openclaw-config

Summary

This workspace provides a flexible configuration and extension setup for OpenClaw. It includes agent definitions, authentication profiles, model providers, skill modules, and persistent memory data for customizing and automating OpenClaw agent behaviors.

Included Assets

• openclaw.json: Global OpenClaw workspace and environment configuration.
• skills/
  • capability-evolver: GEP-powered self-evolution engine for AI agents.
  • deep-research-pro-1.0.2: Multi-source research skill with cited report generation.
  • openclaw-workspace: Utilities for optimizing and managing workspace markdown files.
  • stock-watcher: Skill for tracking and managing Chinese A-share stock watchlists.
• agents/
  • main/agent/: Model and auth profiles for the "main" agent.
  • quant-advisor/agent/: Model and auth profiles for the "quant-advisor" agent.
  • quant-advisor/sessions/: Example agent session data.
• memory/: Log, state, and imported data files for agent memory and persistent context.

How to Use

1. Set up your secrets and authentication
   Update openclaw.json with valid file paths or credentials for your key providers as needed for your environment.

2. Configure agents and models

   • Edit agent-specific configuration files in agents/main/agent/ or agents/quant-advisor/agent/ for model settings and authentication profiles.
   • Update or create new profiles as required by your use case.

3. Extend with skills

   • Place skill code in the skills/ directory.
   • Refer to each skill’s README for installation or usage details.
   • Examples:
     • Use capability-evolver for guided AI prompt evolution.
     • Use deep-research-pro-1.0.2 for thorough research tasks.
     • Use stock-watcher for stock tracking via scripts.

4. Manage workspace and memory

   • Use utilities in skills/openclaw-workspace/ for organizing markdown-based agent state, system prompts, and knowledge files.
   • Memory files under memory/ are used for persistence and bridge agent recall.

Notes

• This workspace does not include a web interface or built-in marketplace integration; all configuration is file-based.
• Ensure secrets and API keys are managed securely and are not committed to shared repositories.
• Refer to individual skill READMEs for prerequisites (e.g., Node.js) and advanced usage.
• Memory files and logs may contain sensitive data—review contents before sharing.
• Configuration and paths are designed to be portable for different OpenClaw deployments.