Token导航 LogoToken导航

Agent Skill · 研究检索

bootstrap-project-documentation(引导项目文档)

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。

bootstrap-project-documentation 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 53,999。

总安装 517周安装 22GitHub Stars 7下载 181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
unknown

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:bootstrap-project-documentation(引导项目文档)
来源仓库:https://github.com/nesnilnehc/ai-cortex
仓库路径:skills/bootstrap-project-documentation
安装命令:
npx skills add https://github.com/nesnilnehc/ai-cortex --skill bootstrap-project-documentation
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/nesnilnehc/ai-cortex --skill bootstrap-project-documentation

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

bootstrap-project-documentation 用于辅助文档、README、Markdown 和内容稿件的整理与改写。适合让 Agent 提炼结构、补齐章节、统一术语或检查链接。使用时应保留项目已有事实、命令和路径,避免写成未确认的结论。涉及对外文案时需控制语气,避免过度营销或夸大能力。

适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/bootstrap-project-documentation

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/nesnilnehc/ai-cortex --skill bootstrap-project…;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

    它需要需要联网,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

bootstrap-project-documentation 怎么安装?

在终端执行 npx skills add https://github.com/nesnilnehc/ai-cortex --skill bootstrap-project-documenta…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

bootstrap-project-documentation 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。

bootstrap-project-documentation 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

bootstrap-project-documentation 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

bootstrap-project-documentation 用的人多吗?

公开数据是累计安装 517 次,近一周安装 22 次,下载 181 次,GitHub 7 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。

bootstrap-project-documentation 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。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 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。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
bootstrap-project-documentation

引导项目文档 · 本页

5177需要联网未明确有提醒

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

Skill: Bootstrap Project Documentation

Purpose

Bootstrap or adapt project documentation using the project-documentation-template structure. Two modes: Initialize (empty project—copy templates and fill placeholders) and Adjust (non-empty—use template as target, propose renames/moves/merges, apply in-place after confirmation). Supports repeatable runs; avoids empty dirs and template files unless requested; enforces strict kebab-case naming. Output paths for skill-generated artifacts (ADR, backlog, design-decisions, calibration) follow spec/artifact-contract.md; project-documentation-template provides content and reference. When creating process-management structure, create docs/process-management/project-board/backlog/ and docs/process-management/decisions/ per contract.


Core Objective

Primary Goal: Produce structured lifecycle documentation aligned with the enterprise template through mode-appropriate bootstrapping or adaptation.

Success Criteria (ALL must be met):

  1. ✅ Mode selected correctly: Initialize for empty projects, Adjust for non-empty projects (or user override applied)
  2. ✅ Template structure applied: Documentation follows project-documentation-template conventions for selected scale
  3. ✅ Placeholders filled: All [...] placeholders replaced with project-specific content (or explicitly marked for later)
  4. ✅ Naming conventions enforced: All paths use strict kebab-case; ADR files follow YYYYMMDD-slug-title.md format
  5. ✅ User confirmation obtained: In Adjust mode, changes applied only after user approval of recommendation list

Acceptance Test: Can a developer navigate the documentation structure and find lifecycle documents without consulting the template repository?


Scope Boundaries

This skill handles:

  • Documentation structure bootstrapping (Initialize mode)
  • Documentation structure adaptation (Adjust mode)
  • Placeholder filling and validation
  • Path naming standardization (kebab-case)
  • ADR generation and indexing

This skill does NOT handle:

  • README generation (use generate-standard-readme)
  • AGENTS.md entry creation (use write-agents-entry)
  • Skill-specific documentation (use refine-skill-design)
  • Content writing beyond template placeholders (user provides domain content)

Handoff point: When documentation structure is established and placeholders filled, hand off to content authoring or project-specific documentation workflows.


Use Cases

  • Empty project: Initialize a full docs skeleton by copying template subsets and filling placeholders for small/medium/large projects.
  • Non-empty project: Use the template as target reference; analyze existing docs, propose renames/moves/merges to align structure and naming; apply in-place changes after user confirmation. Do not create empty dirs or template files unless requested. Can be run repeatedly.
  • Shared workflows: Generate Architecture Decision Records (ADR), update version information across docs, validate placeholders and links.
  • Iterative runs: Run the skill repeatedly to progressively organize and refine docs; each run builds on the current state without requiring a full reinit.

When to use: When a project needs structured lifecycle documentation aligned with the enterprise template, or when existing docs should be aligned with that structure.


Behavior

Mode Selection

First determine the execution mode. User override takes precedence; otherwise:

ModeTriggerBehavior
InitializeNo docs/ or docs/ emptyCopy subset, fill placeholders, create VERSION, output docs skeleton
Adjustdocs/ has ≥1 valid documentScan, compare, output recommendation list; apply after user confirmation

Detection rules:

  • No docs/ or docs/ is empty → Initialize
  • docs/ exists and has ≥1 valid .md file → Adjust
  • User explicitly specifies --mode=initialize or --mode=adjust → use that mode

Initialize Mode Steps

  1. Determine project scale: small, medium, or large (from user or context).
  2. Select the document subset from the template by scale:

- Small: Project Overview, Development Guide, User Guide - Medium: + Architecture, Design, Requirements & Planning - Large: + Process Management, Operations Guide, Compliance, Community & Contributing

  1. Fetch templates from TEMPLATE_BASE_URL (see Appendix) or use a local clone.
  2. Copy only selected docs to the project docs/. Create contract-aligned directories per spec/artifact-contract.md when scale permits: docs/design-decisions/ (medium+), docs/calibration/, docs/process-management/project-board/backlog/, docs/process-management/decisions/ (large). Optionally create docs/ARTIFACT_NORMS.md from contract (per spec/artifact-norms-schema.md) for the user to customize. Do not create other empty dirs unless the user explicitly requests them.
  3. Fill placeholders with project metadata (name, dates, tech stack) and prompt for missing critical data.
  4. Create a VERSION file (e.g. 1.0.0) unless the user explicitly requests no new files.
  5. Validate: no unreplaced placeholders, links valid, tables aligned.

Adjust Mode Steps

  1. Use project-documentation-template as the target reference for structure, conventions, and file/directory names.
  2. Scan existing docs under docs/ (structure, placeholders, links, versions).
  3. Compare to the template and identify gaps:

- Structure and path mismatches (nonstandard dir/file names) - Documents alignable to template (rename, move, or merge) - Unfilled placeholders, broken links, version inconsistencies

  1. Produce a recommendation list and present it to the user; ask for confirmation before applying.
  2. After confirmation, apply changes in-place to existing files. Do not create empty directories or add template files unless the user explicitly requests them.
  3. The skill is idempotent: it can be run repeatedly to iteratively organize and refine docs.

Conventions (from llms.txt)

  • Placeholders: [description], [option1/option2], YYYY-MM-DD, [number]
  • Tables: Preserve column alignment; * marks required fields
  • Version: Use SemVer; update version history at document bottom on changes
  • References: Internal [Name](relative/path); external number
  • Dates: YYYY-MM-DD

File and directory naming (strict)

  • Directories: kebab-case only (e.g. project-overview, development-guide, process-management)
  • Files: kebab-case with .md extension (e.g. goals-and-vision.md, versioning-standards.md)
  • ADR files: docs/process-management/decisions/YYYYMMDD-slug.md per spec/artifact-contract.md
  • No spaces, underscores, or PascalCase; lowercase letters, digits, hyphens only.

Input & Output

Input

  • Project metadata: Name, description, tech stack
  • Scale: small | medium | large (optional; infer from context if absent)
  • Mode override (optional): initialize | adjust

Output

  • Initialize: Filled docs under docs/, VERSION, and a short summary of created files
  • Adjust: A recommendation list (markdown or structured), then—after confirmation—the applied changes and summary

Restrictions

Hard Boundaries

  • Replace all placeholders before finalizing documents; do not leave [description] etc. in final output unless explicitly deferred.
  • Do not add or keep broken internal links; verify relative paths.
  • Use consistent dates (YYYY-MM-DD) and SemVer for versions.
  • In Adjust mode, do not apply changes without user confirmation.
  • Do not remove structural elements (sections, tables) from templates without user approval.
  • Do not create empty directories or add template files unless the user explicitly requests them.
  • Use strict file and directory naming (kebab-case, ADR format YYYYMMDD-title.md) when creating or renaming paths.

Skill Boundaries (Avoid Overlap)

Do NOT do these (other skills handle them):

  • README generation: Creating or updating README.md files → Use generate-standard-readme
  • AGENTS.md entry creation: Writing or updating AGENTS.md files → Use write-agents-entry
  • Skill documentation: Creating or refining SKILL.md files → Use refine-skill-design
  • Content authoring: Writing domain-specific content beyond template placeholders → User provides content

When to stop and hand off:

  • User says "structure is ready" or "placeholders filled" → Documentation structure complete, hand off to content authoring
  • User asks "how do I write the content?" → Structure complete, hand off to domain experts or content workflows
  • User asks "can you generate the README?" → Hand off to generate-standard-readme
  • User asks "can you create AGENTS.md?" → Hand off to write-agents-entry

Self-Check

Core Success Criteria (ALL must be met)

  • Mode selected correctly: Initialize for empty projects, Adjust for non-empty projects (or user override applied)
  • Template structure applied: Documentation follows project-documentation-template conventions for selected scale
  • Placeholders filled: All [...] placeholders replaced with project-specific content (or explicitly marked for later)
  • Naming conventions enforced: All paths use strict kebab-case; ADR files follow YYYYMMDD-slug-title.md format
  • User confirmation obtained: In Adjust mode, changes applied only after user approval of recommendation list

Process Quality Checks

  • Links validated: Do internal links resolve and do external links point to valid resources?
  • Version consistency: Is the version consistent across VERSION and affected documents?
  • Tables aligned: Is Markdown table alignment preserved?
  • No extra dirs/files: Were empty dirs and template files avoided unless user requested them?

Acceptance Test

Can a developer navigate the documentation structure and find lifecycle documents without consulting the template repository?

If NO: Documentation structure is incomplete or unclear. Return to mode-specific steps.

If YES: Documentation structure is complete. Proceed to handoff.


Examples

Example 1: Initialize (Empty Project, Small Scale)

Context: New repo my-service, no docs/ directory.

Steps: Agent selects Initialize; scale = small. Copies Project Overview, Development Guide, User Guide from template. Fills project name, dates, placeholder descriptions. Creates VERSION as 1.0.0. Outputs a summary of created files.

Output snippet: docs/project-overview/goals-and-vision.md, docs/development-guide/..., docs/user-guide/..., VERSION. All placeholders filled with project-specific content.

Example 2: Adjust (Non-Empty Project)

Context: Repo has docs/ with project_overview/goals.md (nonstandard paths). Some placeholders unfilled.

Steps: Agent selects Adjust. Uses project-documentation-template as target. Produces recommendation list:

  • Rename project_overview/ → project-overview/, goals.md → goals-and-vision.md (kebab-case, match template)
  • Unfilled placeholders in goals-and-vision.md: [project description], [target date]
  • Broken link: ../architecture/tech-stack.md (path does not exist)

Agent presents the list and asks: "Apply these changes? (Y/n)". User confirms. Agent renames dirs/files, fixes placeholders and links in-place. No new empty dirs or template files created.

Example 3: Common Workflow—Generate ADR

Context: Any project; user needs an Architecture Decision Record.

Steps: Agent fetches docs/process-management/decisions/ADR-TEMPLATE.md from template. Determines next ADR number (e.g. ADR-001). Fills context, options, rationale, consequences with user input. Saves as docs/process-management/decisions/YYYYMMDD-decision-title.md (kebab-case slug) and updates the decision index if present.


Appendix: Output Contract

Initialize Mode

DeliverableRequired
docs/ with selected template files only (no empty dirs)Yes
VERSION fileYes (unless user explicitly requests no new files)
All placeholders replaced (or marked for later)Yes
Version history table at document bottomPer template

Adjust Mode Recommendation List Format

SectionContent
Target referenceproject-documentation-template
Path/naming issuesCurrent path → recommended path (kebab-case, template alignment)
Alignable documentsExisting docs that can be renamed/moved/merged to match template
Unfilled placeholdersFile path + placeholder text
Broken/outdated linksFile path + link
Version issuesConflicts or missing version refs

Template Source

  • TEMPLATE_BASE_URL (canonical): https://raw.githubusercontent.com/nesnilnehc/project-documentation-template/main/
  • Key files: llms.txt, AGENTS.md, README.md, docs/
  • If fetch fails (network unavailable): prompt the user to provide a local clone path or retry later; do not proceed with stale or missing templates.

平台分布

Codex

37.79%
按下载量换算68

Claude

28.67%
按下载量换算52

Cursor

18.31%
按下载量换算33

Gemini CLI

9.32%
按下载量换算17

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills