This utility is a simple wrapper around the js-xss library that will configure js-xss
to sanitize a value according to the ArcGIS Supported HTML spec. It also
includes a few additional helper methods to validate strings and
prevent XSS attacks.
WARNING: This utility will sanitize and escape a string according to the
ArcGIS Online supported HTML specification. The sanitized string can be inserted
into the DOM
via a method like element.innerHTML = sanitizedHtml
. However,
you should never insert the sanitized string in the following scenarios:
<script>
...NEVER PUT UNTRUSTED DATA HERE...
</script>
Directly in a script
<!--...NEVER PUT UNTRUSTED DATA HERE...-->
Inside an HTML comment
<div ...NEVER PUT UNTRUSTED DATA HERE...="test" />
In an attribute name <NEVER PUT UNTRUSTED DATA HERE... href="/test" /> In a tag
name
<style>
...NEVER PUT UNTRUSTED DATA HERE...
</style>
Directly in CSS
You should also not extend the sanitizer to whitelist the following
tags: script, style, noscript
.
In order to prevent additional attacks, this library should only be used if the following requirements are met: UTF-8 character set, JavaScript environment (NodeJS or Browser), and modern browsers (IE11, Edge, Safari, Chrome, Firefox, Opera, iOS Safari, and Mobile Chrome).
For more information about inserting the sanitized string safely, see this article: https://www.owasp.org/index.php/XSS_(Cross_Site_Scripting)_Prevention_Cheat_Sheet.
Why js-xss
?
js-xss is lightweight (5.5k gzipped) library with an MIT license. It is also highly customizable and works well in both Node.js applications and in the browser.
Using npm:
npm install --save @esri/arcgis-html-sanitizer
Using Yarn:
yarn add @esri/arcgis-html-sanitizer
ES Modules
import { Sanitizer } from "@esri/arcgis-html-sanitizer";
CommonJS
const Sanitizer = require("@esri/arcgis-html-sanitizer").Sanitizer;
AMD (Use UMD version in ./dist/umd folder)
define(['path/to/dist/umd/arcgis-html-sanitizer'], function(Sanitizer) {
...
})
Load as script tag
<!-- Local -->
<script src="path/to/arcgis-html-sanitizer.min.js"></script>
<!-- CDN (Adjust the version as needed) -->
<script src="https://cdn.jsdelivr.net/npm/@esri/arcgis-html-sanitizer@0.7.0/dist/umd/arcgis-html-sanitizer.min.js"></script>
// Instantiate a new Sanitizer object
const sanitizer = new Sanitizer();
// Sanitize a string
const sanitizedHtml = sanitizer.sanitize(
'<img src="https://example.com/fake-image.jpg" onerror="alert(1);" />'
);
// sanitizedHtml => <img src="https://example.com/fake-image.jpg" />
// Check if a string contains invalid HTML
const validation = sanitizer.validate(
'<img src="https://example.com/fake-image.jpg" onerror="alert(1);" />'
);
// validation => {
// isValid: false
// sanitized: '<img src="https://example.com/fake-image.jpg" />'
// }
In addition to sanitizing strings, this utility also allows you to sanitize full JSON objects. This can be useful if you want to sanitize all the app data returned from the server before using it in your application.
If the value passed does not contain a valid JSON data type (String, Number, JSON Object, Array, Boolean, or null), the value will be nullified.
WARNING: You should never concatenate strings from multiple values in the JSON. Strings values may be safe and pass the sanitizer when separated but become dangerous after they are concatenated. This is intended as a convenience method only. You should run the strings through the sanitizer again after they have been concatenated.
// Instantiate a new Sanitizer object
const sanitizer = new Sanitizer();
// Deeply sanitize a JSON object
const sanitizedJSON = sanitizer.sanitize({
sample: ['<img src="https://example.com/fake-image.jpg" onerror="alert(1);\
" />']
});
// sanitizedJSON => {
// "sample": ["<img src=\"https://example.com/fake-image.jpg\" />"]
// }
The ability to sanitize URL strings has been specifically broken out to a public
method, sanitizeUrl
. This can be very useful if you want to sanitize URLs
according to the built-in rules, yet you don't want to necessarily extend the
whitelist and implement custom filtering options to force sanitization of
specific tag attributes.
// Instantiate a new Sanitizer object
const sanitizer = new Sanitizer();
// Sanitize a URL with a valid protocol
const supportedProtocol = sanitizer.sanitizeUrl(
"https://example.com/about/index.html"
);
// supportedProtocol => "https://example.com/about/index.html"
// Sanitize a URL with an invalid protocol
const unsupportedProtocol = sanitizer.sanitizeUrl(
"smb://example.com/path/to/file.html"
);
// unsupportedProtocol => ""
To prevent against XSS, untrusted data going into HTML attribute values need to be
sanitized,
including style
attribute values.
The ability to sanitize HTML attribute values has been specifically broken out to a public
method, sanitizeHTMLAttribute
, specifying the tagname, attribute name, and value.
// Instantiate a new Sanitizer object
const sanitizer = new Sanitizer();
// Sanitize a safe alt attribute value
const unquotedAttribute = sanitizer.sanitizeHTMLAttribute('img', 'alt', 'A picture');
// unquotedAttribute => "A picture"
// Sanitize an unsafe alt attribute value
const quotedAttribute = sanitizer.sanitizeHTMLAttribute('img', 'alt', '"A picture"');
// quotedAttribute => ""A picture""
// Sanitize unsafe style attribute value
const styles = sanitizer.sanitizeHTMLAttribute(
'div',
'style',
'background-image:url("javascript:alert(\"xss\")")'
);
// styles => ""
Optionally, a custom filter for style attributes can be provided as the final argument, which will be used in place of the default js-xss style attribute filter.
const styles = sanitizer.sanitizeHTMLAttribute('div', 'style', 'color:red;', {
process: (value: string) => value.indexOf('color') !== -1 ? '' : value /* or additional filtering here */
});
// styles => ""
The encodeHTML
method encodes a well-defined set of characters within an HTML string to their hexadecimal HTML entity codes to ensure safe display of the string to the end user. This encodes HTML tags for display as text within an HTML page, it does not sanitize the string to prevent XSS. If you are looking to sanitize, use the sanitize
method instead.
&
→8
<
→<
>
→>
"
→"
'
→'
/
→/
const sanitizer = new Sanitizer();
const input = `<a href="javascript:alert(document.cookies)">a link</a>`;
sanitizer.encodeHTML(input);
// => "<a href="javascript:alert(document.cookies)">a link</a>"
The encodeAttrValue
method encodes all non-alphanumeric ASCII characters to their hexadecimal HTML entity code to ensure safe use as an HTML attribute. This method only encodes characters. It does not directly sanitize the string to remove XSS. If you are looking to sanitize your string, use the sanitizeHTMLAttribute
method instead.
const sanitizer = new Sanitizer();
const input = "javascript:alert(document.cookies)";
sanitizer.encodeAttrValue(input);
// => "javascript:alert(document.cookie)"
Override the default XSS filter options by passing a valid js-xss options object as the first parameter of the constructor. Options available here: https://github.com/leizongmin/js-xss#custom-filter-rules.
You can also extend the default options instead of overriding them by passing true
as the second parameter of the constructor. When extending
the filter options whiteList
, the attribute arrays will automatically
be concatenated to the defaults instead of replacing them.
const customSanitizer = new Sanitizer({
whiteList: {
a: ['data-example']
},
escapeHtml: function () {
...
}
}, true /* extend defaults */);
If you override the built-in filtering using custom filter rules, know that you are taking full control of that particular filtering method. To do this safely, you can pass anything you don't want to handle specifically on to the js-xss
filtering method.
In this example, we want to strip any href
for the domain www.mydomain.tld
and pass along everything else to the js-xss
safeAttrValue
method.
import xss from "xss";
const customSanitizer = new Sanitizer({
safeAttrValue: function(tag, name, value, cssFilter) {
// strip anchor href attributes that point to www.mydomain.tld on any protocol,
// and allow any other href attributes
if (tag === "a" && name === "href") {
if (value.indexOf("//www.mydomain.tld") > -1) {
return "";
} else {
return xss.escapeAttrValue(value);
}
}
// pass everything else onto the `js-xss` `safeAttrValue` method
return xss.safeAttrValue(tag, name, value, cssFilter);
}
});
If something isn't working the way you expected, please take a look at previously logged issues first. Have you found a new bug? Want to request a new feature? We'd love to hear from you.
For transparency into the release cycle and in striving to maintain backward compatibility, @esri/arcgis-html-sanitizer is maintained under Semantic Versioning guidelines and will adhere to these rules whenever possible.
For more information on SemVer, please visit http://semver.org/.
Esri welcomes contributions from anyone and everyone. Please see our guidelines for contributing.
Install Dependencies
yarn install
Test
yarn test
Test in development
yarn run test --watch
Build compiled output:
- CommonJS Module outputs to
dist/node/index.js
. - ES Module outputs to
dist/esm/index.js
anddist/esm/index.min.js
. - UMD Module outputs to
dist/umd/arcgis-html-sanitizer.js
anddist/umd/arcgis-html-sanitizer.min.js
.
yarn build
Lint and fix errors
yarn run lint:fix
Copyright 2021 Esri
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
A copy of the license is available in the repository's LICENSE file.