Fun Way to Bypass an Open Redirect Patch
June 20, 2025
4 min read
By Virendra & Monark

Table of Contents

This is a list of all the sections in this post. Click on any of them to jump to that section.

index

In this blog, we will take you through a real-world encounter with a patched Open Redirect vulnerability that turned out to be not-so-patched after all. The target was a modern web application with a typical redirect mechanism using a destination parameter. Such parameter names are actually demanding attention to get exploited as either a Cross-Site Scripting or Open Redirect issue. However, this endpoint appeared to be patched and now comes our game.

May 14, 2025 - Finding Open Redirect

As usual the easiest and foremost simple payloads were utilized to examine the endpoint:

?destination=https://sudarshana.io
?destination=//sudarshana.io
?destination=/\sudarshana.io
?destination=\\sudarshana.io
[...]

All were blocked and responded with 400 Bad Request error. However, the following payload was discovered that bypassed the security in-place:

?destination=///sudarshana.io

image.png

This was reported and accepted by customer.

June 20, 2025 - Bypassing The Patch

The client requested a retest. The previous payload failed and all other common variations failed as well. This was the point were manual fuzzing entered the scene, and replaced wordlist-based payloads.

Instead of relying on tools alone, an Intruder attack was configured with 4 positional placeholders filled with special characters. The idea was to confuse the parser while preserving a valid redirect structure.

image.png

The following characters were used for fuzzing:

$ echo ' ,.,/,\,=,<,>,?,+,&,*,;,:,",{,},|,^,`,#' | tr ',' '\n'
 
.
/
\
=
<
>
?
+
&
*
;
:
"
{
}
|
^
`
#

This 22 letters combination when paired with 4 positions in Cluster Bomb attack configuration generated 234,256 request count. This took a while to run, but results were interesting.

image.png

The results were interesting as following was observed in throwing 302 based redirects:

  • //://x redirected correctly due to the : character

  • //:x//x and //x:x//x were allowed

    image.png

    image.png

  • //x:80//x was blocked, likely due to port validation

    image.png

Based on this observation, it was found that we can not get rid of : and no digits are allowed on right of : . Somehow after trying for a while, the emphasize was put upon how URLs support the username:password@host format. If it would be possible to smuggle sudarshana.io in the host section using this syntax, maybe redirect can be achieved. Tried: //user:pass@sudarshana.io/x and other combinations, but it didn’t work.

image.png

As performed earlier, fuzzing method was utilized and found that @ , /, # were allowed between password and @host keyword.

image.png

This time, the following characters were used for fuzzing. It contains more characters then before mentioned list to not miss on anything:

!
"
#
$
%
&
'
(
)
*
+
,
-
.
/
0
1
2
3
4
5
6
7
8
9
:
;
<
=
>
?
@
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
[
\
]
^
_
`
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
s
t
u
v
w
x
y
z
{
|
}
~
 

image.png

Of all, all other characters throw ERR_INVALID_REDIRECT error, except @ hence indicating trial of double @ characters.

image.png

Surpisingly, this payload worked.

?destination=//user:pass@@sudarshana.io/x

This worked due to browser considering first instance of @ as part of password, while the other @ was assumed to be separator to mark the beginning of the domain name.

Takeaways:

This bypass was not just a win over the patch. This is a reminder of how manual fuzzing, creativity, and understanding of obscure URL parsing rules can outsmart knowledge and wordlist based attacks. Today, AI and tools like ChatGPT operate based on knowledge, but intelligence, curiosity, and experimentation of a human cannot be completely perceived.

Looking to strengthen your security posture? Whether it’s a one-time assessment, continuous monitoring, or a deeper dive into your external attack surface, we are here to help. Our team of OSCP certified experts brings years of hands-on experience and would love to chat about anything in the cybersecurity space. Reach out & let’s make things more secure together.

#OpenRedirect #BugBounty #WebSecurity #URLFuzzing #SecurityResearch #ManualFuzzing #VulnerabilityBypass #OffensiveSecurity #Sudarshana #QuickRead