Token导航 LogoToken导航

Agent Skill · 研究检索

bootstrap(Bootstrap 初始化)

bootstrap 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。

bootstrap 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,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 55,679。

总安装 470周安装 20GitHub Stars 35下载 165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(Bootstrap 初始化)
来源仓库:https://github.com/petekp/claude-code-setup
仓库路径:skills/bootstrap
安装命令:
npx skills add https://github.com/petekp/claude-code-setup --skill bootstrap
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/petekp/claude-code-setup --skill bootstrap

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。

用于查找、检索和筛选相关信息。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。适合根据关键词或任务场景快速定位候选结果。可结合来源仓库和原始 README 核验具体用法。安装前建议确认权限范围和维护状态。

bootstrap 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/bootstrap

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/petekp/claude-code-setup --skill bootstrap;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 怎么安装?

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

bootstrap 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 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 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 用的人多吗?

公开数据是累计安装 470 次,近一周安装 20 次,下载 165 次,GitHub 35 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 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

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。
  • 第三方扫描已通过,常见的高风险行为没有命中。

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

Bootstrap 初始化 · 本页

47035需要联网未明确通过

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

Bootstrap Prompt Generator

Generate a detailed context prompt that enables seamless continuation of work in a fresh Claude Code session.

Purpose

Sessions accumulate valuable context: task understanding, decisions made, patterns discovered, gotchas encountered, and progress achieved. When starting fresh, this knowledge is lost. A bootstrap prompt preserves the essential context needed to continue effectively.

Core Principle: Intelligent Context Selection

Do NOT mechanically dump everything. Analyze what actually matters for continuation:

  • What would confuse a fresh Claude? Include it.
  • What decisions took deliberation? Document the reasoning.
  • What gotchas caused wasted time? Warn about them.
  • What's the critical path forward? Clarify it.

Omit: routine operations, obvious context, resolved dead-ends, standard patterns.

Analysis Process

1. Assess Session Significance

Determine what kind of work occurred:

  • Exploration: Learning codebase, investigating options
  • Implementation: Building features, fixing bugs
  • Debugging: Tracking down issues, testing hypotheses
  • Planning: Designing architecture, making decisions

Each type has different handoff needs.

2. Identify Essential Context

Project State

  • Working directory and project identity
  • Tech stack if non-obvious
  • Branch and git state (use scripts/gather-git-state.sh)

Task Context

  • Original objective (what the user asked for)
  • Current interpretation (what we understood it to mean)
  • Scope decisions (what's in/out)

Progress Assessment

  • What's completed and working
  • What's in progress and current state
  • What's remaining (check todo list)
  • What's blocked and why

Critical Knowledge

  • Architectural decisions and their rationale
  • Patterns established in this session
  • Gotchas and workarounds discovered
  • Files that are central to the work

Next Actions

  • Immediate next step
  • Remaining work outline
  • Known unknowns to investigate

3. Determine Depth

Scale detail to session complexity:

Light session (quick task, simple fix): Brief summary, next step Medium session (feature work, debugging): Key decisions, progress, gotchas Deep session (architecture, complex debug): Full context with reasoning

Output Format

Generate a markdown document structured as a prompt for a fresh session:

# Bootstrap: [Brief Task Description]

> Generated: [timestamp]
> Project: [project path]
> Branch: [branch name]

## Context

[2-4 sentences on what this project/task is about]

## Session Summary

[What happened in the session - decisions, progress, discoveries]

## Current State

[Where things stand right now - what works, what's in progress]

## Key Files

[List of files central to the work with brief descriptions]

## Decisions Made

[Important choices with brief rationale - only if non-obvious]

## Gotchas & Warnings

[Things that caused problems or need careful handling]

## Next Steps

[Prioritized list of what to do next]

## Resume Instructions

[Specific guidance on how to continue - commands to run, files to open, etc.]

Omit sections that aren't relevant. A simple task might only need Context, Current State, and Next Steps.

Execution Steps

  1. Analyze the session - Review conversation, understand what happened
  2. Run git state script - Execute scripts/gather-git-state.sh to capture repository state
  3. Check todo list - Review current todos for progress context
  4. Identify key files - Determine which files are central to the work
  5. Draft bootstrap prompt - Write the document following the format above
  6. Determine output path - Use .claude/handoffs/{project-name}-{YYYYMMDD-HHMMSS}.md
  7. Save the file - Write the bootstrap prompt to the handoffs directory
  8. Copy to clipboard - Execute scripts/copy-to-clipboard.sh {filepath} to copy contents

File Locations

  • Output directory: .claude/handoffs/ (create if doesn't exist)
  • Filename pattern: {project-name}-{YYYYMMDD-HHMMSS}.md
  • Project name: Derive from git remote, directory name, or package.json

Scripts

scripts/gather-git-state.sh

Collects repository state: branch, recent commits, uncommitted changes, modified files. Run this first to include accurate git context in the bootstrap prompt.

scripts/copy-to-clipboard.sh

Copies file contents to system clipboard (macOS pbcopy). Run after saving the bootstrap prompt file.

Quality Checklist

Before finalizing, verify:

  • Fresh Claude could understand the task without prior context
  • Decisions include enough rationale to avoid re-litigating
  • Gotchas are specific enough to be actionable
  • Next steps are concrete and prioritized
  • No unnecessary detail that obscures the important parts

Example Bootstrap Prompts

Light Session Example

# Bootstrap: Fix API rate limiting bug

> Generated: 2025-01-15 14:30
> Project: /Users/dev/acme-api
> Branch: fix/rate-limiter

## Context

Fixing a bug where rate limiting wasn't being applied to the `/search` endpoint.

## Current State

Found the issue - the rate limiter middleware was added after the route registration.
Fix is ready but untested.

## Next Steps

1. Run test suite: `npm test`
2. If passing, commit with message "Fix rate limiter middleware order for /search"

Deep Session Example

See references/deep-session-example.md for a complex multi-day project handoff.

平台分布

Codex

32.33%
按下载量换算53

Claude

32.21%
按下载量换算53

Cursor

19.94%
按下载量换算33

Gemini CLI

9.77%
按下载量换算16

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

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。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