← 返回 Skills 市场
golikegod

Workspace Guardian

作者 Golikegod · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
57
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install lobster-workspace-guardian
功能描述
Enforce workspace structure, file placement, naming conventions, memory lifecycle, and safety for AI project organization and cleanup tasks.
使用说明 (SKILL.md)

Workspace Guardian

Enforce workspace discipline — consistent structure, naming, memory lifecycle, and safety boundaries.


Quick Reference

workspace/
├── projects/   YYYYNNNN_name     — source code only
├── knowledge/  topics|articles|research|assets — knowledge files only
├── output/     wechat|docs|slides|preview — final deliverables
├── memory/     YYYY-MM-DD.md → archive/ (14d+) — never deleted
├── logs/       run-time logs
├── .temp/      temporary files — delete after use
└── slides/     presentations

Core Rules (6 Tenets)

1. System Files Stay Outside

No workspace files in OpenClaw's own directories (~/.openclaw/ except workspace/, ~/.openclaw/canvas/, etc.).

2. Workspace Is the Only Writable Zone

All AI-generated files → workspace/ subtree only. Never write to Desktop/, Downloads/, ~/, or system paths.

3. Scattered Files → Immediate Redirect

Any file not in its designated directory → move to correct location, not delete.

4. One Config Backup Only

Configuration or .json-based projects keep one archived/expired backup, not piles of timestamps.

5. Temp Files → Never Overnight

.temp/ files: delete after task completion. >7 days stale → cleanup on next session start.

6. Skills Source Separation

ClawHub-installed skills in skills/. User-built local skills also in skills/. No double installation.


File Routing Rules

Content Type Destination
Source code/projects projects/YYYYNNNN_name/
Knowledge/research `knowledge/topics
WeChat articles output/wechat/
Word/PDF docs output/docs/
Presentations output/slides/
Preview files output/preview/ (delete after publish)
Daily logs memory/YYYY-MM-DD.md
Temp work .temp/temp_* (delete after use)

Naming Standards

Rule Pattern
Project dirs YYYYNNNN_project-name
Output files {description}_v1|v2|final.{ext}
Avoid spaces, Chinese characters in dir names
Use - (hyphen) as separator
Temp files temp_{description}_{uuid}.{ext}

Authorization Levels

Level Action
✅ Autonomous File reads, scheduling, memory updates, knowledge base management
⚠️ Need Auth External sends (email/social), destructive ops (delete/move/cleanup), sub-agent creation
🚫 Never System security changes, private data leaks, impersonation

Red Lines

  • No private data leaks
  • No destructive commands without explicit confirmation
  • No external sends without authorization
  • No safety settings modification

Scripts

  • scripts/workspace-cleanup.py — Scan and clean temp/stale files
  • scripts/validate-workspace.py — Validate workspace structure against these rules

References

  • references/naming-conventions.md — Detailed naming rules
  • references/memory-tiering.md — Memory lifecycle management
  • references/safety-boundaries.md — Detailed safety rules
安全使用建议
Install this only if you want an agent to enforce a specific workspace layout. Before allowing cleanup, moves, or deletion, review the dry-run output and confirm the exact workspace path. Consider overriding the Chinese-character naming restriction if it conflicts with your content or team conventions.
能力评估
Purpose & Capability
The stated purpose, directory rules, memory tiering, naming guidance, validation script, and cleanup script fit together as workspace hygiene functionality.
Instruction Scope
The trigger list includes broad phrases such as clean up and project setup, and the skill discusses moving and deleting files; however destructive actions, cleanup, external sends, and sub-agent creation are explicitly marked as requiring authorization.
Install Mechanism
The artifact consists of SKILL.md, reference markdown files, and two Python helper scripts, with no declared dependencies, package install steps, credential setup, or network service configuration.
Credentials
The cleanup helper is dry-run by default and with --apply deletes only stale files under the selected workspace .temp directory, but users should confirm the target workspace path before applying changes.
Persistence & Privilege
No background service, privilege escalation, credential/session access, external transmission, or persistent auto-start mechanism was found.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install lobster-workspace-guardian
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /lobster-workspace-guardian 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
lobster-workspace-guardian v1.0.0 - Initial release for enforcing workspace discipline among AI agents - Defines strict directory structure, naming conventions, and memory management processes - Implements routing rules for all major file/content types - Specifies temp file handling and backup procedures to reduce clutter - Establishes authorization levels and red lines for safety and privacy - Includes scripts for workspace validation and cleanup - Provides reference documents for naming, memory tiering, and safety
元数据
Slug lobster-workspace-guardian
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Workspace Guardian 是什么?

Enforce workspace structure, file placement, naming conventions, memory lifecycle, and safety for AI project organization and cleanup tasks. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 57 次。

如何安装 Workspace Guardian?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install lobster-workspace-guardian」即可一键安装,无需额外配置。

Workspace Guardian 是免费的吗?

是的,Workspace Guardian 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Workspace Guardian 支持哪些平台?

Workspace Guardian 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Workspace Guardian?

由 Golikegod(@golikegod)开发并维护,当前版本 v1.0.0。

💬 留言讨论