A robust multi-agent orchestration framework utilizing historical governance models to provide fault tolerance and separation of concerns.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install tang-agents
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install tang-agents using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Tang Agents: Multi-Agent Collaboration System?
Tang Agents is a sophisticated multi-agent collaboration system inspired by the Tang Dynasty's administrative structure, specifically the Three Departments and Six Ministries system. This contribution to Openclaw Skills provides a framework for managing complex AI tasks through a clear separation of powers, ensuring that policy drafting, review, and execution are handled by distinct, specialized agents.
By implementing this historical model, the framework provides a structured approach to agentic workflows where reliability is paramount. The system is designed to prevent single points of failure through parallel execution and a rigorous review process, making it an essential tool for developers building resilient Openclaw Skills applications.
Tang Agents: Multi-Agent Collaboration System Use Cases
- Orchestrating complex software development tasks requiring multi-stage verification.
- Managing high-stakes content generation where compliance and security checks are mandatory.
- Automating resource-heavy deployments that require parallel processing and failure isolation.
- Implementing human-in-the-loop approval workflows for critical business logic.
How Tang Agents: Multi-Agent Collaboration System Works
- The Zhongshu agent drafts a policy or plan based on the initial user request.
- The Menxia agent performs a review and can exercise a veto, triggering up to three retries to ensure quality.
- Upon approval, the Shangshu agent dispatches the task to the relevant administrative branches.
- Six specialized Ministries (Libu, Hubu, Bingbu, etc.) execute specific sub-tasks in parallel, covering permissions, resources, and security.
- The system consolidates the output for the final user response or human approval.
Tang Agents: Multi-Agent Collaboration System Setup
To integrate this system into your project, install the dependencies using the following command:
pip install -r requirements.txt
Initialize the system within your Python script to start leveraging Tang Agents for your Openclaw Skills development.
Tang Agents: Multi-Agent Collaboration System Data Schema & Taxonomy
The system organizes data through an Edict-based lifecycle, tracking metadata across different ministry jurisdictions:
| Component | Responsibility | Data Tracked |
|---|---|---|
| Zhongshu | Policy | Task definitions and initial intent |
| Menxia | Review | Retry counts and veto logs |
| Six Ministries | Execution | Resource usage, security status, and deployment logs |
| Dashboard | Monitoring | Real-time status of active Edict IDs |
Tang Agents: Multi-Agent Collaboration System Advanced Features
- Intelligent Circuit Breaker that automatically pauses the system after three consecutive agent failures.
- Independent Timeout Handling for every agent (variable between 3-15 seconds) to prevent system hangs.
- Graceful Degradation support that returns pre-defined template data instead of failing entirely.
- High-concurrency parallel execution across the Six Ministries to maximize throughput within Openclaw Skills environments.
- Comprehensive Failure Isolation ensuring that a crash in one ministry does not interrupt the operations of others.
SKILL.md
Loading
Related Openclaw Skills
Yuque Knowledge Base Manager
A comprehensive integration for AI agents to search, read, create, and organize documents within the Yuque knowledge management platform.
AI Promo
A specialized utility for discovering and managing promotions, token discounts, and referral rewards across various AI model platforms.
China Commerce Copilot
An intelligent routing agent that directs China-based shopping and delivery requests to the most effective platform-specific specialized skills.
AI Companion Market Research Automator
A fully automated agent skill for generating comprehensive market research reports on the global AI companion software sector.
Xueqiu Ops
An automated interaction and content creation tool for the Xueqiu investment community powered by Openclaw Skills.
WeCom Workplan Summary
A professional utility for summarizing team work plans and generating actionable reports from WeCom Smart Sheets or pasted data.
Featured*
