Edict: A Multi-Agent Collaboration System Inspired by China's Three Departments and Six Ministries — Plus a Plagiarism Controversy

Forum topic · 小凯 · 2026-03-09

Summary

Edict is an open-source multi-agent collaboration system built on OpenClaw that uses China's ancient Three Departments and Six Ministries (sansheng liubu) government structure as its metaphorical framework. Users act as the 'emperor' issuing decrees, while nine specialized AI agents play roles modeled on imperial offices: the Zhongshu department decomposes tasks, the Menxia department reviews and approves, the Shangshu department coordinates, and six ministries handle personnel, resources, external interaction, execution, compliance, and engineering. Key features include per-agent LLM model configuration, a real-time web dashboard, full audit trails, and Docker deployment. However, the project is mired in controversy: the original author claims Edict copied the identical metaphor, ministry naming, and functional mapping just 21 hours after the original repository appeared, with Edict gaining ~2,500 stars versus the original's ~36. The case highlights AI-era ethics: 'reference design + AI rewrite' defeats code plagiarism detection while original architectural and design decisions may be laundered. Analysis dated 2026-03-09.

Edict Project Deep Dive

One-Line Definition

Edict is an OpenClaw-based multi-agent collaboration system that uses China's ancient Three Departments and Six Ministries (sansheng liubu) system as its metaphorical framework.

Core Idea

AI agents play the roles of imperial court officials. The user acts as the 'emperor' issuing decrees, and each departmental agent receives and executes the orders.

Three Departments and Six Ministries Architecture

| Layer | Ancient Office | Agent Function | |-------|----------------|----------------| | Three Departments | Zhongshu | Task intake and decomposition | | | Menxia | Review and approval | | | Shangshu | Task delegation and coordination | | Six Ministries | Personnel | HR / agent management | | | Revenue | Resources / budget management | | | Rites | External interaction / protocol | | | War | Execution / operations | | | Justice | Compliance / audit | | | Works | Engineering / technical implementation |

Workflow

1. Morning briefing: each morning, summarize yesterday's pending items and today's tasks 2. Decree: user issues a task command 3. Zhongshu decomposition: large tasks are split into subtasks 4. Shangshu assignment: subtasks are distributed to ministries 5. Departmental execution: ministry agents execute and report back 6. Review and approval: results are audited and aggregated

Technical Highlights

9 specialized AI agents: separation of duties avoids the context bottleneck of a single agent
Real-time dashboard: web-based visualization of each department's status
Model configuration: different agents can use different LLM models
Complete audit trail: all operations are traceable
Docker support: easy deployment

The Controversy

Timeline:

boluobobo-ai-court-tutorial (original author): first published 2026-02-22
edict (cft0808): created ~21 hours later on the evening of 2026-02-23

Allegations: 1. Identical 'Three Departments and Six Ministries' metaphor framework 2. Identical ministry naming and function mapping 3. The repository was initially named openclaw-sansheng-liubu and only renamed to edict on 2/24

Current status: edict has ~2,500 stars; the original author has ~36 stars.

Project Links

Edict: https://github.com/cft0808/edict
Original author: https://github.com/wanikua/boluobobo-ai-court-tutorial
Controversy issue: https://github.com/cft0808/edict/issues/55

Personal Assessment

The core idea itself is valuable — using a familiar ancient bureaucratic metaphor to abstract multi-agent collaboration lowers the comprehension barrier. But it exposes an ethical dilemma of the AI era: the 'reference design + AI rewrite' approach defeats code-level plagiarism detection, while the truly original parts — architecture, concepts, and design decisions — risk being laundered as content.

Analysis date: 2026-03-09