danghuangshang: a hierarchical multi-agent collaboration system that organizes AI agents into a structured court or corporate hierarchy
danghuangshang: a hierarchical multi-agent collaboration system that organizes AI agents into a structured court or corporate hierarchy
What it solves
This project provides a ready-to-use multi-agent collaboration system that transforms the experience of managing multiple AI agents into a structured organizational hierarchy. Instead of interacting with a single general-purpose AI, users can manage a specialized "AI Court" where different agents (acting as government ministers or corporate executives) handle specific domains like software engineering, finance, marketing, and legal compliance, reducing the friction of manual prompt engineering and task delegation.
How it works
Built on the OpenClaw framework, the system implements a hierarchical command structure based on historical or modern organizational models (such as the Ming Dynasty cabinet system).
- Command Flow: A user (the "Emperor") issues a command. A dispatcher agent (e.g., the Director of Ceremonies) receives it and sends it to a planning agent (the Cabinet) to optimize the prompt and create an execution plan. The dispatcher then assigns tasks to specialized agents (the Six Ministries) for execution.
- Execution: Specialized agents use a library of over 60 built-in skills (GitHub, Notion, web browsing, etc.) to complete tasks.
- Quality Control: A supervisory agent (the Censorate) automatically reviews code pushes to GitHub to ensure quality.
- ** keluarga Memory**: Each agent maintains its own independent workspace and memory, allowing them to become more specialized over time.
- Interfaces: The system integrates with Discord, Feishu, Slack, and Telegram, allowing users to command their AI team via chat.
Who it’s for
- Individuals and developers who want a structured way to manage complex projects using multiple AI agents without writing extensive Python code.
- Project managers looking for a template for AI-driven organizational management.
- Users who prefer a chat-based interface (like Discord) to interact with a team of specialized AI agents.
Highlights
- Pre-configured Regimes: Choose between Ming Dynasty, Tang Dynasty, or Modern Corporate structures.
- Prompt Optimization: Automatic prompt enhancement via a dedicated "Cabinet" agent before task distribution.
- Extensive Tooling: 60+ integrated skills including GitHub, Notion, and browser automation.
- Multi-platform Support: Native integration with Discord, Feishu, Slack, and Telegram.
- Zero-Code Deployment: One-click installation scripts for rapid setup on cloud servers.
- Integrated GUI: A Web Dashboard for monitoring agent status, token usage, and system health.
Sources
- undefinedwanikua/danghuangshang