Files
muzi_skills/crawler-reverse/README.md
2026-03-10 08:58:27 +08:00

4.6 KiB
Raw Permalink Blame History

crawler-reverse

中文 | English

一个适用于 OpenClaw 风格技能仓库 的可复用技能包,用于在合法授权前提下进行网页抓包分析、前端 JS 混淆排查、请求签名定位、反爬链路梳理,以及浏览器辅助逆向分析。

这个技能能做什么

当你需要下面这些能力时,可以使用 crawler-reverse

  • 分析页面请求链路
  • 查找 signtokentimestampnonce 或自定义 Header 的生成位置
  • 比较浏览器请求与脚本请求差异
  • 排查与请求相关的前端 JS 混淆逻辑
  • 分析 Cookie / localStorage / sessionStorage / Header 依赖
  • 复现一个已观察到的请求流程,并输出最小验证脚本

安全边界

这个技能仅用于合法授权、正当测试、自有系统调试、教学演示或明确获准的分析场景

不应用于:

  • 未授权访问
  • 绕过登录、权限、付费墙、验证码或限流
  • 撞库、账号滥用
  • 未经授权的大规模采集
  • 为攻击性滥用提供规避安全控制方案

如果授权范围不明确,应该先确认再继续。

仓库内容

  • SKILL.md — 技能主说明
  • skill.json — 基础元数据,可用于索引/注册
  • examples/example.md — 示例提示词与使用方式
  • LICENSE — MIT 许可证

推荐用法

典型分析流程:

  1. 在浏览器中复现用户操作
  2. 观察 XHR / fetch / websocket / document 请求
  3. 识别动态参数
  4. 追踪这些参数的生成位置
  5. 对比浏览器请求与脚本请求
  6. 产出最小验证脚本

推荐配套工具

这个技能适合与以下工具配合使用:

  • 浏览器自动化 / 浏览器检查工具
  • 本地文件读取工具
  • shell / grep / ripgrep
  • 小型 Python / JavaScript 验证脚本

安装方式

将该目录复制到你的 OpenClaw 兼容 skills 目录,或根据你的 OpenClaw 配置将该 GitHub 仓库作为自定义技能来源引入。

技能摘要

  • 名称: crawler-reverse
  • 分类: web-analysis / reverse-engineering / debugging
  • 主要输出: 请求链路分析、参数来源说明、安全复现步骤

说明

这个仓库目前采用通用 GitHub skill 仓库布局生成,后续如果需要适配某个 OpenClaw 技能注册中心或特定格式,可以再进一步调整。


English

A reusable OpenClaw-style skill package for authorized web traffic analysis, JS deobfuscation support, request-signature tracing, anti-bot workflow inspection, and browser-assisted reverse engineering.

What this skill is for

Use crawler-reverse when you need to:

  • inspect a page's request chain
  • locate where sign, token, timestamp, nonce, or custom headers are generated
  • compare browser requests with script requests
  • analyze obfuscated frontend JS related to requests
  • understand cookie / localStorage / sessionStorage / header dependencies
  • reproduce an observed request flow with a minimal script

Safety boundary

This skill is intended only for authorized, defensive, educational, self-owned, or explicitly permitted analysis.

It must not be used for:

  • unauthorized access
  • bypassing authentication, paywalls, permissions, captchas, or rate limits
  • credential stuffing / account abuse
  • large-scale scraping in violation of authorization
  • evasion of security controls for abusive purposes

If authorization is unclear, ask first.

Package contents

  • SKILL.md — full skill instructions
  • skill.json — basic metadata for registry/indexing
  • examples/example.md — example invocation patterns
  • LICENSE — MIT

Suggested usage

Typical workflow:

  1. Reproduce the user action in a browser
  2. Observe XHR / fetch / websocket / document requests
  3. Identify dynamic parameters
  4. Trace where they are generated
  5. Compare browser and script requests
  6. Produce a minimal validation script

This skill is designed to pair well with tools such as:

  • browser automation / browser inspection tools
  • local file readers
  • shell / grep / ripgrep
  • small Python or JavaScript validation scripts

Install

Copy this folder into your OpenClaw-compatible skills directory, or add it as a GitHub-hosted custom skill source depending on your OpenClaw setup.

Skill summary

  • Name: crawler-reverse
  • Category: web-analysis / reverse-engineering / debugging
  • Primary output: request-chain analysis, parameter-origin explanation, safe reproduction steps

Publishing note

This package was generated in a generic GitHub skill-repo layout so it can be adapted to a specific OpenClaw registry format later if needed.