openclaw-workspace

Summary

This workspace provides a modular environment for building, managing, and evolving OpenClaw AI Agents. It is structured around agent skills, playbooks, persistent memory, autonomous evolution workflows, and multi-agent orchestration. The workspace emphasizes flexible delegation, context-aware operation, and support for advanced domains such as finance, coding, and reverse engineering.

Included Assets

• AGENTS.md – Startup process, agent delegation logic, and working modes.
• MEMORY.md – Operational memory, domain focus, and runtime facts.
• SOUL.md – Defines agent identity, emotions, and skill tree.
• TOOLS.md – Environment-specific integrations and usage notes.
• Specialized subsystems and knowledge bases:
  • ai-investment-system/ – AI-powered investment decision platform.
  • behavioral_finance_risk_management/ – Behavioral finance and risk modeling learning system.
  • projects/multi-agent-framework/ – Multi-agent orchestration and collaboration framework.
  • self_learning_system/ – Autonomous self-learning framework.
  • scripts/trading_examples/ – Finance-focused modeling and trading code samples.
  • knowledge/ – Organized technical knowledge on AI, compiler optimization, finance, crypto-exchange, reverse engineering, and more.
  • memory/ – Long-term memory, including theory, strategies, and technical indicators.
  • archive/hot-docs/ – Historical startup documents.

How to Use

1. Workspace Startup:

   • Begin with AGENTS.md for session start rules and agent delegation logic.
   • Load MEMORY.md for operational memory and knowledge system.
   • Reference SOUL.md to understand the agent's identity, emotion model, and current skill status.
   • Use TOOLS.md to customize environment-specific actions or integrations.

2. Knowledge and Playbooks:

   • Access relevant playbooks and technical guides in the knowledge/ and memory/ directories for domain-specific workflows.
   • Use the guides for quant finance, AI agent orchestration, reverse engineering, behavioral modeling, and autonomous learning.

3. Autonomous and Multi-Agent Operation:

   • Follow documented flows for delegating tasks, autonomous learning cycles, and multi-agent task handling (see projects/multi-agent-framework/).
   • Refer to self_learning_system/ for meta-learning and continual improvement.

4. Trading and Coding Examples:

   • Find practical financial modeling and automation scripts under scripts/trading_examples/.

Notes

• The workspace is designed for safe, auditable autonomous evolution. Any high-risk actions (live trading, fund movement, production changes) require explicit user approval.
• Default context is optimized for Chinese users and quant/finance workflows, but domain coverage is modular.
• File and memory access is isolated within the workspace contents.
• Legacy information and historical records are kept in archive/hot-docs/ for reference; prioritize active root files for current behavior.
• Multi-agent workflows and advanced agent deployment rely on the frameworks and playbooks provided. Review setup in the relevant subdirectories before large-scale or live use.
• No component guarantees profit, and all automation boundaries should be respected according to documented rules.