Token导航 LogoToken导航TokenDH.com
效率只读clawhub未标认证来源可访问clear审计提醒

talonforge-workspace-hygiene塔隆福奇工作空间卫生

Agent Skill

talonforge-workspace-hygiene 用于整理文档、README、Markdown 和说明材料,适合在 OpenClaw 中需要把零散信息整理成结构清晰的文档时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

1,944

周安装

81

GitHub Stars

公开资料未说明

下载量

648
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:talonforge-workspace-hygiene(塔隆福奇工作空间卫生)
来源仓库:https://github.com/casperzinou/talonforge-workspace-hygiene
安装命令:
openclaw skills install talonforge-workspace-hygiene
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install talonforge-workspace-hygiene

简介

自动实施工作区文件大小限制、归档陈旧或臃肿的文件,并组织根内容以维护高效的 AI 代理上下文。

SKILL.md

Workspace Hygiene Skill

Auto-maintains workspace cleanliness. Prevents context bloat that burns tokens on every session boot.

Triggers

  • Cron: runs daily at 06:00 Europe/Berlin
  • Manual: when Zinou says "clean up", "audit workspace", "hygiene check"
  • Auto: after any session that creates 10+ files

Rules

1. BOOT FILE SIZE LIMITS (hard limits, enforce or trim)

  • AGENTS.md: max 50 lines
  • MEMORY.md: max 80 lines
  • STATE.md: max 60 lines
  • SOUL.md: no limit (identity, changes rarely)
  • USER.md: max 40 lines
  • IDENTITY.md: max 30 lines
  • TOOLS.md: no limit (reference, not loaded often)
  • HEARTBEAT.md: max 20 lines

If any file exceeds its limit:

  1. Identify bloat (action logs, duplicate info, stale references)
  2. Move old content to memory/archive/YYYY-MM-DD-<topic>.md
  3. Leave a 1-line pointer: See archive/YYYY-MM-DD-<topic>.md
  4. Log the trim in today's daily note

2. ROOT WORKSPACE RULES

  • NO image/video files in workspace root → archive/frames/ or archive/media/
  • NO package.json/package-lock.json in root (belongs in subprojects)
  • NO temp scripts in root → /tmp/ for those
  • NO .secrets in root → always in talonforge/.secrets/
  • Only .md files and directories allowed in workspace root

3. MEMORY/ RULES

  • Only keep files actively needed this week
  • Daily notes older than 3 days → archive (except today's)
  • Stale files (not updated in 7+ days, not referenced by STATE.md) → archive
  • Fragment daily notes (multiple per day) → consolidate into one
  • claude-task.md → archive when >7 days stale, keep latest 500 lines
  • Never delete, always archive to memory/archive/

4. AUTO-ARCHIVE PATTERNS

After any session:

  • If temp files created in /tmp/ → leave them (auto-cleaned)
  • If files created in workspace root → move to proper location
  • If memory/ grows past 10 active files → audit for archiving

5. BLOAT SIGNALS (check for these)

  • MEMORY.md action log >5 entries → consolidate to archive
  • Same info in 2+ files → keep one, archive the other
  • Files with "TODO" or "DRAFT" older than 3 days → flag or archive
  • Empty files (<10 bytes) → delete
  • Duplicate daily notes → merge

6. ARCHIVE STRUCTURE

memory/archive/
  action-log-day1-3.md
  day3-fragments/
  frames/
  media/
  <date>-<topic>.md

7. REPORT FORMAT

After each run, log to today's daily notes:

## Workspace Hygiene (HH:MM)
- Trimmed: X files, Y lines removed
- Archived: Z files
- Boot load: N lines (was M lines)
- Status: CLEAN / NEEDS ATTENTION

Cron Setup

Run: 0 6 * * * (daily at 06:00 Europe/Berlin) This is a lightweight check — most days it does nothing. Only acts when bloat exists.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

95.01%
按下载量换算616

安全审计

VirusTotal

通过

ClawScan

可疑

Static analysis

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills