Transcript
Content Security PolicyOR HOW TO MAKE DEVELOPERS EVEN MORE LAZIER
RIYAZ WALIKAR
whoami
Security evangelist
Do not work at a Big 4
One of the 3 OWASP Bangalore chapter leaders
Extremely talkative
Same Origin Policy
So you own http://banana.com
Code from http://potato.com should not be able to access data from http://banana.com
Browser’s sandbox and Origin protection
XSS to bypass SOP
For the love of XSS
Reflected, Stored, DOM based
Content Security Policy
The core issue exploited by XSS attacks is the browser’s inability to distinguish between script that’s intended to be part of your application, and script that’s been maliciously injected by a third-party.
http://www.html5rocks.com/en/tutorials/security/content-security-policy/
I had you at Header
Content Security Policy (CSP) defines the Content-Security-Policy HTTP header
Whitelist script sources of trusted content
Even if vulnerable to XSS, injected script will not trigger due to header definition
Building the policy
So you trust scripts only from http://banana.com and your own domain (non inline)
Content-Security-Policy: script-src 'self' http://banana.com
So you want to load images only from http://potato.com and flash content from your own domain. Also, absolutely no scripts.
Content-Security-Policy: script-src 'none'; img-src http://potato.com; object-src 'self'
CSP Directives
default-src
script-src
style-src
img-src
connect-src
font-src
object-src
media-src
child-src
sandbox
report-uri
The default-src is the default policy for loading content such as JavaScript, Images, CSS, fonts, AJAX requests, Frames and HTML5 Media
Defines valid sources of JavaScriptDefines valid sources of css (stylesheets)Defines valid sources of imagesDefines sources to which XMLHTTPRequest (AJAX), WebSocket or EventSource can fetch data from
Defines valid sources of fontsDefines valid sources of plugins (for example: flash, embed tag, applet etc.)
Defines valid source of audio and videoDefines valid source for workers and embedded frame contents.
frame-src is deprecated. child-src should be used.
More about this laterInstructs the browser to POST a reports of policy failures to a specified URI.
CSP Source Declarations
Source Value Meaning
* Wildcard, allows all origins.
'self' Allow same origin (current origin).
'none'Don't allow any resources of this directive to load.
domain.example.com Allow a domain (explicit declaration)
*.example.com Allow all subdomains on a domain. Exclude TLD.
https://example.com Exact match including protocol
https: Load from any domain but https
data: Allow data uri (eg: Base64 encoded image)
unsafe-inline
When script-src or style-src are declared, inline script tags and css are disabled
You can specify 'unsafe-inline' to execute inline script but that is precisely what CSP was designed to prevent!
unsafe-eval
CSP disables the JavaScript function eval() by design
To enable this explicitly, add 'unsafe-eval' to a script-src directive
Not advised!
sandbox
If present, browser treats the page as if it loaded inside an iframe with a sandbox attribute
The browser severely restricts the page’s functionality, disabling JS, form submissions, plugins and objects
You can keep the sandbox value empty to keep all restrictions in place, or add values: allow-forms allow-same-origin allow-scripts, and allow-top-navigation
DEMOTIME
Sources: caniuse.com/contentsecuritypolicy & Mozilla
CSP 2.0!
Several new enhancements including support for inline scripts in combination with a cryptographic nonce or hash sharing of the script itself
Content-Security-Policy: script-src 'nonce-AY778asa229b2DEADBEEF'
http://www.w3.org/TR/CSP2/
I read the following to make this presentation
http://www.w3.org/TR/2012/CR-CSP-20121115/
http://www.html5rocks.com/en/tutorials/security/content-security-policy/
http://content-security-policy.com/
http://caniuse.com/#feat=contentsecuritypolicy
https://html.spec.whatwg.org/multipage/browsers.html#sandboxing-flag-set
http://www.w3.org/TR/CSP2/
top related