openclaw-workspace

Summary

This workspace provides organized templates, guides, and operational notes for running agents, managing memory, defining "soul"/personality, and tracking tool/integration details in OpenClaw-based projects. It is structured for use as an operator's or agent's primary working directory, enabling continuity through daily logs and persistent long-term knowledge.

Included Assets

• AGENTS.md — Session routines, guide to identity/context files, memory usage instructions
• SOUL.md — Defines assistant tone, boundaries, core truths, and behavioral style
• MEMORY.md — Curated summary of long-term context and recurring operational jobs
• TOOLS.md — Workspace-specific tool/configuration notes, technical limitations, integrations
• memory/ — Time-stamped daily logs with detailed records (YYYY-MM-DD.md, flow docs)

How to Use

1. Session Startup:

   • Read SOUL.md to understand assistant purpose and constraints.
   • Consult IDENTITY.md and USER.md for agent and user context.
   • Check current and previous day memory/YYYY-MM-DD.md for continuity.
   • In main/direct chat sessions, additionally load MEMORY.md for full background.

2. Daily Logging:

   • Add a new file to memory/ using the format YYYY-MM-DD.md to capture notable events, decisions, and system changes.
   • Summarize and curate critical knowledge to MEMORY.md as needed.

3. Tooling and Integrations:

   • Refer to TOOLS.md before running technical integrations (Node path, Discord, API keys, etc.).
   • Use provided cheatsheets for cameras, SSH, TTS, and limit notes specific to the environment.

4. Agent Operation:

   • Adhere to practices and boundaries in SOUL.md.
   • Use AGENTS.md for runbook instructions and memory handling.
   • Reference daily logs for recent actions and ongoing tasks.

Notes

• Memory Handling:

  • Daily logs in memory/ preserve short-term continuity; MEMORY.md provides long-term, curated context.
  • Security: MEMORY.md is only to be loaded in direct sessions to avoid leaks of sensitive info.

• Cron Jobs/Automation:

  • Several recurring pipeline tasks are tracked and described in MEMORY.md (job search, daily review, content creation, email dispatch).
  • Workflow and schedule specifics (times, models, agent IDs) are documented in MEMORY.md and reflected in daily logs.

• Operational Boundaries:

  • All external actions should be cautious and intentional.
  • No automated messaging on behalf of users without approval.
  • "Plan-first" approaches may be exempt for strictly scheduled cron jobs (see daily logs for exceptions).

• Technical Restrictions:

  • File writes are sandboxed to the workspace.
  • Some integrations (e.g., YouTube, Gmail, Discord) require API keys/tokens configured as described in TOOLS.md.

• Change Tracking:

  • All significant workflow and process changes are recorded in the memory/ daily logs.
  • Use logs for historical context and to inform updates to other workspace files.

This workspace is designed for clarity, security, and continuity — use it as your agent's home base for all OpenClaw operations.