Skip to content
Tutorials Center
My learning

Topics

Lesson 13 of 38 ยท Reconnaissance

Web Fingerprinting and Technology Discovery

Learn how testers identify web servers, frameworks and CMS products from HTTP headers, cookies, error pages and files, and how defenders reduce what they reveal.

Before you can judge whether a website is vulnerable, you need to know what it is built with. A WordPress site with a vulnerable plugin, an old Apache version, or a framework left in debug mode each call for different tests. Fingerprinting is the craft of identifying software from the small clues it leaves in its responses. Most of it needs nothing more than an ordinary HTTP request, which makes it a light active technique. Run it only against in-scope targets.

Clues in HTTP responses

Response headers

Servers often announce themselves.

TEXT
HTTP/1.1 200 OK
Server: Apache/2.4.41 (Ubuntu)
X-Powered-By: PHP/7.4.3
Set-Cookie: PHPSESSID=3d9f...; path=/; HttpOnly
Content-Type: text/html; charset=UTF-8

This is example output in the real format. It reveals the web server and its version, the language, and (through the cookie name) the session mechanism. A version number lets you look up known vulnerabilities and check whether the software is still supported.

Security headers matter too. The absence of Content-Security-Policy, Strict-Transport-Security, X-Content-Type-Options and a sensible Referrer-Policy is itself a low-severity finding and tells you about the team's hardening habits.

Cookies

Cookie names are strong fingerprints: PHPSESSID (PHP), JSESSIONID (Java), ASP.NET_SessionId (ASP.NET), csrftoken (Django), _rails_session (Ruby on Rails), laravel_session (Laravel). Cookie flags show care: Secure, HttpOnly and SameSite should be present for session cookies.

Page source and files

  • The <meta name="generator"> tag can name the CMS.
  • Paths like /wp-content/, /wp-json/, /sites/default/, /skin/frontend/ indicate WordPress, Drupal or Magento.
  • JavaScript file names and comments reveal frameworks and libraries with versions.
  • robots.txt and sitemap.xml list paths the owner thinks are interesting. /.well-known/security.txt shows how to report issues.

Error pages and behavior

A request for a missing page, an invalid method or malformed input returns an error whose format differs between products. A default error page with a stack trace identifies the framework and may leak file paths. Production sites should show generic errors.

TLS and other layers

TLS certificates show the issuer and names. The set of supported protocol versions and cipher suites shows how the server is configured. A web application firewall (WAF) or CDN may sit in front and add its own headers; knowing it exists affects how you test.

Where fingerprints hide

Headers Server, X-Powered-By.
Cookies Names and flags.
Page source Generator tags, paths.
Error pages Default templates.

One clue is a hint; three agreeing clues are a fact.

1 / 2

Read-only: example tool output

Dedicated tools combine many checks, for example whatweb or browser extensions. The output below is an example of the format only.

Terminal
whatweb -v https://www.example.test
TEXT
WhatWeb report for https://www.example.test
Status    : 200 OK
Title     : Example Corp
IP        : 192.0.2.10
Country   : RESERVED, ZZ

Summary   : Apache[2.4.41], PHP[7.4.3], WordPress[5.8.1], Cookies[PHPSESSID]

Always cross-check. A banner can be edited, hidden or misleading, and version numbers do not prove that a vulnerability is present, because distributions often backport fixes without changing the visible version.

Confidence and verification

Treat fingerprints with graded confidence:

  1. High: several independent clues agree (header, file path and generator tag).
  2. Medium: one specific clue, such as a cookie name.
  3. Low: a guess from page style or default colors.

Then verify before reporting. Check whether the version is current, whether security fixes were backported, and whether the specific vulnerable component is actually exposed.

Stand-in: fingerprint a stored response

The first example inspects a saved response (an invented one) and lists technology clues. This is the real analysis logic applied to sample data, so it needs no network.

Python Extract technology clues from a response

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The second example audits a response's security headers and cookie flags. This is a standard checklist, and a missing header is a small but real finding.

Python Audit security headers

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Defender takeaways

  • Remove or genericize Server and X-Powered-By headers (this is minor hardening, not protection by itself: patching matters more).
  • Use custom error pages and turn off debug modes in production.
  • Set the security headers and cookie flags above.
  • Keep software patched; hiding versions does not fix vulnerabilities.
  • Remove unused default files, sample applications and old admin paths.

Where it leads

Fingerprinting gives you hypotheses to test: which CMS plugin versions to look up, which framework features to probe, which default credentials may exist. It continues in Service and Version Enumeration, and the human side of the attack surface is covered next in Human Factors and Social Engineering.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. A response includes Set-Cookie: JSESSIONID=... What does it suggest?

    JSESSIONID is the default session cookie name of Java servlet containers.
  2. 2. Why should a fingerprinted version number be verified before reporting a vulnerability?

    Distributions often patch security issues in place, so the visible version does not prove exposure.
  3. 3. Which finding is typically low severity but worth reporting?

    A missing hardening header lowers defense in depth but is not directly exploitable alone.
  4. 4. How does hiding the Server header help a defender?

    Hiding banners is minor hardening. It does not remove vulnerabilities.

Summary

  • Fingerprinting identifies servers, languages, frameworks and CMS products from headers, cookies, source, files and error pages.
  • Combine several clues for confidence; verify before reporting, since banners can mislead.
  • Missing security headers and cookie flags are small but real findings.
  • Defenders genericize errors, set headers, patch and remove default content.

Next: Human Factors and Social Engineering