
Katana Crawl
- 30 installs
- 1.6k repo stars
- Updated July 19, 2026
- wgpsec/aboutsecurity
Helps with ai & agent building tasks during AI-assisted development.
About
katana-crawl is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.
- katana-crawl
- AI & Agent Building
- AI-coding skill
Katana Crawl by the numbers
- 30 all-time installs (skills.sh)
- +5 installs in the week ending Jul 27, 2026 (Skillselion tracking)
- Ranked #9,284 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/wgpsec/aboutsecurity --skill katana-crawlAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 30 |
|---|---|
| repo stars | ★ 1.6k |
| Last updated | July 19, 2026 |
| Repository | wgpsec/aboutsecurity ↗ |
What it does
Helps with ai & agent building tasks during AI-assisted development.
Files
katana Web 爬虫方法论
katana 是 ProjectDiscovery 出品的下一代 Web 爬虫框架。核心优势:JavaScript 解析(发现动态端点)+ Headless 模式(渲染 SPA)+ 管道友好 + 表单自动填充。
项目地址:https://github.com/projectdiscovery/katana
Phase 1: 基本爬取
# 爬取单个目标(默认深度 3)
katana -u http://target.com
# 指定爬取深度
katana -u http://target.com -d 5
# 静默输出(只显示 URL)
katana -u http://target.com -silent
# 输出到文件
katana -u http://target.com -silent -o urls.txtPhase 2: JavaScript 解析(关键能力)
现代 Web 应用大量使用 JS 框架,传统爬虫无法发现 JS 中硬编码的 API 端点:
# 启用 JS 端点解析(推荐始终开启)
katana -u http://target.com -jc
# 启用 jsluice 深度 JS 解析(更全面,内存消耗大)
katana -u http://target.com -jsluice
# Headless 模式(渲染页面后爬取,适合 SPA/React/Vue)
katana -u http://target.com -headless
# Headless + 使用系统 Chrome
katana -u http://target.com -headless -system-chromePhase 3: 批量爬取
# 从文件读取目标
katana -list urls.txt -silent
# 从 stdin 管道
cat urls.txt | katana -silent
# 配合其他工具
subfinder -d target.com -silent | httpx -silent | katana -silent -jcPhase 4: 输出过滤
# 只输出特定扩展名
katana -u http://target.com -em php,jsp,asp -silent
# 排除静态资源
katana -u http://target.com -ef png,jpg,css,gif,svg,woff -silent
# 正则匹配(只输出包含 api 的 URL)
katana -u http://target.com -mr "api|admin|login" -silent
# 正则过滤(排除包含 logout 的 URL)
katana -u http://target.com -fr "logout|static" -silent
# 提取表单元素
katana -u http://target.com -form-extraction -jsonlPhase 5: 范围控制
# 限制在同一域名内(默认行为)
katana -u http://target.com
# 显示外部端点(不爬取,只显示)
katana -u http://target.com -display-out-scope
# 自定义范围(正则)
katana -u http://target.com -crawl-scope "target\.com|api\.target\.com"
# 排除特定路径
katana -u http://target.com -crawl-out-scope "logout|signout"
# 无范围限制(危险,慎用)
katana -u http://target.com -no-scopePhase 6: 高级配置
# 限时爬取(比赛场景重要)
katana -u http://target.com -crawl-duration 60s
# 控制并发和速率
katana -u http://target.com -c 20 -rl 100
# 自定义 Header(带 Cookie 访问需要认证的页面)
katana -u http://target.com -H "Cookie: session=abc123"
# 使用代理
katana -u http://target.com -proxy http://127.0.0.1:8080
# JSONL 输出(包含完整请求信息)
katana -u http://target.com -jsonl -o results.jsonl
# 已知文件扫描(robots.txt, sitemap.xml)
katana -u http://target.com -known-files all -d 3
# 技术检测
katana -u http://target.com -tech-detect
# XHR 请求提取(Headless 模式)
katana -u http://target.com -headless -xhr-extractionPhase 7: 管道集成
# 爬虫 → 漏洞扫描
katana -u http://target.com -jc -silent | nuclei -severity critical,high
# 爬虫 → 筛选 JS 文件 → 敏感信息提取
katana -u http://target.com -jc -silent -em js | while read url; do
curl -s "$url" | grep -iE "api_key|password|secret|token"
done
# 爬虫 → 参数提取 → SQLi 测试
katana -u http://target.com -jc -silent | grep "?" | sort -u > params.txt
# 子域名 → 存活 → 爬虫 → 漏洞扫描(完整链)
subfinder -d target.com -silent | httpx -silent | katana -jc -silent | nuclei爬取策略选择
| 场景 | 策略 | 命令 |
|---|---|---|
| 快速发现(比赛) | 浅爬 + JS | katana -u URL -d 2 -jc -silent |
| 全面爬取 | 深爬 + Headless | katana -u URL -d 5 -headless -jc |
| API 端点发现 | JS 解析 + 过滤 | katana -u URL -jc -em js,json -silent |
| 认证后爬取 | 带 Cookie | katana -u URL -H "Cookie: ..." -jc |
| SPA 应用 | Headless 必须 | katana -u URL -headless -system-chrome |
{
"skill_name": "katana-crawl",
"evals": [
{
"id": 1,
"name": "js-endpoint-discovery",
"prompt": "目标是一个 React SPA 单页应用,传统爬虫爬不到什么内容,怎么用 katana 发现隐藏的 API 端点?",
"expected_output": "katana -u http://target -headless -jc -system-chrome",
"expectations": ["katana", "-headless", "-jc|jsluice|js-crawl"]
},
{
"id": 2,
"name": "crawl-with-auth",
"prompt": "我已经登录了目标网站拿到了 session cookie,怎么用 katana 爬取认证后的页面?",
"expected_output": "katana -u http://target -H 'Cookie: session=xxx' -jc -silent",
"expectations": ["katana", "-H.*Cookie|Header.*Cookie", "session"]
}
]
}
Related skills
AI & Agent Buildingagents