ClawLodge Mock Publish

Summary

This is an OpenClaw workspace for building and customizing intelligent agent workflows. It includes agents, modular skills, and memory components. The workspace is designed for personal assistant scenarios, with a focus on managing job searches, integrating email/calendar, and self-improvement through continuous learning.

Included Assets

• Agents and character definition

  • AGENTS.md: Main workspace home and short-term/long-term memory policies
  • SOUL.md: Defines agent personality, boundaries, and style

• Memory and continuity

  • MEMORY.md: Long-term, user-curated memory
  • memory/YYYY-MM-DD.md: Daily session logs (e.g., memory/2026-03-14.md)
  • memory/resume-jiaxuan.md: User resume memory

• Tools and environment

  • TOOLS.md: Local environment and integration notes

• Skills

  • skills/github/: Automate GitHub interactions via the gh CLI
  • skills/gog/: Google Workspace CLI for Gmail, Calendar, Drive, Contacts, Sheets, Docs
  • skills/linkedin_apify_jobs/: Fetch, filter, and rank LinkedIn jobs (with Apify fallback, Notion optional)
  • skills/self-improving-agent/: Log errors, feature requests, and learnings for agent self-improvement

How to Use

1. First setup

   • Read SOUL.md to understand agent philosophy and boundaries.
   • Review AGENTS.md for guidance on session start, memory refresh, and context capture.

2. Session startup

   • On each wake, consult today's and yesterday’s memory/YYYY-MM-DD.md files for session context.
   • Only load MEMORY.md in direct (main) sessions for privacy.

3. Extend or configure skills

   • Credentials and secrets (e.g., Notion tokens, Google OAuth) are managed locally—see TOOLS.md for placement and integration.
   • Skill modules provide markdown guides (e.g., SKILL.md) for usage examples.

4. Job search and reporting

   • Use skills/linkedin_apify_jobs/ to automate job discovery, ranking, and reporting.
   • Optionally sync job recommendations to Notion (credentials required).

5. Self-improvement and error tracking

   • Log new learnings, corrections, or failed commands to .learnings/ markdown files in skills/self-improving-agent/.

Notes

• Workspace does not include external orchestration, UI, or deployment scripts—intended for use with OpenClaw-compatible infrastructure.
• All sensitive information (tokens, secrets, private details) should be managed in local .secrets/ files, never hardcoded or committed.
• Memory is user-visible, editable, and should be reviewed regularly for relevancy.
• Agents are opinionated and improve through real feedback—review/update .learnings/ logs often.
• Each skill is self-contained, with installation/setup instructions in its SKILL.md.
• Designed for single-user or trusted small-group contexts; group-chat boundaries are explicitly defined in SOUL.md and AGENTS.md.
• See each markdown file for more detailed, role-specific instructions.

This workspace is for building deeply personalized digital agents, balancing utility, adaptability, and strong user boundaries.