Security
Sanitizing User Input
For components such as ion-alert
developers can allow for custom or user-provided content. This content can be plain text or HTML and should be considered untrusted. As with any untrusted input, it is important to sanitize it before doing anything else with it. In particular, using things like innerHTML
without sanitization provides an attack vector for bad actors to input malicious content and potentially launch a Cross Site Scripting attack (XSS).
Ionic comes built in with a basic sanitization implementation for the components it provides. However, it is not a comprehensive solution. It is up to the developer to make sure all data that is passed is sanitized. Different frameworks have different solutions for sanitizing user input, so developers should familiarize themselves with what their specific framework offers.
For developers who are not using a framework, or for developers whose framework does not provide the sanitization methods they need, we recommend using sanitize-html. This package provides a simple HTML sanitizer that allows the developer to specify the exact tags and attributes that they want to allow in their application.
Angular
Angular comes built in with the DomSanitizer
class. This helps prevent XSS issues by ensuring that values are safe to be used in the DOM. By default, Angular will mark any values it deems unsafe. For example, the following link would be marked as unsafe by Angular because it would attempt to execute some JavaScript.
public myUrl: string = 'javascript:alert("oh no!")';
...
<a [href]="myUrl">Click Me!</a>
To learn more about the built-in protections that Angular provides, see the Angular Security Guide.
React
React DOM escapes values embedded in JSX before rendering them by converting them to strings. For example, the following would be safe as name
is converted to a string before being rendered:
const name = values.name;
const element = <h1>Hello, {name}!</h1>;
However, this does not stop someone from injecting JavaScript into places such as the href
attribute of an anchor element. The following is unsafe and can potentially allow an XSS attack to occur:
const userInput = 'javascript:alert("Oh no!")';
const element = <a href={userInput}>Click Me!</a>;
If the developer needs to achieve more comprehensive sanitization, they can use the sanitize-html package.
To learn more about the built-in protections that React and JSX provide, see the React JSX Documentation.
Vue
Vue does not provide any type of sanitizing methods built in. It is recommended that developers use a package such as sanitize-html.
To learn more about the security recommendations for binding to directives such as v-html
, see the Vue Syntax Guide.
Enabling Custom HTML Parsing via innerHTML
ion-alert
, ion-infinite-scroll-content
, ion-loading
, ion-refresher-content
, and ion-toast
can accept custom HTML as strings for certain properties. These strings are added to the DOM using innerHTML
and must be properly sanitized by the developer. This behavior is disabled by default which means values passed to the affected components will always be interpreted as plaintext. Developers can enable this custom HTML behavior by setting innerHTMLTemplatesEnabled: true
in the IonicConfig.
Ejecting from the built-in sanitizer
For developers who wish to add complex HTML to components such as ion-toast
, they will need to eject from the sanitizer that is built into Ionic Framework. Developers can either disable the sanitizer across their entire app or bypass it on a case-by-case basis.
Bypassing sanitization functionality can make your application vulnerable to XSS attacks. Please exercise extreme caution when disabling the sanitizer.
Disabling the sanitizer via config
Ionic Framework provides an application config option called sanitizerEnabled
that is set to true
by default. Set this value to false
to globally disable Ionic Framework's built in sanitizer. Please note that this does not disable any sanitizing functionality provided by other frameworks such as Angular.
Bypassing the sanitizer on a case-by-case basis
Developers can also choose to eject from the sanitizer in certain scenarios. Ionic Framework provides the IonicSafeString
class that allows developers to do just that.
In order to bypass the sanitizer and use unsanitized custom HTML in the relevant Ionic components, innerHTMLTemplatesEnabled
must be set to true
in the Ionic config. See Enabling Custom HTML Parsing for more information.
Usage
- Angular
- JavaScript
- React
import { IonicSafeString, ToastController } from '@ionic/angular';
...
constructor(private toastController: ToastController) {}
async presentToast() {
const toast = await this.toastController.create({
message: new IonicSafeString('<ion-button>Hello!</ion-button>'),
duration: 2000
});
toast.present();
}
import { IonicSafeString } from '@ionic/core';
...
const async presentToast = () => {
const toast = document.createElement('ion-toast');
toast.message = new IonicSafeString('<ion-button>Hello!</ion-button>');
toast.duration = 2000;
document.body.appendChild(toast);
return toast.present();
}
import React, { useState } from 'react';
import { Animation, IonButton, IonContent, IonicSafeString, IonToast } from '@ionic/react';
export const ToastExample: React.FC = () => {
const [showToast, setShowToast] = useState(false);
return (
<IonContent>
<IonButton onClick={() => setShowToast(true)} expand="block">Show Toast</IonButton>
<IonToast
isOpen={showToast}
onDidDismiss={() => setShowToast(false)}
message={new IonicSafeString('<ion-button>Hello!</ion-button>')}
duration={2000}
/>
</IonContent>
)
};
Content Security Policies (CSP)
A Content Security Policy (CSP) is a security mechanism that helps protect web applications against certain types of attacks, such as cross-site scripting (XSS) and data injection. It is implemented through an HTTP header that instructs the browser on which sources of content, such as scripts, stylesheets, and images, are allowed to be loaded and executed on a web page.
The main purpose of a CSP is to mitigate the risks associated with code injection attacks. By defining a policy, web developers can specify from which domains or sources the browser should allow the loading and execution of various types of content. This effectively limits the potential damage that can be caused by malicious scripts or unauthorized content.
Enabling CSPs
Developers can assign a CSP to their application by setting a meta tag with the policy details and the expected nonce value on script and style tags.
<meta
http-equiv="Content-Security-Policy"
content="default-src 'self'; script-src 'self' 'nonce-randomNonceGoesHere'; style-src 'self' 'nonce-randomNonceGoesHere';"
/>
Ionic and CSP
Ionic Framework provides a function to help developers set the nonce value used when constructing the web component stylesheets. This function should be called before any Ionic components are loaded. This is required to pass the nonce value to the web components so that they can be used in a CSP environment.
import { setNonce } from '@ionic/core/loader';
setNonce('randomNonceGoesHere');
In Angular this can be called in the main.ts
file, before the application is bootstrapped.
For more information on how to use CSPs with Stencil web components, see the Stencil documentation.
Angular
Starting in Angular 16, Angular provides two options for setting the nonce value.
- Set the
ngCspNonce
attribute on the root application element as<app ngCspNonce="randomNonceGoesHere"></app>
. Use this approach if you have access to server-side templating that can add the nonce both to the header and theindex.html
when constructing the response. - Provide the nonce using the
CSP_NONCE
injection token. Use this approach if you have access to the nonce at runtime and you want to be able to cache theindex.html
.
If providing the CSP_NONCE
injection token, set the provider in your AppModule
for module projects or within the bootstrapApplication
for standalone projects.
For more information on how to use CSPs with Angular, see the Angular documentation.