridge. The system includes a DeepSeek Harness variant that runs natively inside the harness environment, offering role-pinned delegation tools where each role operates on its own model with its own expert persona, along with features like model panels for side-by-side independent responses and live teammates that persist across conversation turns.
Developers should adopt this tool if they need to coordinate multiple AI models for complex workflows where different models excel at different tasks. It suits projects where intent analysis and dynamic strategy selection are valuable, and where keeping one model in control while leveraging specialized capabilities of others makes sense. The tool is available as both a standalone CLI and integrated into DeepSeek Harness, giving teams flexibility in how they deploy it.
The maintainers respond to new issues and pull requests within hours. Work in the issue tracker centers on bug fixes and enhancements.
