Skip to main content

䷀ Yi-Jing Agent — I Ching Six Lines AI Agent Lifecycle Framework

License: MIT CI PRs Welcome PyPI

Fusing ancient I Ching wisdom with modern AI agent architecture. A structured lifecycle framework that prevents goal drift, enables graceful degradation, and brings philosophical depth to AI agent execution.


🧠 What is Yi-Jing Agent?

Modern AI agents suffer from three structural problems:

  1. Goal Drift — Agents forget the original intent mid-task
  2. Blind Impulse — Agents rush to call APIs without proper planning
  3. Crash with No Recovery — One error cascades into total failure

The I Ching (易經) Six Lines framework solves this by imposing a 6-stage lifecycle on every task, where each stage (爻) has strict behavioral rules and output gates:

Line           Agent Behavior              Output Gate
────────────────────────────────────────────────────────────────
初爻 (Hidden Dragon) → Pure understanding, NO actions → Task Graph
二爻 (Dragon in Field) → Sandbox & prototyping   → Feasibility Report
三爻 (Alert Dragon) → 3D Reflection Engine      → Safety Report
四爻 (Leaping Dragon) → Authorization Gate       → Confirm / Pivot
五爻 (Flying Dragon) → Full execution            → Core Payload
上爻 (Regretful Dragon) → Memory compression     → LTM Write

🔄 Dynamic Fault Tolerance with Hexagram Mutation

When any line encounters an error, the system triggers dynamic hexagram mutation (動爻變卦) — XOR flipping that line's bit and looking up the new hexagram's strategy:

Initial (111111 ䷀ Qian) ──[API Timeout at Line 5]──→ flip bit5
                                                    ↓
Mutation (111011 ䷉ Lu) ──→ Strategy: degrade to Line 2 sandbox

🧿 3D Reflection Engine (錯綜互卦)

At Line 3, the agent performs a mandatory three-dimensional reflection:

Dimension I Ching Concept Question
🔍 Interlocking (互卦) Hidden motive What does the user really need?
⚔️ Opposite (錯卦) Adversarial view Where would this plan fail if everything is wrong?
👁️ Reversed (綜卦) User perspective How will the end-user experience this output?

🚀 Quick Start

Prerequisites

  • Python 3.10+
  • An AI agent platform (Hermes Agent, OpenClaw, AutoGPT, etc.)

Install

git clone https://github.com/yayasasapig/yi-jing-agent.git
cd yi-jing-agent
# Copy the skill to your agent's skill directory
cp -r skill/* ~/.hermes/skills/
# Or use the Python library directly
pip install -e .

Basic Usage (Python)

from src.agent_state import YiJingAgentState
from src.executor import YiJingAgentExecutor
import asyncio

async def main():
    executor = YiJingAgentExecutor()
    result = await executor.execute("Analyze this website's pricing strategy")
    print(f"Status: {result['status']}")
    print(f"Hexagram path: {' → '.join(
        t.transition_name for t in result['hexagram_history']
    )}")

asyncio.run(main())

For Hermes/OpenClaw Agents

Load the skill in your SOUL.md or task prompt:

Load the `yi-jing-agent` skill to activate the 六爻 lifecycle.

📂 Repository Structure

yi-jing-agent/
├── README.md                 # ← You are here
├── LICENSE                   # MIT
├── CONTRIBUTING.md           # How to contribute
├── docs/
│   ├── 六爻AI-Agent架構設計書.md  # Original Chinese design document (37KB)
│   └── architecture-overview.md  # English architecture summary
├── skill/
│   └── SKILL.md              # Hermes/OpenClaw skill definition
├── src/
│   ├── agent_state.py        # Core YiJingAgentState
│   ├── executor.py           # Full lifecycle executor
│   ├── hexagram_table.py     # 64 hexagrams → strategies
│   ├── reflection.py         # 3D reflection engine
│   └── yao_positions.py      # Line definitions & enums
├── examples/
│   └── six-yao-lifecycle-demo.md  # Walkthrough demo
└── .github/
    └── workflows/
        └── ci.yml            # GitHub Actions CI

📖 Documentation

Resource Language Description
Design Document 🇭🇰 Chinese (Cantonese) Full 37KB architectural blueprint
Architecture Overview 🇬🇧 English English summary of the framework
Skill Definition 🇭🇰 Chinese Ready-to-use agent skill
Lifecycle Demo 🇭🇰 Chinese Real-world walkthrough with摘日

🌍 Community & Contributions

This project is open for everyone to use, remix, and improve!

  • Share your hexagram strategies — Found a new failure mode? Add it to the table!
  • Translate — Help translate the docs to Japanese, Korean, English
  • Build tools — Dashboard, VSCode extension, visualization
  • Report issues — Found a bug in the lifecycle? Open an issue!

See CONTRIBUTING.md for guidelines.


📜 License

MIT — Free for personal and commercial use.


🙏 Credits

  • Concept: yayafu × Gemini conversation distillation
  • Implementation: Orchestrator Nova (Hermes Agent)
  • Platform: OpenClaw Agent Team

「An agent that knows its time uses wisdom to move mountains.」 「識得時機嘅 Agent,用智慧四兩撥千斤。」

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

yi_jing_agent-0.1.0.tar.gz (35.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

yi_jing_agent-0.1.0-py3-none-any.whl (23.3 kB view details)

Uploaded Python 3

File details

Details for the file yi_jing_agent-0.1.0.tar.gz.

File metadata

  • Download URL: yi_jing_agent-0.1.0.tar.gz
  • Upload date:
  • Size: 35.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for yi_jing_agent-0.1.0.tar.gz
Algorithm Hash digest
SHA256 9e5f4300dc45b3ab81cd17b5ebb4849c222059e3b9cba43818ee910632148103
MD5 352a07c914837d4be87f0898e61a0b32
BLAKE2b-256 2be39a0cfa75d7c299bda28d2b75d675c8aae106bef21aaa3e4a24cc5c6e357f

See more details on using hashes here.

Provenance

The following attestation bundles were made for yi_jing_agent-0.1.0.tar.gz:

Publisher: publish.yml on yayasasapig/yi-jing-agent

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file yi_jing_agent-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: yi_jing_agent-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 23.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for yi_jing_agent-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 7830815e3f28dbe984b2a9d6ee3171defdb087892468418a271ab137ffdcbee8
MD5 731867fd6009ac70943a71d4cf80c37b
BLAKE2b-256 7db668ee3348cd58704c050efbe7d2c2d32c43f3a5ac21c2799c054898f092d1

See more details on using hashes here.

Provenance

The following attestation bundles were made for yi_jing_agent-0.1.0-py3-none-any.whl:

Publisher: publish.yml on yayasasapig/yi-jing-agent

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page