Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计提醒

workspace-guide工作区指南

Agent Skill

workspace-guide 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。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

294

周安装

8

GitHub Stars

公开资料未说明

下载量

65
CodexClaudeCursorGemini CLI

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

unknown

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:workspace-guide(工作区指南)
来源仓库:https://github.com/tacuchi/workspace-tools
仓库路径:skills/workspace-guide
安装命令:
npx skills add https://github.com/tacuchi/workspace-tools --skill workspace-guide
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 npx skills 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。

skills.shnpx skills
npx skills add https://github.com/tacuchi/workspace-tools --skill workspace-guide

简介

workspace-guide 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。

  • 适用于根据关键词或任务场景从来源线索中组织工作区相关内容。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装并使用。
  • 建议确认权限范围和维护状态,避免触发联网或文件读写操作。
  • 可结合原始 README 进一步核验具体功能和调用方式。

SKILL.md

Workspace Conventions

These conventions apply whenever you detect workspace markers in the project:

  • Directories: docs/, tools/
  • Instruction files: AGENTS.md, CLAUDE.md, GEMINI.md
  • Numbered files: NNN-*.ext pattern (e.g., 001-arquitectura.md)

If any of these markers exist, follow all conventions below automatically.

1. Where Does This Go?

What are you creating?Destination
Architecture/design documentationdocs/arquitectura/NNN-<name>.md
Work plan or analysisdocs/planes/NNN-<name>.md
Architecture Decision Recorddocs/decisiones/NNN-<name>.md
Manual, setup, or config guidedocs/manuales/NNN-<name>.md
Functional/technical specdocs/especificaciones/NNN-<name>.md
SQL script or data analysisdocs/scripts/<topic>/NNN-<name>.sql
Executable tool (scripts, CLIs)tools/scripts/<name>/ (with SKILL.md)
Instruction-only guidancetools/skills/<name>/ (SKILL.md only)

2. Directory Structure

docs/ — Project documentation

docs/
├── arquitectura/       ← architecture and design (mandatory)
├── planes/             ← work plans and specs
├── decisiones/         ← ADRs (Architecture Decision Records)
├── manuales/           ← installation, configuration, deployment guides
├── especificaciones/   ← functional and technical specs
└── scripts/            ← SQL, analysis scripts, DDL
    ├── ddl/
    └── analisis/

tools/ — Scripts and skills

tools/
├── scripts/            ← executable code (Python, Bash, etc.) + SKILL.md
│   └── <tool-name>/
│       ├── SKILL.md
│       └── 001-script.sh
└── skills/             ← instruction-only (no executable code), SKILL.md only
    └── <skill-name>/
        └── SKILL.md

3. File Numbering

Format: NNN-descriptive-name.ext (3-digit: 001, 002,...) Before creating a file, check the highest existing number and increment by 1.

4. Rules

  • Every tool in tools/scripts/ or tools/skills/ MUST have a SKILL.md with description in frontmatter.
  • SKILL.md can also exist inside src/, lib/, etc. when the user requests it.
  • Save work plans in docs/planes/NNN-<plan>.md.
  • Save architectural decisions in docs/decisiones/NNN-<decision>.md.

5. NEVER Do This

  • NEVER create unnumbered files in docs/ — breaks chronological ordering, causes confusion when switching agents
  • NEVER put executable code in tools/skills/ — skill discovery treats skills/ as instruction-only; executables won't be found by runners
  • NEVER create a SKILL.md without a description in frontmatter — the tool becomes invisible to all agent skill discovery systems
  • NEVER duplicate content between the instruction file and docs/ — creates drift; instruction file should only reference docs/
  • NEVER create docs/planes/ or docs/decisiones/ during initial setup — they fill on demand; empty dirs clutter the workspace
  • NEVER create a numbered file without checking the current highest number (ls docs/<subdir>/ first)
  • NEVER use a number already taken — always increment from the highest existing number

6. Existing Tools

Before creating a new tool, check these locations for existing ones:

  1. tools/scripts/ — project executable tools
  2. tools/skills/ — project instruction skills
  3. .agents/skills/ — synced project skills
  4. .claude/skills/ — Claude-specific skills
  5. .gemini/skills/ — Gemini-specific skills

Do not create duplicates. If a similar tool exists, update it instead.

7. References

When creating a tool, follow references/skill-template.md
For instruction file format, see references/instruction-file-guide.md

8. Related Tools

workspace-setup (initial step)

If the project has a monolithic instruction file (>200 lines from /init), suggest:

The instruction file is large. You can split it into docs/arquitectura/ using workspace-setup: npx skills add tacuchi/workspace-tools@workspace-setup

workspace-update (after creating tools)

After creating tools in tools/scripts/ or tools/skills/, suggest syncing:

New tools created. Sync SKILL.md to .agents/skills/ with: npx workspace-update sync

This makes project tools discoverable by all compatible agents.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

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

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

Codex

33.15%
按下载量换算22

Claude

31.17%
按下载量换算20

Cursor

19.76%
按下载量换算13

Gemini CLI

8.72%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

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