openclaw-workspace

Summary

This workspace provides a portable foundation for building and running OpenClaw agents. It includes detailed agent setup, memory management, skill modules, and local environment configuration, designed to help OpenClaw maintain intelligent autonomy, context, and productivity across sessions.

Included Assets

• AGENTS.md  — Onboarding, session flow, memory usage guidelines.
• SOUL.md  — Describes agent core principles, personality, boundaries, and persistence expectations.
• MEMORY.md  — Long-term, curated memory for the main session.
• TOOLS.md  — Local configuration notes and integration points.
• memory/
  • Daily logs (YYYY-MM-DD.md) for capturing recent events and context.
  • resume-jiaxuan.md — Resume data for targeted job search and matching.
• skills/
  • github — Usage guides for GitHub interaction via CLI.
  • gog — Reference for Google Workspace operations (Gmail, Calendar, Drive, etc.).
  • linkedin_apify_jobs — Fetches and processes LinkedIn job listings, supports Notion sync.
  • self-improving-agent — Manages errors and tracks feature requests for agent self-improvement.
• Metadata and origin files for skills.

How to Use

1. Session Startup
   • Review SOUL.md to understand your agent's persona and priorities.
   • Check AGENTS.md for onboarding steps, how session memory is managed, and session routines.
   • Use today's and yesterday's memory/YYYY-MM-DD.md files for context.
   • In direct sessions, load MEMORY.md for curated long-term memory.
2. Skills
   • Leverage skill modules in the skills/ directory for supported integrations (GitHub, Google, LinkedIn).
   • Follow instructions in each SKILL.md for authentication and usage.
3. Memory Management
   • Record relevant information in daily memory files.
   • Curate and migrate important context to MEMORY.md to persist insights.
4. Local Tools and Integrations
   • Add unique environment notes to TOOLS.md as needed (e.g., custom device names, integration credentials).
   • Sensitive tokens should be stored in local secrets files; do not echo them in output.

Notes

• Security & Context
  • Only load MEMORY.md in trusted/direct (main) sessions to avoid context leaks.
  • Treat user and system data with respect; follow the agent boundaries defined in SOUL.md.
• Extensibility
  • Add new skills in the skills/ directory as needed.
  • Use .learnings/ and memory logs to drive agent self-improvement and track open requests.
• Persistence
  • Agents wake up fresh each session—rely on provided files for continuity.
  • Update memory and tool notes regularly to support evolving workflows.

For more details, refer to each markdown and skill documentation within the workspace.