openclaw-workspace

Summary

A workspace for building, managing, and customizing AI agent capabilities in OpenClaw. Organize your agent’s memory, tools, and personality in one place, with skill backups to extend functionality for tasks like blog writing, research, interview design, documentation integration, and social media scheduling.

Included Assets

• AGENTS.md — Main workspace operating guide; session and memory procedures.
• MEMORY.md — Long-term memory file with detailed rules and critical user instructions.
• SOUL.md — Agent “personality,” operating principles, and boundaries.
• TOOLS.md — Local environment notes and tools cheat sheet.
• Skills and Backups:
  • Blog Writer
  • Feishu Doc integration
  • Interview Designer
  • Obsidian-Ontology Sync
  • Research Paper Writer
  • Social Media Scheduler
  • Supabase Postgres Best Practices

How to Use

1. First Run: Follow onboarding notes found in AGENTS.md if present.
2. Every Session:
   • Read SOUL.md (for agent’s guiding principles).
   • Read USER.md (if present, for current user context).
   • Review memory/YYYY-MM-DD.md and MEMORY.md for session continuity.
3. Backing Skills: Each skill provides its own README with usage instructions (see skills-backup/agents-skills/ and skills-backup/all-skills/agents-skills/).
4. Customize Tools: Edit TOOLS.md to reflect devices, environment, and local configuration.

Notes

• Memory and continuity depend on consistently updating MEMORY.md and session logs.
• Strict boundaries and security are defined in MEMORY.md — always follow user deletion and action rules.
• Skills are modular and may require additional configuration as documented in their respective directories.
• Tooling notes for python and automation (like uv usage) are provided in TOOLS.md for reference.
• Back up your custom changes outside skills-backup/ before updating skills.

This workspace is best suited for power users who want detailed control and context persistence for their OpenClaw agent’s operation.