Token导航 LogoToken导航

Agent Skill · AI 工具

ln-111-root-docs-creator(ln 111 根文档创建者)

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

ln-111-root-docs-creator 是AI 工具类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为可写文件;许可证为 MIT,第三方扫描有提醒项,按安装量在AI 工具类的 2,536 个 Skill 里排第 354。

总安装 6,610周安装 270GitHub Stars 437下载 2,138核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

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:ln-111-root-docs-creator(ln 111 根文档创建者)
来源仓库:https://github.com/levnikolaevich/claude-code-skills
仓库路径:skills/ln-111-root-docs-creator
安装命令:
npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-111-root-docs-creator
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-111-root-docs-creator

这个 Skill 是做什么的

AI 工具类 Skill 用来扩展助手本身的能力,例如接模型、接检索、接外部服务。

用于辅助文档和 Markdown 内容的整理与改写。适合提炼结构、补齐章节、统一术语或检查链接有效性。使用时应保留项目已有事实和路径, 不要把未确认信息写成确定结论;ln-111-root-docs-creator 属于AI 工具类 Skill,可作为该场景下的辅助能力补充。涉及对外文案时需避免过度营销。

来源仓库内路径:skills/ln-111-root-docs-creator

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-1…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要可写文件,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

ln-111-root-docs-creator 怎么安装?

在终端执行 npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-111-root-do…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

ln-111-root-docs-creator 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

ln-111-root-docs-creator 需要什么权限?

权限级别是可写文件。该 Skill 可能写入或修改本地文件,使用前需要确认目标目录和修改范围。

ln-111-root-docs-creator 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

ln-111-root-docs-creator 用的人多吗?

公开数据是累计安装 6,610 次,近一周安装 270 次,下载 2,138 次,GitHub 437 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

ln-111-root-docs-creator 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。

先想清楚的地方

  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理AI 工具类任务、又不想每次都手写提示词的人。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

AI 工具类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
executing-plans

执行计划 · Codex、Claude

1,515,336173,110只读MIT通过
verification-before-completion

完成前的验证 · Codex、Claude

1,319,472173,121权限未标注MIT通过
subagent-driven-development

子 Agent 驱动开发 · Codex、Claude

1,312,416173,138只读MIT通过
copy-editing

复制编辑 · Codex、Claude

1,209,57625,632只读MIT通过
paywall-upgrade-cro

付费专区升级 cro · Codex、Claude

1,030,17625,617只读MIT通过
ln-111-root-docs-creator

ln 111 根文档创建者 · 本页

6,610437可写文件MIT有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Paths: File paths (shared/, references/, ../ln-*) are relative to skills repo root. If not found at CWD, locate this SKILL.md directory and go up one level for repo root. If shared/ is missing, fetch files via WebFetch from https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/{path}.

Root Documentation Creator

Type: L3 Worker

L3 Worker that creates 7 root documentation files using templates and Context Store from coordinator.

Purpose & Scope

  • Creates 7 root documentation files (entry points for AI agents)
  • Receives Context Store from ln-110-project-docs-coordinator
  • Replaces placeholders with project-specific data
  • Self-validates structure and content (22 questions)
  • Never gathers context itself; uses coordinator input

Inputs

From coordinator:

  • contextStore: Key-value pairs with all placeholders

- PROJECT_NAME, PROJECT_DESCRIPTION - TECH_STACK_SUMMARY - DEV_COMMANDS (from package.json scripts) - DATE (current date) - ENABLE_WORKFLOW_PRINCIPLES (optional boolean; default false) — when true, expands the {{WORKFLOW_PRINCIPLES_BLOCK}} placeholder in AGENTS.md with the content of references/templates/agents_md_workflow_principles.md

  • targetDir: Project root directory

MANDATORY READ: Load shared/references/docs_quality_contract.md, shared/references/docs_quality_rules.json, and shared/references/agent_instructions_writing_guide.md (the canonical writing guide for AGENTS.md / CLAUDE.md).

Documents Created

FileTarget SectionsQuestions
AGENTS.mdQuick Navigation, Agent Entry, Critical Rules, (optional) Workflow Principles, Development Commands, MaintenanceQ1-Q6
CLAUDE.md@AGENTS.md import + ## Claude Code delta (≤20 lines total)Q1-Q6
docs/README.mdQuick Navigation, Agent Entry, Documentation Map, MaintenanceQ7-Q13
docs/documentation_standards.mdQuick Reference (60+ requirements), 12 main sections, MaintenanceQ14-Q16
docs/principles.mdCore Principles (8), Decision Framework, Anti-Patterns, Verification, MaintenanceQ17-Q22

Workflow

Phase 1: Receive Context

  1. Parse Context Store from coordinator
  2. Validate required keys present (PROJECT_NAME, PROJECT_DESCRIPTION)
  3. Set defaults for missing optional keys (ENABLE_WORKFLOW_PRINCIPLES defaults to false)

Phase 2: Create Documents

For each document (AGENTS.md, CLAUDE.md, docs/README.md, docs/documentation_standards.md, docs/principles.md):

  1. Check if file exists (idempotent)
  2. If exists: skip with log
  3. If not exists:

- Copy template from references/templates/ - For AGENTS.md: enforce the shared header contract (SCOPE, DOC_KIND, DOC_ROLE: canonical, READ_WHEN, SKIP_WHEN, PRIMARY_SOURCES) and the top-section contract (## Quick Navigation, ## Agent Entry, ## Critical Rules, ## Maintenance) - For AGENTS.md: if ENABLE_WORKFLOW_PRINCIPLES=true, replace the {{WORKFLOW_PRINCIPLES_BLOCK}} placeholder with the full content of references/templates/agents_md_workflow_principles.md; otherwise strip the placeholder line and its leading comment - For CLAUDE.md: use the import-stub template (claude_md_template.md). It contains DOC_ROLE: derived, a single @AGENTS.md line, and a bounded harness-specific delta. Do not inline any AGENTS.md content - For docs/principles.md: prefer normalized principle inputs already present in Context Store - If project-specific principles are absent, keep the template structure and fill only facts supported by current project sources - Replace {{PLACEHOLDER}} tokens with Context Store values - Never leave template markers in published root docs - If data is missing: omit the claim or use a concise neutral fallback, but do NOT emit [TBD:...] - Write file

Root entrypoint rule (canonical model):

  • AGENTS.md is the single canonical source of content. It holds the Critical Rules table, MCP Tool Preferences, Navigation, Development Commands, and optional Workflow Principles.
  • CLAUDE.md is a Claude Code-specific stub that contains @AGENTS.md plus a ## Claude Code delta with harness-specific rules (≤20 lines total, ≤50 absolute max). Do not duplicate AGENTS.md content.
  • Claude Code expands @path imports into the session context at launch, so the imported AGENTS.md is available automatically.
  • Harness delta content: command terminology (/compact, /memory show), storage pointers (~/.claude/projects/<project>/memory/), and Claude-only features (.claude/rules/ with paths: frontmatter, nested on-demand loading).

Phase 3: Self-Validate

For each created document:

  1. Check SCOPE tag in first 12 lines
  2. Check metadata markers (DOC_KIND, DOC_ROLE, READ_WHEN, SKIP_WHEN, PRIMARY_SOURCES)
  3. For AGENTS.md: check Quick Navigation, Agent Entry, Critical Rules, Maintenance
  4. For CLAUDE.md: check @AGENTS.md import line is present and the file is ≤50 lines
  5. Check required sections (from questions_root.md)
  6. Check docs-quality contract compliance (no forbidden placeholders, no leaked template metadata)
  7. Check POSIX endings (single newline at end)
  8. Auto-fix issues where possible

Phase 4: Return Status

Return to coordinator:

{
  "created_files": ["AGENTS.md", "CLAUDE.md", "docs/README.md", "docs/documentation_standards.md", "docs/principles.md"],
  "skipped_files": [],
  "quality_inputs": {
    "doc_paths": ["AGENTS.md", "CLAUDE.md", "docs/README.md", "docs/documentation_standards.md", "docs/principles.md"],
    "owners": {
      "AGENTS.md": "ln-111-root-docs-creator",
      "CLAUDE.md": "ln-111-root-docs-creator",
      "docs/README.md": "ln-111-root-docs-creator",
      "docs/documentation_standards.md": "ln-111-root-docs-creator",
      "docs/principles.md": "ln-111-root-docs-creator"
    }
  },
  "validation_status": "passed"
}

Critical Notes

Core Rules

  • Idempotent: Never overwrite existing files; skip and log
  • No context gathering: All data comes from coordinator's Context Store
  • Publishable output: Root docs must not contain [TBD:...], TODO, or leaked template metadata
  • Language: All root docs in English (universal standards)
  • SCOPE tags: Required in first 10 lines of each file (HTML comments are stripped from Claude Code's injected context but stay visible to maintainers and auditors)
  • Canonical-stub root model: AGENTS.md is the single source; CLAUDE.md is an @AGENTS.md import plus bounded harness delta. See shared/references/agent_instructions_writing_guide.md for rationale and anti-patterns.

NO_CODE_EXAMPLES Rule (MANDATORY)

Root documents define navigation and standards, NOT implementations:

  • FORBIDDEN: Code blocks, implementation snippets
  • ALLOWED: Tables, links, command examples (1 line)
  • TEMPLATE RULE: All templates include <!-- NO_CODE_EXAMPLES:... --> tag - FOLLOW IT

Stack Adaptation Rule (MANDATORY)

  • All external links must match project stack (detected in Context Store)
  • .NET project → Microsoft docs; Node.js → MDN, npm docs; Python → Python docs
  • Never mix stack references (no Python examples in.NET project)

Format Priority (MANDATORY)

Tables/ASCII > Lists (enumerations only) > Text (last resort)

Runtime Summary Artifact

MANDATORY READ: Load shared/references/docs_generation_summary_contract.md

Accept optional summaryArtifactPath.

Summary kind:

  • docs-generation

Required payload semantics:

  • worker = "ln-111"
  • status
  • created_files
  • skipped_files
  • quality_inputs
  • validation_status
  • warnings

Write the summary to the provided artifact path or return the same envelope in structured output.

Definition of Done

  • Context Store received and validated
  • Root documents created (or skipped if exist)
  • All placeholders replaced; no [TBD:...] markers or template metadata remain in root docs
  • CLAUDE.md contains exactly one @AGENTS.md line and a bounded harness delta (≤50 lines total)
  • If ENABLE_WORKFLOW_PRINCIPLES=true: the workflow principles shard is expanded in AGENTS.md; if false: the placeholder line is stripped
  • Self-validation passed (SCOPE, metadata markers, top sections, Maintenance, POSIX)
  • Actuality verified: all document facts match current code (paths, functions, APIs, configs exist and are accurate)
  • Status returned

Reference Files

  • Templates: references/templates/agents_md_template.md, references/templates/claude_md_template.md, references/templates/agents_md_workflow_principles.md, references/templates/docs_root_readme_template.md, references/templates/documentation_standards_template.md, references/templates/principles_template.md
  • Questions: references/questions_root.md (Q1-Q22)
  • Writing guide: shared/references/agent_instructions_writing_guide.md (canonical rationale for the @AGENTS.md import pattern, size budgets, anti-patterns)
  • Environment state: shared/references/environment_state_contract.md (detection and bootstrap pattern)

Version: 2.1.0 Last Updated: 2025-01-12

平台分布

Claude Code

29.27%
按下载量换算626

Gemini CLI

23.31%
按下载量换算498

Codex

16.74%
按下载量换算358

OpenCode

10.73%
按下载量换算229

Antigravity

6.76%
按下载量换算145

windsurf

3.48%
按下载量换算74

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

可写文件

该 Skill 可能写入或修改本地文件,使用前需要确认目标目录和修改范围。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。

继续浏览同类 Skills