security-talent/bugskill-ai

web-fingerprinting

Identify web server type/version, framework, and application entry points via banner grabbing, HTTP header analysis (Server, X-Powered-By, X-Generator), cookie names (CAKEPHP, laravelsession, wp-settings), HTML meta generators, robots.txt, source map files…

소스 보기

사용 방법

  1. 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
  2. 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
  3. 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/security-talent/bugskill-ai --skill web-fingerprinting

시작 전 확인

  • 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
  • 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서

원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.

Web Fingerprinting and Information Gathering

What Is Broken and Why

Web applications leak technology stack details through HTTP response headers, HTML comments, cookie names, error messages, and static file paths. Search engines and archives may index sensitive configuration files, credentials, and internal network diagrams. Collectively this intelligence narrows an attacker's target surface dramatically before any active exploitation begins, enabling precise selection of known CVEs for the identified stack.

Key Signals

  • Server: header reveals product and version (e.g., Apache/2.4.41, nginx/1.17.3)
  • X-Powered-By: or X-Generator: headers expose backend language/framework
  • Framework-specific cookies: CAKEPHP, laravel_session, wp-settings, fe_typo_user
  • HTML <meta name="generator"> tags (WordPress, Joomla, Drupal)
  • Accessible paths: /wp-admin/, /wp-content/, /wp-includes/
  • HTML comments embedding SQL queries, credentials, or internal IPs
  • JS files with hardcoded API keys, DB connection strings, AWS credentials
  • .map source map files exposing full source trees and internal file paths
  • robots.txt Disallow entries revealing hidden paths
  • Error responses leaking stack traces, framework names, file paths
  • __VIEWSTATE in forms = ASP.NET; <!-- ZK = ZK framework; .cfm = ColdFusion

Methodology

  1. Passive recon: Run Google dorks against the target domain using site:, inurl:,

filetype:, intitle:, intext: operators. Check Internet Archive Wayback Machine and Shodan.

  1. Banner grabbing: Send raw HTTP/HTTPS requests and record Server:, X-Powered-By:,

ETag, header ordering, and error page footers.

  1. Malformed request probing: Send invalid HTTP version strings to trigger version-revealing

error pages.

  1. Cookie analysis: Capture Set-Cookie headers across all application sections; match names

against known framework signatures.

  1. HTML/JS source review: Spider the application, download all .js files, search for

credentials patterns, internal hostnames, API endpoints.

  1. Source map detection: Append .map to discovered JS filenames; if 200 OK, parse sources

array for internal paths.

  1. Entry point mapping: Use intercepting proxy to record all GET/POST parameters, hidden

fields, custom headers, and WebSocket frames.

  1. Dirbusting: Request framework-specific paths identified in step 4; compare HTTP response

codes to confirm technology.

Payloads & Tools

bash
# Banner grabbing over HTTP
curl -I http://TARGET/

# Banner grabbing over HTTPS
openssl s_client -connect TARGET:443 -quiet | head -20

# Malformed request to trigger error page
printf 'GET / SANTA CLAUS/1.1\r\nHost: TARGET\r\n\r\n' | nc TARGET 80

# Nikto scan
nikto -h http://TARGET/

# WhatWeb fingerprint
whatweb -a 3 http://TARGET/

# Nmap service/version detection
nmap -sV -p 80,443,8080,8443 TARGET

# Find source maps
curl -s http://TARGET/static/app.js.map | python3 -m json.tool | grep sources

# Google dork examples (replace TARGET with actual domain)
# site:TARGET filetype:env
# site:TARGET inurl:admin
# site:TARGET intitle:"index of"
# site:TARGET filetype:sql

# Check robots.txt
curl -s http://TARGET/robots.txt

# Retrieve archived URLs
waybackurls TARGET | grep -E '\.(php|asp|aspx|jsp|config|bak|sql|env)'

Bypass Techniques

  • Servers with header-suppression (mod_security, mod_headers ServerTokens Prod) still leak via

header ordering, ETag format, and error page HTML structure.

  • WAFs hiding Server: headers may still reveal technology in cookies or X-AspNet-Version.
  • Source maps may be accessible even when JS is minified/obfuscated.
  • Wayback Machine and Common Crawl may preserve old versions of pages that no longer exist live.

Exploitation Scenarios

Scenario 1 — Version to CVE Setup: Target running Apache/2.4.41 revealed via Server: header. Trigger: Cross-reference with CVE database; identify known RCE for that minor version. Impact: Targeted exploitation without broad scanning.

Scenario 2 — Hardcoded AWS Key Setup: JS source map exposes src/config/aws.js with accessKeyId and secretAcccessKey. Trigger: Extract keys, configure AWS CLI with stolen credentials. Impact: Full cloud account takeover, data exfiltration, infrastructure control.

Scenario 3 — Hidden Admin Panel Setup: robots.txt lists Disallow: /admin-legacy/. Framework cookies confirm WordPress. Trigger: Navigate to /admin-legacy/; attempt default or leaked credentials. Impact: Unauthenticated or low-effort admin access.

False Positives

  • Generic Server: Apache without version may indicate a well-hardened install, not a vulnerable one.
  • X-Powered-By can be spoofed to mislead fingerprinting; corroborate with multiple signals.
  • Source map sources paths may be Docker container paths with no direct filesystem relevance.
  • robots.txt Disallow entries may point to already-deleted or protected paths.

Fix Patterns

  • Set ServerTokens Prod and ServerSignature Off in Apache; server_tokens off in Nginx.
  • Remove X-Powered-By with Header unset X-Powered-By or framework-level config.
  • Strip or randomize ETag format to avoid version inference.
  • Never ship .map files to production; configure build pipeline to exclude them.
  • Audit all JS bundles for secrets before deployment; use environment variables at runtime.
  • Set framework-neutral cookie names.
  • Implement Cache-Control: no-store on error pages to prevent caching of stack traces.

Related Skills

Fingerprinting is the entry point that directs all further testing. Once the stack is identified, move to [[auth-bypass]] if an admin panel is exposed, or [[sql-injection]] if the framework version has a known SQLi vulnerability. If a CMS login is found, [[default-credentials]] is the immediate next step. Source maps and hardcoded secrets discovered here can feed directly into [[business-logic-flaws]] by revealing undocumented API endpoints.

같은 저장소의 Skills

더 많은 Skills

모든 Skills