eslint (latest)
Published 21 Jul, 2018
No known vulnerabilities in eslint
Security wise, eslint seems to be a safe package to use.
Over time, new vulnerabilities may be disclosed on eslint and other packages. To easily find, fix and prevent such vulnerabilties, protect your repos with Snyk!
Vulnerable versions of eslint
Fixed in 4.18.2
Regular Expression Denial of Service (ReDoS)
Detailed paths
- Introduced through: eslint@4.18.1
Overview
eslint
is an AST-based pattern checker for JavaScript.
Affected versions of the package are vulnerable to Regular Expression Denial of Service (ReDoS) attacks. This can cause an impact of about 10 seconds matching time for data 100k characters long.
Details
Denial of Service (DoS) describes a family of attacks, all aimed at making a system inaccessible to its original and legitimate users. There are many types of DoS attacks, ranging from trying to clog the network pipes to the system by generating a large volume of traffic from many machines (a Distributed Denial of Service - DDoS - attack) to sending crafted requests that cause a system to crash or take a disproportional amount of time to process.
The Regular expression Denial of Service (ReDoS) is a type of Denial of Service attack. Regular expressions are incredibly powerful, but they aren't very intuitive and can ultimately end up making it easy for attackers to take your site down.
Let’s take the following regular expression as an example:
regex = /A(B|C+)+D/
This regular expression accomplishes the following:
A
The string must start with the letter 'A'(B|C+)+
The string must then follow the letter A with either the letter 'B' or some number of occurrences of the letter 'C' (the+
matches one or more times). The+
at the end of this section states that we can look for one or more matches of this section.D
Finally, we ensure this section of the string ends with a 'D'
The expression would match inputs such as ABBD
, ABCCCCD
, ABCBCCCD
and ACCCCCD
It most cases, it doesn't take very long for a regex engine to find a match:
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCD")'
0.04s user 0.01s system 95% cpu 0.052 total
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCX")'
1.79s user 0.02s system 99% cpu 1.812 total
The entire process of testing it against a 30 characters long string takes around ~52ms. But when given an invalid string, it takes nearly two seconds to complete the test, over ten times as long as it took to test a valid string. The dramatic difference is due to the way regular expressions get evaluated.
Most Regex engines will work very similarly (with minor differences). The engine will match the first possible way to accept the current character and proceed to the next one. If it then fails to match the next one, it will backtrack and see if there was another way to digest the previous character. If it goes too far down the rabbit hole only to find out the string doesn’t match in the end, and if many characters have multiple valid regex paths, the number of backtracking steps can become very large, resulting in what is known as catastrophic backtracking.
Let's look at how our expression runs into this problem, using a shorter string: "ACCCX". While it seems fairly straightforward, there are still four different ways that the engine could match those three C's:
- CCC
- CC+C
- C+CC
- C+C+C.
The engine has to try each of those combinations to see if any of them potentially match against the expression. When you combine that with the other steps the engine must take, we can use RegEx 101 debugger to see the engine has to take a total of 38 steps before it can determine the string doesn't match.
From there, the number of steps the engine must use to validate a string just continues to grow.
String | Number of C's | Number of steps |
---|---|---|
ACCCX | 3 | 38 |
ACCCCX | 4 | 71 |
ACCCCCX | 5 | 136 |
ACCCCCCCCCCCCCCX | 14 | 65,553 |
By the time the string includes 14 C's, the engine has to take over 65,000 steps just to see if the string is valid. These extreme situations can cause them to work very slowly (exponentially related to input size, as shown above), allowing an attacker to exploit this and can cause the service to excessively consume CPU, resulting in a Denial of Service.
Remediation
Upgrade eslint
to version 4.18.2 or higher.
References
Fixed in 4.0.0-beta.0
Command Injection
Detailed paths
- Introduced through: eslint@4.0.0-alpha.2 > shelljs@0.7.8
Overview
shelljs
is a portable Unix shell commands for Node.js.
Affected version of this package are vulnerable to Command Injection. It is possible to invoke commands from shell.exec()
from external sources, allowing an attacker to inject arbitrary commands.
Remediation
There is no fix version for shelljs
.
References
Fixed in 2.1.0
Prototype Pollution
Detailed paths
- Introduced through: eslint@2.0.0-rc.1 > inquirer@0.11.4 > lodash@3.10.1
Overview
lodash is a javaScript utility library delivering modularity, performance & extras.
Affected versions of this package are vulnerable to Prototype Pollution.
The utilities function allow modification of the Object
prototype. If an attacker can control part of the structure passed to this function, they could add or modify an existing property.
PoC by Olivier Arteau (HoLyVieR)
var _= require('lodash');
var malicious_payload = '{"__proto__":{"oops":"It works !"}}';
var a = {};
console.log("Before : " + a.oops);
_.merge({}, JSON.parse(malicious_payload));
console.log("After : " + a.oops);
Remediation
Upgrade lodash
to version 4.17.5 or higher.
References
Fixed in 1.9.0
Regular Expression Denial of Service (DoS)
Detailed paths
- Introduced through: browser-sync@1.8.0 > foxy@7.1.0 > resp-modifier@1.0.2 > minimatch@2.0.10
- Introduced through: browser-sync@1.8.0 > resp-modifier@1.0.2 > minimatch@2.0.10
- Introduced through: browser-sync@1.8.0 > glob-watcher@0.0.7 > gaze@0.5.2 > globule@0.1.0 > minimatch@0.2.14
- Introduced through: browser-sync@1.8.0 > glob-watcher@0.0.7 > gaze@0.5.2 > globule@0.1.0 > glob@3.1.21 > minimatch@0.2.14
- Introduced through: eslint@1.8.0 > minimatch@2.0.10
Overview
minimatch
is a minimalistic matching library used for converting glob expressions into JavaScript RegExp objects.
Affected versions of this package are vulnerable to Regular Expression Denial of Service (ReDoS) attacks.
The Regular expression Denial of Service (ReDoS) is a type of Denial of Service attack. Many Regular Expression implementations may reach edge cases that causes them to work very slowly (exponentially related to input size), allowing an attacker to exploit this and can cause the program to enter these extreme situations by using a specially crafted input and cause the service to excessively consume CPU, resulting in a Denial of Service.
An attacker can provide a long value to the minimatch
function, which nearly matches the pattern being matched. This will cause the regular expression matching to take a long time, all the while occupying the event loop and preventing it from processing other requests and making the server unavailable (a Denial of Service attack).
You can read more about Regular Expression Denial of Service (ReDoS)
on our blog.
Remediation
Upgrade minimatch
to version 3.0.2
or greater.
References
Fixed in 1.4.0
Regular Expression Denial of Service (DoS)
Detailed paths
- Introduced through: bower@1.3.1 > handlebars@1.3.0 > uglify-js@2.3.6
- Introduced through: jade@1.3.1 > transformers@2.1.0 > uglify-js@2.2.5
- Introduced through: jade@1.3.1 > with@3.0.1 > uglify-js@2.4.24
- Introduced through: jade@1.3.1 > constantinople@2.0.1 > uglify-js@2.4.24
- Introduced through: uglify-js@1.3.1
- Introduced through: sinopia@1.3.1 > handlebars@2.0.0 > uglify-js@2.3.6
- Introduced through: eslint@1.3.1 > handlebars@3.0.3 > uglify-js@2.3.6
Overview
The parse()
function in the uglify-js
package prior to version 2.6.0 is vulnerable to regular expression denial of service (ReDoS) attacks when long inputs of certain patterns are processed.
Denial of Service (DoS) describes a family of attacks, all aimed at making a system inaccessible to its original and legitimate users. There are many types of DoS attacks, ranging from trying to clog the network pipes to the system by generating a large volume of traffic from many machines (a Distributed Denial of Service - DDoS - attack) to sending crafted requests that cause a system to crash or take a disproportional amount of time to process.
The Regular expression Denial of Service (ReDoS) is a type of Denial of Service attack. Regular expressions are incredibly powerful, but they aren't very intuitive and can ultimately end up making it easy for attackers to take your site down.
Let’s take the following regular expression as an example:
regex = /A(B|C+)+D/
This regular expression accomplishes the following:
A
The string must start with the letter 'A'(B|C+)+
The string must then follow the letter A with either the letter 'B' or some number of occurrences of the letter 'C' (the+
matches one or more times). The+
at the end of this section states that we can look for one or more matches of this section.D
Finally, we ensure this section of the string ends with a 'D'
The expression would match inputs such as ABBD
, ABCCCCD
, ABCBCCCD
and ACCCCCD
It most cases, it doesn't take very long for a regex engine to find a match:
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCD")'
0.04s user 0.01s system 95% cpu 0.052 total
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCX")'
1.79s user 0.02s system 99% cpu 1.812 total
The entire process of testing it against a 30 characters long string takes around ~52ms. But when given an invalid string, it takes nearly two seconds to complete the test, over ten times as long as it took to test a valid string. The dramatic difference is due to the way regular expressions get evaluated.
Most Regex engines will work very similarly (with minor differences). The engine will match the first possible way to accept the current character and proceed to the next one. If it then fails to match the next one, it will backtrack and see if there was another way to digest the previous character. If it goes too far down the rabbit hole only to find out the string doesn’t match in the end, and if many characters have multiple valid regex paths, the number of backtracking steps can become very large, resulting in what is known as catastrophic backtracking.
Let's look at how our expression runs into this problem, using a shorter string: "ACCCX". While it seems fairly straightforward, there are still four different ways that the engine could match those three C's:
- CCC
- CC+C
- C+CC
- C+C+C.
The engine has to try each of those combinations to see if any of them potentially match against the expression. When you combine that with the other steps the engine must take, we can use RegEx 101 debugger to see the engine has to take a total of 38 steps before it can determine the string doesn't match.
From there, the number of steps the engine must use to validate a string just continues to grow.
String | Number of C's | Number of steps |
---|---|---|
ACCCX | 3 | 38 |
ACCCCX | 4 | 71 |
ACCCCCX | 5 | 136 |
ACCCCCCCCCCCCCCX | 14 | 65,553 |
By the time the string includes 14 C's, the engine has to take over 65,000 steps just to see if the string is valid. These extreme situations can cause them to work very slowly (exponentially related to input size, as shown above), allowing an attacker to exploit this and can cause the service to excessively consume CPU, resulting in a Denial of Service.
Remediation
Upgrade to version 2.6.0
or greater.
If a direct dependency update is not possible, use snyk wizard
to patch this vulnerability.
References
Cross-site Scripting (XSS)
Detailed paths
- Introduced through: bower@1.3.1 > handlebars@1.3.0
- Introduced through: sinopia@1.3.1 > handlebars@2.0.0
- Introduced through: eslint@1.3.1 > handlebars@3.0.3
Overview
handlebars provides the power necessary to let you build semantic templates.
When using attributes without quotes in a handlebars template, an attacker can manipulate the input to introduce additional attributes, potentially executing code. This may lead to a Cross-site Scripting (XSS) vulnerability, assuming an attacker can influence the value entered into the template. If the handlebars template is used to render user-generated content, this vulnerability may escalate to a persistent XSS vulnerability.
Details
Cross-Site Scripting (XSS) attacks occur when an attacker tricks a user’s browser to execute malicious JavaScript code in the context of a victim’s domain. Such scripts can steal the user’s session cookies for the domain, scrape or modify its content, and perform or modify actions on the user’s behalf, actions typically blocked by the browser’s Same Origin Policy.
These attacks are possible by escaping the context of the web application and injecting malicious scripts in an otherwise trusted website. These scripts can introduce additional attributes (say, a "new" option in a dropdown list or a new link to a malicious site) and can potentially execute code on the clients side, unbeknown to the victim. This occurs when characters like \< > \" \' are not escaped properly.
There are a few types of XSS:
- Persistent XSS is an attack in which the malicious code persists into the web app’s database.
- Reflected XSS is an which the website echoes back a portion of the request. The attacker needs to trick the user into clicking a malicious link (for instance through a phishing email or malicious JS on another page), which triggers the XSS attack.
- DOM-based XSS is an that occurs purely in the browser when client-side JavaScript echoes back a portion of the URL onto the page. DOM-Based XSS is notoriously hard to detect, as the server never gets a chance to see the attack taking place.
Example:
Assume handlebars was used to display user comments and avatar, using the following template:
<img src={{avatarUrl}}><pre>{{comment}}</pre>
If an attacker spoofed their avatar URL and provided the following value:
http://evil.org/avatar.png onload=alert(document.cookie)
The resulting HTML would be the following, triggering the script once the image loads:
<img src=http://evil.org/avatar.png onload=alert(document.cookie)><pre>Gotcha!</pre>
References
Improper minification of non-boolean comparisons
Detailed paths
- Introduced through: bower@1.3.1 > handlebars@1.3.0 > uglify-js@2.3.6
- Introduced through: jade@1.3.1 > transformers@2.1.0 > uglify-js@2.2.5
- Introduced through: uglify-js@1.3.1
- Introduced through: sinopia@1.3.1 > handlebars@2.0.0 > uglify-js@2.3.6
- Introduced through: eslint@1.3.1 > handlebars@3.0.3 > uglify-js@2.3.6
Overview
uglify-js
is a JavaScript parser, minifier, compressor and beautifier toolkit.
Tom MacWright discovered that UglifyJS versions 2.4.23 and earlier are affected by a vulnerability which allows a specially crafted Javascript file to have altered functionality after minification. This bug was demonstrated by Yan to allow potentially malicious code to be hidden within secure code, activated by minification.
Details
In Boolean algebra, DeMorgan's laws describe the relationships between conjunctions (&&
), disjunctions (||
) and negations (!
).
In Javascript form, they state that:
!(a && b) === (!a) || (!b)
!(a || b) === (!a) && (!b)
The law does not hold true when one of the values is not a boolean however.
Vulnerable versions of UglifyJS do not account for this restriction, and erroneously apply the laws to a statement if it can be reduced in length by it.
Consider this authentication function:
function isTokenValid(user) {
var timeLeft =
!!config && // config object exists
!!user.token && // user object has a token
!user.token.invalidated && // token is not explicitly invalidated
!config.uninitialized && // config is initialized
!config.ignoreTimestamps && // don't ignore timestamps
getTimeLeft(user.token.expiry); // > 0 if expiration is in the future
// The token must not be expired
return timeLeft > 0;
}
function getTimeLeft(expiry) {
return expiry - getSystemTime();
}
When minified with a vulnerable version of UglifyJS, it will produce the following insecure output, where a token will never expire:
( Formatted for readability )
function isTokenValid(user) {
var timeLeft = !( // negation
!config // config object does not exist
|| !user.token // user object does not have a token
|| user.token.invalidated // token is explicitly invalidated
|| config.uninitialized // config isn't initialized
|| config.ignoreTimestamps // ignore timestamps
|| !getTimeLeft(user.token.expiry) // > 0 if expiration is in the future
);
return timeLeft > 0
}
function getTimeLeft(expiry) {
return expiry - getSystemTime()
}
Remediation
Upgrade UglifyJS to version 2.4.24
or higher.
References
Fixed in 1.1.0
Arbitrary Code Injection
Detailed paths
- Introduced through: eslint@1.0.0-rc-3 > mocha@2.5.3 > growl@1.9.2
Overview
growl
is a package adding Growl support for Nodejs.
Affected versions of the package are vulnerable to Arbitrary Code Injection due to unsafe use of the eval()
function. Node.js provides the eval()
function by default, and is used to translate strings into Javascript code. An attacker can craft a malicious payload to inject arbitrary commands.
Remediation
Upgrade growl
to version 1.10.0 or higher.
References
Regular Expression Denial of Service (ReDoS)
Detailed paths
- Introduced through: eslint@1.0.0-rc-3 > mocha@2.5.3 > debug@2.2.0
Overview
debug
is a JavaScript debugging utility modelled after Node.js core's debugging technique..
debug
uses printf-style formatting. Affected versions of this package are vulnerable to Regular expression Denial of Service (ReDoS) attacks via the the %o
formatter (Pretty-print an Object all on a single line). It used a regular expression (/\s*\n\s*/g
) in order to strip whitespaces and replace newlines with spaces, in order to join the data into a single line. This can cause a very low impact of about 2 seconds matching time for data 50k characters long.
Details
Denial of Service (DoS) describes a family of attacks, all aimed at making a system inaccessible to its original and legitimate users. There are many types of DoS attacks, ranging from trying to clog the network pipes to the system by generating a large volume of traffic from many machines (a Distributed Denial of Service - DDoS - attack) to sending crafted requests that cause a system to crash or take a disproportional amount of time to process.
The Regular expression Denial of Service (ReDoS) is a type of Denial of Service attack. Regular expressions are incredibly powerful, but they aren't very intuitive and can ultimately end up making it easy for attackers to take your site down.
Let’s take the following regular expression as an example:
regex = /A(B|C+)+D/
This regular expression accomplishes the following:
A
The string must start with the letter 'A'(B|C+)+
The string must then follow the letter A with either the letter 'B' or some number of occurrences of the letter 'C' (the+
matches one or more times). The+
at the end of this section states that we can look for one or more matches of this section.D
Finally, we ensure this section of the string ends with a 'D'
The expression would match inputs such as ABBD
, ABCCCCD
, ABCBCCCD
and ACCCCCD
It most cases, it doesn't take very long for a regex engine to find a match:
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCD")'
0.04s user 0.01s system 95% cpu 0.052 total
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCX")'
1.79s user 0.02s system 99% cpu 1.812 total
The entire process of testing it against a 30 characters long string takes around ~52ms. But when given an invalid string, it takes nearly two seconds to complete the test, over ten times as long as it took to test a valid string. The dramatic difference is due to the way regular expressions get evaluated.
Most Regex engines will work very similarly (with minor differences). The engine will match the first possible way to accept the current character and proceed to the next one. If it then fails to match the next one, it will backtrack and see if there was another way to digest the previous character. If it goes too far down the rabbit hole only to find out the string doesn’t match in the end, and if many characters have multiple valid regex paths, the number of backtracking steps can become very large, resulting in what is known as catastrophic backtracking.
Let's look at how our expression runs into this problem, using a shorter string: "ACCCX". While it seems fairly straightforward, there are still four different ways that the engine could match those three C's:
- CCC
- CC+C
- C+CC
- C+C+C.
The engine has to try each of those combinations to see if any of them potentially match against the expression. When you combine that with the other steps the engine must take, we can use RegEx 101 debugger to see the engine has to take a total of 38 steps before it can determine the string doesn't match.
From there, the number of steps the engine must use to validate a string just continues to grow.
String | Number of C's | Number of steps |
---|---|---|
ACCCX | 3 | 38 |
ACCCCX | 4 | 71 |
ACCCCCX | 5 | 136 |
ACCCCCCCCCCCCCCX | 14 | 65,553 |
By the time the string includes 14 C's, the engine has to take over 65,000 steps just to see if the string is valid. These extreme situations can cause them to work very slowly (exponentially related to input size, as shown above), allowing an attacker to exploit this and can cause the service to excessively consume CPU, resulting in a Denial of Service.
Remediation
Upgrade debug
to version 2.6.9, 3.1.0 or higher.
References
Regular Expression Denial of Service (ReDoS)
Detailed paths
- Introduced through: eslint@1.0.0-rc-3 > mocha@2.5.3 > debug@2.2.0 > ms@0.7.1
Overview
ms
is a tiny millisecond conversion utility.
Affected versions of this package are vulnerable to Regular Expression Denial of Service (ReDoS) due to an incomplete fix for previously reported vulnerability npm:ms:20151024. The fix limited the length of accepted input string to 10,000 characters, and turned to be insufficient making it possible to block the event loop for 0.3 seconds (on a typical laptop) with a specially crafted string passed to ms()
function.
Proof of concept
ms = require('ms');
ms('1'.repeat(9998) + 'Q') // Takes about ~0.3s
Note: Snyk's patch for this vulnerability limits input length to 100 characters. This new limit was deemed to be a breaking change by the author. Based on user feedback, we believe the risk of breakage is very low, while the value to your security is much greater, and therefore opted to still capture this change in a patch for earlier versions as well. Whenever patching security issues, we always suggest to run tests on your code to validate that nothing has been broken.
For more information on Regular Expression Denial of Service (ReDoS)
attacks, go to our blog.
Disclosure Timeline
- Feb 9th, 2017 - Reported the issue to package owner.
- Feb 11th, 2017 - Issue acknowledged by package owner.
- April 12th, 2017 - Fix PR opened by Snyk Security Team.
- May 15th, 2017 - Vulnerability published.
- May 16th, 2017 - Issue fixed and version
2.0.0
released. - May 21th, 2017 - Patches released for versions
>=0.7.1, <=1.0.0
.
Details
Denial of Service (DoS) describes a family of attacks, all aimed at making a system inaccessible to its original and legitimate users. There are many types of DoS attacks, ranging from trying to clog the network pipes to the system by generating a large volume of traffic from many machines (a Distributed Denial of Service - DDoS - attack) to sending crafted requests that cause a system to crash or take a disproportional amount of time to process.
The Regular expression Denial of Service (ReDoS) is a type of Denial of Service attack. Regular expressions are incredibly powerful, but they aren't very intuitive and can ultimately end up making it easy for attackers to take your site down.
Let’s take the following regular expression as an example:
regex = /A(B|C+)+D/
This regular expression accomplishes the following:
A
The string must start with the letter 'A'(B|C+)+
The string must then follow the letter A with either the letter 'B' or some number of occurrences of the letter 'C' (the+
matches one or more times). The+
at the end of this section states that we can look for one or more matches of this section.D
Finally, we ensure this section of the string ends with a 'D'
The expression would match inputs such as ABBD
, ABCCCCD
, ABCBCCCD
and ACCCCCD
It most cases, it doesn't take very long for a regex engine to find a match:
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCD")'
0.04s user 0.01s system 95% cpu 0.052 total
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCX")'
1.79s user 0.02s system 99% cpu 1.812 total
The entire process of testing it against a 30 characters long string takes around ~52ms. But when given an invalid string, it takes nearly two seconds to complete the test, over ten times as long as it took to test a valid string. The dramatic difference is due to the way regular expressions get evaluated.
Most Regex engines will work very similarly (with minor differences). The engine will match the first possible way to accept the current character and proceed to the next one. If it then fails to match the next one, it will backtrack and see if there was another way to digest the previous character. If it goes too far down the rabbit hole only to find out the string doesn’t match in the end, and if many characters have multiple valid regex paths, the number of backtracking steps can become very large, resulting in what is known as catastrophic backtracking.
Let's look at how our expression runs into this problem, using a shorter string: "ACCCX". While it seems fairly straightforward, there are still four different ways that the engine could match those three C's:
- CCC
- CC+C
- C+CC
- C+C+C.
The engine has to try each of those combinations to see if any of them potentially match against the expression. When you combine that with the other steps the engine must take, we can use RegEx 101 debugger to see the engine has to take a total of 38 steps before it can determine the string doesn't match.
From there, the number of steps the engine must use to validate a string just continues to grow.
String | Number of C's | Number of steps |
---|---|---|
ACCCX | 3 | 38 |
ACCCCX | 4 | 71 |
ACCCCCX | 5 | 136 |
ACCCCCCCCCCCCCCX | 14 | 65,553 |
By the time the string includes 14 C's, the engine has to take over 65,000 steps just to see if the string is valid. These extreme situations can cause them to work very slowly (exponentially related to input size, as shown above), allowing an attacker to exploit this and can cause the service to excessively consume CPU, resulting in a Denial of Service.
Remediation
Upgrade ms
to version 2.0.0 or higher.
References
Fixed in 0.8.2
Regular Expression Denial of Service (ReDoS)
Detailed paths
- Introduced through: eslint@0.8.1 > js-yaml@3.0.2 > argparse@0.1.16 > underscore.string@2.4.0
Overview
underscore.string is a String manipulation helpers for javascript.
Affected versions of this package are vulnerable to Regular Expression Denial of Service (ReDoS). It parses dates using regex strings, which may cause a slowdown of 2 seconds per 50k characters.
Details
Denial of Service (DoS) describes a family of attacks, all aimed at making a system inaccessible to its original and legitimate users. There are many types of DoS attacks, ranging from trying to clog the network pipes to the system by generating a large volume of traffic from many machines (a Distributed Denial of Service - DDoS - attack) to sending crafted requests that cause a system to crash or take a disproportional amount of time to process.
The Regular expression Denial of Service (ReDoS) is a type of Denial of Service attack. Regular expressions are incredibly powerful, but they aren't very intuitive and can ultimately end up making it easy for attackers to take your site down.
Let’s take the following regular expression as an example:
regex = /A(B|C+)+D/
This regular expression accomplishes the following:
A
The string must start with the letter 'A'(B|C+)+
The string must then follow the letter A with either the letter 'B' or some number of occurrences of the letter 'C' (the+
matches one or more times). The+
at the end of this section states that we can look for one or more matches of this section.D
Finally, we ensure this section of the string ends with a 'D'
The expression would match inputs such as ABBD
, ABCCCCD
, ABCBCCCD
and ACCCCCD
It most cases, it doesn't take very long for a regex engine to find a match:
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCD")'
0.04s user 0.01s system 95% cpu 0.052 total
$ time node -e '/A(B|C+)+D/.test("ACCCCCCCCCCCCCCCCCCCCCCCCCCCCX")'
1.79s user 0.02s system 99% cpu 1.812 total
The entire process of testing it against a 30 characters long string takes around ~52ms. But when given an invalid string, it takes nearly two seconds to complete the test, over ten times as long as it took to test a valid string. The dramatic difference is due to the way regular expressions get evaluated.
Most Regex engines will work very similarly (with minor differences). The engine will match the first possible way to accept the current character and proceed to the next one. If it then fails to match the next one, it will backtrack and see if there was another way to digest the previous character. If it goes too far down the rabbit hole only to find out the string doesn’t match in the end, and if many characters have multiple valid regex paths, the number of backtracking steps can become very large, resulting in what is known as catastrophic backtracking.
Let's look at how our expression runs into this problem, using a shorter string: "ACCCX". While it seems fairly straightforward, there are still four different ways that the engine could match those three C's:
- CCC
- CC+C
- C+CC
- C+C+C.
The engine has to try each of those combinations to see if any of them potentially match against the expression. When you combine that with the other steps the engine must take, we can use RegEx 101 debugger to see the engine has to take a total of 38 steps before it can determine the string doesn't match.
From there, the number of steps the engine must use to validate a string just continues to grow.
String | Number of C's | Number of steps |
---|---|---|
ACCCX | 3 | 38 |
ACCCCX | 4 | 71 |
ACCCCCX | 5 | 136 |
ACCCCCCCCCCCCCCX | 14 | 65,553 |
By the time the string includes 14 C's, the engine has to take over 65,000 steps just to see if the string is valid. These extreme situations can cause them to work very slowly (exponentially related to input size, as shown above), allowing an attacker to exploit this and can cause the service to excessively consume CPU, resulting in a Denial of Service.
Remediation
There is no fix version for underscore.string
.