RUN $ python3 -m labs.web_exposure.check --demo SOURCE ↗

Lab 9: Web Exposure

Why it matters: Forgotten admin pages, backup folders, .git/ directories, and debug configs are among the most common ways web servers leak information. An exposed .git/ can hand over your entire source history. Checking your own servers for them is basic hygiene.

What you’ll practice

  • Recognizing commonly exposed web paths
  • Interpreting HTTP status codes with care
  • Building safety guardrails into security tooling

Run it

From the root of the CyberSHIELD Labs repository:

python3 -m labs.web_exposure.check --demo

--demo starts a temporary practice server on 127.0.0.1 (on a random free port) with synthetic files, sends one HEAD request for each of 18 common paths, prints the results, and shuts the server down. To check a server you run yourself, pass its URL instead: python3 -m labs.web_exposure.check http://127.0.0.1:8000.

Example output:

$ python3 -m labs.web_exposure.check --demo
Checking http://127.0.0.1:55823 ...

[+] Access controlled (might exist) (403): http://127.0.0.1:55823/admin/
[+] Found (200): http://127.0.0.1:55823/.git/
[+] Found (200): http://127.0.0.1:55823/robots.txt
[+] Found (200): http://127.0.0.1:55823/config.json
[+] Found (200): http://127.0.0.1:55823/backup/

What the checks look for

CheckTriggers when
FoundThe path returned 200.
Access controlledThe path returned 401 or 403. That’s only a response code; it doesn’t prove the resource exists.
Redirect not followedThe path returned a 3xx redirect, which the checker reports but never follows.

Try this

  1. Run --demo, then start python3 -m http.server 8000 --bind 127.0.0.1 in an empty folder and check that. Compare the results.
  2. Try https://example.com. The checker refuses it. Read normalize_base_url to see how the loopback-only guard works.
  3. Some servers return 200 for every path. How would you tell a real hit from a catch-all page?

What this lab doesn’t do

The checker accepts only localhost or loopback IP addresses and refuses credentials, query strings, and fragments in the URL. It ignores HTTP(S)_PROXY settings so requests never leave your machine, doesn’t follow redirects, and uses a short timeout. It is not a vulnerability scanner or a substitute for an authorized assessment. Don’t modify the local-only safeguard.

Takeaway

Keep deployable artifacts out of web roots, block version-control and backup paths at the server, and assess only systems you own or are explicitly authorized to test.

Source: labs/web_exposure on GitHub.