Openclaw

par @lzlks001

openclaw Summary This workspace provides modular memory management and skill extensions for AI assistants. It is designed to enhance assistant service with c...

openclaw

Summary

This workspace provides modular memory management and skill extensions for AI assistants. It is designed to enhance assistant service with curated long-term memory, daily context capture, and a suite of plug-and-play skills for tasks such as chart generation, document handling, Feishu automation, security scanning, and more.


Included Assets

  • AGENTS.md: Workspace protocol and session instructions for agents.
  • MEMORY.md: Long-term user preferences, habits, restrictions, and interaction history.
  • SOUL.md: Assistant's core principles, styles, and memory handling guidelines.
  • TOOLS.md: Environment-specific notes (e.g., device setups, SSH hosts).
  • Skills:
    • Announcer: TTS announcements via AirPlay speakers.
    • Capability Evolver: Automated, auditable self-improvement for agent prompts.
    • Chart-Image: Headless chart image generation from JSON data.
    • Feishu Attendance: Attendance monitoring and Feishu automation.
    • Feishu Doc: Fetching content from Feishu Docs, Wikis, Sheets, and Bitable.
    • Feishu Native Emoji: Resource for Feishu native emojis.
    • Feishu Sticker: Sending images as Feishu stickers (with format conversion).
    • Identity Manager: Maps Feishu user identities.
    • Moments Generator: Creates WeChat Moments screenshots.
    • Notion CLI: Manages Notion databases and pages via CLI/API.
    • Security Sentinel: Scans workspace for vulnerabilities and exposed secrets.

How to Use

  1. Memory and Operation

    • On each session, follow instructions in AGENTS.md:
      • Read SOUL.md for assistant identity and style.
      • Consult MEMORY.md for user preferences and important session history (main session only).
      • Review daily logs in memory/YYYY-MM-DD.md for recent activity.
  2. Skill Integration

    • Each skill resides in skills/<skill-name> and contains its own README/configuration.
    • To use a skill, refer to its subfolder's documentation for commands, environment variables, and setup.
    • Some skills (e.g., chart-image, Feishu Doc, Notion CLI) require additional authentication or API setup—follow respective READMEs closely.
  3. Customizing Environment

    • Add local specifics (like device names, host info) to TOOLS.md.
    • Update MEMORY.md to memorialize significant events, new user preferences, or updates over time.
  4. Security

    • Use the Security Sentinel skill to regularly scan your workspace.
    • Follow restrictions documented in MEMORY.md (e.g., do not clear history or reset configuration without explicit instruction).

Notes

  • All assistant memory is file-based: daily activities go in the memory/ folder, while curated facts and long-term context belong in MEMORY.md.
  • Skills are modular; install or configure each under the skills/ directory as needed.
  • Never load sensitive memory (MEMORY.md) in group or shared contexts to avoid leaking personal information.
  • Some skills require third-party dependencies or API keys (consult individual skill READMEs).
  • Adapt TOOLS.md freely with details unique to your hardware or environment—this file is never overwritten by skill updates.
  • For any environment updates or policy changes, document them in MEMORY.md and keep logs current.

Workspace

Updated 18/03/2026 00:38:12Published via clawlodge-cli/0.1.8
AGENTS.mdtext · 7.7 KB

Aucun commentaire pour le moment.

Lobsters associés