Difference between revisions of "REST Security Cheat Sheet"

From OWASP
Jump to: navigation, search
m (fixed broken links)
 
(77 intermediate revisions by 13 users not shown)
Line 1: Line 1:
= DRAFT CHEAT SHEET - WORK IN PROGRESS =
+
__NOTOC__
= Introduction =
+
<div style="width:100%;height:160px;border:0,margin:0;overflow: hidden;">[[File:Cheatsheets-header.jpg|link=]]</div>
  
[http://en.wikipedia.org/wiki/Representational_state_transfer REST] or REpresentational State Transfer is a means of expressing specific entities in a system by URL path elements, REST is not an architecture but it is an architectural style to build services on top of the Web. REST allows interaction with a web-based system via simplified URL's rather than complex request body or <tt>POST</tt> parameters to request specific items from the system. This document serves as a guide (although not exhaustive) of best practices to help REST-based services.
+
{| style="padding: 0;margin:0;margin-top:10px;text-align:left;" |-
 +
| style="border-right: 1px dotted gray;padding-right:25px;" valign="top" |
 +
Last revision (mm/dd/yy): '''{{REVISIONMONTH}}/{{REVISIONDAY}}/{{REVISIONYEAR}}'''  
  
= Authentication and Session Management =
+
= Introduction =
 
+
__TOC__
RESTful web services should use session based authentication, either by establishing a session token via a POST, or using an API key as a POST body argument or as a cookie. Usernames and passwords, session tokens and API keys should not appear in the URL, as this can be captured in web server logs and makes them intrinsically valuable.
 
 
 
OK:
 
  
* [https://example.com/resourceCollection/123/action https://example.com/resourceCollection/<id>/action]
+
[http://en.wikipedia.org/wiki/Representational_state_transfer REST] (or REpresentational State Transfer) is an architectural style first described in [https://en.wikipedia.org/wiki/Roy_Fielding Roy Fielding]'s Ph.D. dissertation on [https://www.ics.uci.edu/~fielding/pubs/dissertation/top.htm Architectural Styles and the Design of Network-based Software Architectures]. It evolved as Fielding wrote the HTTP/1.1 and URI specs and has been proven to be well-suited for developing distributed hypermedia applications. While REST is more widely applicable, it is most commonly used within the context of communicating with services via HTTP.
* https://twitter.com/vanderaj/lists
 
  
NOT OK:
+
The key abstraction of information in REST is a resource. A REST API resource is identified by a URI, usually a HTTP URL. REST components use connectors to perform actions on a resource by using a representation to capture the current or intended state of the resource and transferring that representation. The primary connector types are client and server, secondary connectors include cache, resolver and tunnel. In order to implement flows with REST APIs, resources are typically created, read, updated and deleted. For example, an ecommerce site may offer methods to create an empty shopping cart, to add items to the cart and to check out the cart.
  
* [https://example.com/controller/123/action?apiKey=a53f435643de32 https://example.com/controller/<id>/action?apiKey=a53f435643de32] (API Key in URL)
+
Another key feature of REST applications is the use of standard HTTP verbs and error codes in the pursuit or removing unnecessary variation among different services.
* [http://example.com/controller/123/action?apiKey=a53f435643de32 http://example.com/controller/<id>/action?apiKey=a53f435643de32] (transaction not protected by TLS and API Key in URL)
 
  
== Protect Session State ==
+
Another key feature of REST applications is the use of HATEOS or Hypermedia as the Engine of Application State. This provides REST applications a self-documenting nature making it easier for developers to interact with a REST service without a priori knowledge.
  
Many web services are written to be as stateless as possible. This usually ends up with a state blob being sent as part of the transaction.  
+
= HTTPS =
 +
Secure REST services must only provide HTTPS endpoints. This protects authentication credentials in transit, for example passwords, API keys or JSON Web Tokens. It also allows clients to authenticate the service and guarantees integrity of the transmitted data.
  
* Consider just using the session token or API key to maintain client state in a server side cache. This is directly equivalent to how normal web apps do it, and there's a reason why this is moderately safe.
+
See the [[Transport Layer Protection Cheat Sheet]] for additional information.
* Anti-replay. Attackers will cut and paste a blob and become someone else.  Consider using a time limited encryption key, keyed against the session token or API key, date and time and incoming IP address. In general, implement some protection of local client storage of the authentication token to mitigate replay attacks.
 
* Don't make it easy to decrypt and change the internal state to be much better than it should be.  
 
  
In short, even if you have a brochureware web site, don't put in https://example.com/users/2313/edit?isAdmin=false&debug=false&allowCSRPanel=false as you will quickly end up with a lot of admins, and help desk helpers, and "developers".
+
Consider the use of mutually authenticated client-side certificates to provide additional protection for highly privileged web services.
  
= Authorization =
+
= Access Control =
 +
Non-public REST services must perform access control at each API endpoint. Web services in monolithic applications implement this by means of user authentication, authorisation logic and session management. This has several drawbacks for modern architectures which compose multiple micro services following the RESTful style.
 +
* in order to minimise latency and reduce coupling between services, the access control decision should be taken locally by REST endpoints
 +
* user authentication should be centralised in a Identity Provider (IdP), which issues access tokens
  
== Anti-farming ==
+
= JWT =
 +
There seems to be a convergence towards using [https://tools.ietf.org/html/rfc7519 JSON Web Tokens] (JWT) as the format for security tokens. JWTs are JSON data structures containing a set of claims that can be used for access control decisions. A cryptographic signature or message authentication code (MAC) can be used to protect the integrity of the JWT. 
 +
* Ensure JWTs are integrity protected by either a signature or a MAC. Do not allow the unsecured JWTs: {"alg":"none"}. See https://tools.ietf.org/html/rfc7519#section-6.1
 +
* In general, signatures should be preferred over MACs for integrity protection of JWTs.
 +
If MACs are used for integrity protection, every service that is able to validate JWTs can also create new JWTs using the same key. This means that all services using the same key have to mutually trust each other. Another consequence of this is that a compromise of any service also compromises all other services sharing the same key. See https://tools.ietf.org/html/rfc7515#section-10.5 for additional information.
  
Many RESTful web services are put up, and then farmed, such as a price matching website or aggregation service. There's no technical method of preventing this use, so strongly consider means to encourage it as a business model by making high velocity farming is possible for a fee, or contractually limiting service using terms and conditions. CAPTCHAs and similar methods can help reduce simpler adversaries, but not well funded or technically competent adversaries. Using mutually assured client side TLS certificates may be a method of limiting access to trusted organizations, but this is by no means certain, particularly if certificates are posted deliberately or by accident to the Internet.
+
The relying party or token consumer validates a JWT by verifying its integrity and claims contained.
 +
* A relying party must verify the integrity of the JWT based on its own configuration or hard-coded logic. It must not rely on the information of the JWT header to select the verification algorithm. See https://www.chosenplaintext.ca/2015/03/31/jwt-algorithm-confusion.html and https://www.youtube.com/watch?v=bW5pS4e_MX8
 +
Some claims have been standardised and should be present in JWT used for access controls. At least the following of the standard claims should be verified:
 +
* 'iss' or issuer - is this a trusted issuer? Is it the expected owner of the signing key?
 +
* 'aud' or audience - is the relying party in the target audience for this JWT?
 +
* 'exp' or expiration time - is the current time before the end of the validity period of this token?
 +
* 'nbf' or not before time - is the current time after the start of the validity period of this token?
  
== Protect HTTP methods ==
+
= API Keys =
 +
Public REST services without access control run the risk of being farmed leading to excessive bills for bandwidth or compute cycles. API keys can be used to mitigate this risk. They are also often used by organisation to monetize APIs; instead of blocking high-frequency calls, clients are given access in accordance to a purchased access plan.
  
RESTful API often use GET (read), POST (create), PUT (replace/update) and DELETE (to delete a record). Not all of these are valid choices for every single resource collection, user, or action. Make sure the  incoming HTTP method is valid for the session token/API key and associated resource collection, action, and record. For example, if you have an RESTful API for a library, it's not okay to allow anonymous users to DELETE book catalog entries, but it's fine for them to GET a book catalog entry, whereas for the librarian, both of these are valid uses.  
+
API keys can reduce the impact of denial-of-service attacks. However, when they are issued to third-party clients, they are relatively easy to compromise.
 +
* Require API keys for every request to the protected endpoint.
 +
* Return 429 "Too Many Requests" HTTP response code if requests are coming in too quickly.
 +
* Revoke the API key if the client violates the usage agreement.
 +
* Do not rely exclusively on API keys to protect sensitive, critical or high-value resources.
  
== Whitelist Allowable Methods ==
+
= Restrict HTTP methods =
 +
* Apply a whitelist of permitted HTTP Methods e.g. GET, POST, PUT
  
It is common with RESTful services to allow multiple methods for a given URL for different operations on that entity. For example, a <tt>GET</tt> request might read the entity while <tt>POST</tt> would update an existing entity, <tt>PUT</tt> would create a new entity, and <tt>DELETE</tt> would delete an existing entity. It is important for the service to properly restrict the allowable verbs such that only the allowed verbs will work, all others return a proper response code (for example, a <tt>403 Forbidden</tt>).
+
* Reject all requests not matching the whitelist with HTTP response code 405 Method not allowed
 +
* Make sure the caller is authorised to use the incoming HTTP method on the resource collection, action, and record
 +
In Java EE in particular, this can be difficult to implement properly. See [https://packetstormsecurity.com/files/66779/Bypassing_VBAAC_with_HTTP_Verb_Tampering.pdf.html Bypassing Web Authentication and Authorization with HTTP Verb Tampering] for an explanation of this common misconfiguration.
  
In Java EE in particular, this can be difficult to implement properly. See [https://www.aspectsecurity.com/wp-content/plugins/download-monitor/download.php?id=18 Bypassing Web Authentication and Authorization with HTTP Verb Tampering] for an explanation of this common misconfiguration.
+
= Input validation =
 +
* Do not trust input parameters/objects
 +
* Validate input: length / range / format and type
 +
* Achieve an implicit input validation by using strong types like numbers, booleans, dates, times or fixed data ranges in API parameters
 +
* Constrain string inputs with regexps
 +
* Reject unexpected/illegal content
 +
* Make use of validation/sanitation libraries or frameworks in your specific language
 +
* Define an appropriate request size limit and reject requests exceeding the limit with HTTP response status 413 Request Entity Too Large
 +
* Consider logging input validation failures. Assume that someone who is performing hundreds of failed input validations per second is up to no good.
 +
* Have a look at input validation cheat sheet for comprehensive explanation
  
== Protect privileged actions and sensitive resource collections ==
+
* Use a secure parser for parsing the incoming messages. If you are using XML, make sure to use a parser that is not vulnerable to [https://www.owasp.org/index.php/XML_External_Entity_(XXE)_Processing XXE] and similar attacks.
  
Not every user has a right to every web service. This is vital, as you don't want administrative web services to be misused:
+
= Validate content types =
 +
A REST request or response body should match the intended content type in the header. Otherwise this could cause misinterpretation at the consumer/producer side and lead to code injection/execution.
 +
* Document all supported content types in your API
  
* https://example.com/admin/exportAllData
+
== Validate request content types ==
 +
* Reject requests containing unexpected or missing content type headers with HTTP response status 406 Unacceptable or 415 Unsupported Media Type
  
The session token or API key should be sent along as a cookie or body parameter to ensure that privileged collections or actions are properly protected from unauthorized use.
+
* For XML content types ensure appropriate XML parser hardening, see the [[XML External Entity (XXE) Prevention Cheat Sheet|cheat shee]]<nowiki/>t
  
== Protect against cross-site request forgery ==
+
* Avoid accidentally exposing unintended content types by explicitly defining content types e.g. [https://jersey.github.io/ Jersey] (Java) ''@consumes("application/json"); @produces("application/json")''. This avoids XXE-attack vectors for example.
  
For resources exposed by RESTful web services, it's important to make sure any PUT, POST and DELETE request is protected from Cross Site Request Forgery. Typically one would use a token based approach. See [[Cross-Site Request Forgery (CSRF) Prevention Cheat Sheet]] for more information on how to implement CSRF-protection.  
+
== Send safe response content types ==
 +
It is common for REST services to allow multiple response types (e.g. "application/xml" or "application/json", and the client specifies the preferred order of response types by the Accept header in the request.  
 +
* Do NOT simply copy the Accept header to the Content-type header of the response.
 +
* Reject the request (ideally with a 406 Not Acceptable response) if the Accept header does not specifically contain one of the allowable types.
 +
Services including script code (e.g. JavaScript) in their responses must be especially careful to defend against header injection attack.
 +
* ensure sending intended content type headers in your response matching your body content e.g. "application/json" and not "application/javascript"
  
CSRF is easily achieved even using random tokens if any XSS exists within your application, so please make sure you understand [[XSS (Cross Site Scripting) Prevention Cheat Sheet|how to prevent XSS]].
+
= Management endpoints =
 +
* Avoid exposing management endpoints via Internet.
 +
* If management endpoints must be accessible via the Internet, make sure that users must use a strong authentication mechanism, e.g. multi-factor.
 +
* Expose management endpoints via different HTTP ports or hosts preferably on a different NIC and restricted subnet.
 +
* Restrict access to these endpoints by firewall rules  or use of access control lists.
  
== Direct object references ==
+
= Error handling =
 +
* Respond with generic error messages - avoid revealing details of the failure unnecessarily
 +
* Do not pass technical details (e.g. call stacks or other internal hints) to the client
  
It may seem obvious, but if you had a bank account REST web service, you have to make sure there is adequate checking of primary and foreign keys:
+
= Audit logs =
 +
* Write audit logs before and after security related events
 +
* Consider logging token validation errors in order to detect attacks
 +
* Take care of log injection attacks by sanitising log data beforehand
  
* https://example.com/account/325365436/transfer?amount=$100.00&toAccount=473846376
+
= Security headers =
  
In this case, it would be possible to transfer money from any account to any other account, which is clearly insane. Not even a random token makes this safe.  
+
To make sure the content of a given resources is interpreted correctly by the browser, the server should always send the Content-Type header with the correct Content-Type, and preferably the Content-Type header should include a charset. The server should also send an <tt>X-Content-Type-Options: nosniff</tt> to make sure the browser does not try to detect a different Content-Type than what is actually sent (can lead to XSS).
  
* https://example.com/invoice/2362365
+
Additionally the server should send an <tt>X-Frame-Options: deny</tt> to protect against drag'n drop clickjacking attacks in older browsers.
  
In this case, it would be possible to get a copy of all invoices.  
+
== CORS ==
 +
Cross-Origin Resource Sharing (CORS) is a W3C standard to flexibly specify what cross-domain requests are permitted. By delivering appropriate CORS Headers your REST API signals to the browser which domains, AKA origins, are allowed to make JavaScript calls to the REST service.
 +
* Disable CORS headers if cross-domain calls are not supported
 +
* Be as specific as possible and as general as necessary when setting the origins of cross-domain calls
 +
In Spring Boot (Java), for example, CORS support is disabled by default and is only enabled once the ''endpoints.cors.allowed-origins'' property has been set. The configuration below permits GET and POST calls from the example.com domain:<blockquote>endpoints.cors.allowed-origins=<nowiki>https://example.com</nowiki></blockquote><blockquote>endpoints.cors.allowed-methods=GET,POST</blockquote>
  
Please make sure you understand how to protect against [[Top_10_2010-A4-Insecure_Direct_Object_References|direct object references]] in the OWASP Top 10 2010.
+
= Sensitive information in HTTP requests =
 +
RESTful web services should be careful to prevent leaking credentials. Passwords, security tokens, and API keys should not appear in the URL, as this can be captured in web server logs, which makes them intrinsically valuable.
 +
* In POST/PUT requests sensitive data should be transferred in the request body or request headers
 +
* In GET requests sensitive data should be transferred in an HTTP Header
 +
OK:
  
= Input Validation =
+
* <nowiki>https://example.com/resourceCollection/</nowiki><id>/action
 +
* <nowiki>https://twitter.com/vanderaj/lists</nowiki>
  
== Input validation 101 ==
+
NOT OK:
 
 
Everything you know about input validation applies to RESTful web services, but add 10% because automated tools can easily fuzz your interfaces for hours on end at high velocity. So:
 
 
 
* Assist the user > Reject input > Sanitize (filtering) > No input validation
 
 
 
Assisting the user makes the most sense, as the most common scenario is "problem exists between keyboard and computer" (PEBKAC). Help the user input high quality data into your web services, such as ensuring a Zip code makes sense for the supplied address, or the date makes sense. If not, reject that input. If they continue on, or it's a text field or some other difficult to validate field, input sanitization is a losing proposition but still better than XSS or SQL injection. If you're already reduced to  sanitization or no input validation, make sure output encoding is very strong for your application.
 
 
 
Log input validation failures, particularly if you assume that client side code you wrote is going to call your web services. The reality is that anyone can call your web services, so assume that someone who is performing hundreds of failed input validations per second is up to no good. Also consider rate limiting the API to a certain number of requests per hour or day to prevent abuse.
 
 
 
== Strong typing ==
 
 
 
It's difficult to perform most attacks if the only allowed values are true or false, or a number, or one of a small number of acceptable values. Strongly type incoming data as quickly as possible.
 
 
 
== Validate Incoming Content-Types ==
 
 
 
When POSTing or PUTing new data, the client will specify the a Content-Type (e.g. <tt>application/xml</tt> or <tt>application/json</tt>) of the incoming data. The client should never assume the Content-Type, but always check that the Content-Type header and the content is of the same type. A lack of Content-Type header or an unexpected Content-Type header, should result in the server rejecting the Content with a <tt>406 Not Acceptable</tt> response.
 
 
 
== Validate Response Types ==
 
 
 
It is common for REST services to allow multiple response types (e.g. <tt>application/xml</tt> or <tt>application/json</tt>, and the client specifies the preferred order of response types by the <tt>Accept</tt> header in the request. '''Do NOT''' simply copy the <tt>Accept</tt> header to the <tt>Content-type</tt> header of the response. Reject the request (ideally with a <tt>406 Not Acceptable</tt> response) if the <tt>Accept</tt> header does not specifically contain one of the allowable types.
 
 
 
Because there are many MIME types for the typical response types, it's important to document for clients specifically which MIME types should be used.
 
 
 
== XML Input Validation ==
 
 
 
XML-based services must ensure that they are protected against common XML based attacks by using secure XML-parsing. This typically means XML External Entity attacks, XML-signature wrapping etc. See  [http://ws-attacks.org http://ws-attacks.org] for examples of such attacks.
 
 
 
= Output Encoding =
 
 
 
== Send security headers ==
 
 
 
To make sure the content of a given resources is interpreted correctly by the browser, the server should always send the Content-Type header with the correct Content-Type. The server should also send an <tt>X-Content-Type-Options: nosniff</tt> to make sure the browser does not try to detect a different Content-Type than what is actually sent (can lead to XSS).
 
 
 
Additionally the client should send an <tt>X-Frame-Options: deny</tt> to protect against drag'n drop clickjacking attacks in older browsers.
 
 
 
== JSON encoding ==
 
 
 
A key concern with JSON encoders is to prevent arbitrary JavaScript remote code execution within the browser, or if you're using node.js, on the server. It's vital that you use the JSON library of your choice to encode user supplied data properly to prevent the execution of user supplied input on the browser.
 
 
 
When sending values to the browser, strongly consider using .value rather than .innerHTML updates, as this protects against simple DOM XSS attacks.
 
 
 
== XML encoding ==
 
 
 
XML should never be hand concatenated, it should always be constructed using an XML parser. This ensures that the XML content sent to the browser is parseable, and should not contain XML injection. For more information, please see the [[Web Service Security Cheat Sheet]].
 
 
 
= Cryptography =
 
 
 
== Data in transit ==
 
  
Unless completely read only public information, the use of TLS should be mandated, particularly where credentials, updates, deletions and any value transactions are performed. The overhead of TLS is negligible on modern hardware, with a minor latency increase that is more than compensated by safety for the end user.  
+
* <nowiki>https://example.com/controller/</nowiki><id>/action?apiKey=a53f435643de32 (API Key in URL)
  
Consider the use of mutually authenticated client side certificates to provide additional protection for highly privileged web services.  
+
= HTTP Return Code =
 +
HTTP defines status codes [https://en.wikipedia.org/wiki/List_of_HTTP_status_codes].
 +
When designing REST API, don't just use 200 for success or 404 for error. Always use the semantically appropriate status code for the response.
  
== Data in storage ==
+
Here is a non-exhaustive selection of security related REST API status codes. Use it to ensure you return the correct code. 
 +
{| class="wikitable"
 +
!Status code
 +
!Message
 +
!Description
 +
|-
 +
|200
 +
|OK
 +
|Response to a successful REST API action. The HTTP method can be GET, POST, PUT, PATCH or DELETE
 +
|-
 +
|201
 +
|Created
 +
|The request has been fulfilled and resource created. A URI for the created resource is returned in the Location header
 +
|-
 +
|202
 +
|Accepted
 +
|The request has been accepted for processing, but processing is not yet complete
 +
|-
 +
|400
 +
|Bad Request
 +
|The request is malformed, such as message body format error
 +
|-
 +
|401
 +
|Unauthorized
 +
|Wrong or no authentication ID/password provided
 +
|-
 +
|403
 +
|Forbidden
 +
|It's used when the authentication succeeded but authenticated user doesn't have permission to the request resource
 +
|-
 +
|404
 +
|Not Found
 +
|When a non-existent resource is requested
 +
|-
 +
|406
 +
|Unacceptable
 +
|The client presented a content type in the Accept header which is not supported by the server API
 +
|-
 +
|405
 +
|Method Not Allowed
 +
|The error for an unexpected HTTP method. For example, the REST API is expecting HTTP GET, but HTTP PUT is used
 +
|-
 +
|413
 +
|Payload too large
 +
|Use it to signal that the request size exceeded the given limit e.g. regarding file uploads
 +
|-
 +
|415
 +
|Unsupported Media Type
 +
|The requested content type is not supported by the REST service
 +
|-
 +
|429
 +
|Too Many Requests
 +
|The error is used when there may be DOS attack detected or the request is rejected due to rate limiting
 +
|-
 +
|500
 +
|Internal Server Error
 +
|An unexpected condition prevented the server from fulfilling the request. Be aware that the response should not reveal internal information that helps an attacker, e.g. detailed error messages or stack traces.
 +
|-
 +
|501
 +
|Not Implemented
 +
|The REST service does not implement the requested operation yet
 +
|-
 +
|503
 +
|Service Unavailable
 +
|The REST service is temporarily unable to process the request. Used to inform the client it should retry at a later time.
 +
|}
  
Leading practices are recommended as per any web application when it comes to correctly handling stored sensitive or regulated data. For more information, please see [[Top 10 2010-A7|OWASP Top 10 2010 - A7 Insecure Cryptographic Storage]].
+
= Authors and primary editors  =
  
= Related Articles =
+
Erlend Oftedal - erlend.oftedal@owasp.org<br />
 +
Andrew van der Stock - vanderaj@owasp.org<br />
 +
Tony Hsu Hsiang Chih- Hsiang_chihi@yahoo.com<br />
 +
Johan Peeters - yo@johanpeeters.com<br />
 +
Jan Wolff - jan.wolff@owasp.org<br />
 +
Rocco Gränitz - rocco.graenitz@owasp.org
  
{{Cheatsheet_Navigation}}
+
= Other Cheatsheets =
  
= Authors and Primary Editors  =
+
{{Cheatsheet_Navigation_Body}}
  
Erlend Oftedal - erlend.oftedal@owasp.org<br/>
+
|}
Andrew van der Stock - vanderaj@owasp.org<br/>
 
  
 
[[Category:Cheatsheets]]
 
[[Category:Cheatsheets]]

Latest revision as of 09:02, 28 December 2018

Cheatsheets-header.jpg

Last revision (mm/dd/yy): 12/28/2018

Introduction

REST (or REpresentational State Transfer) is an architectural style first described in Roy Fielding's Ph.D. dissertation on Architectural Styles and the Design of Network-based Software Architectures. It evolved as Fielding wrote the HTTP/1.1 and URI specs and has been proven to be well-suited for developing distributed hypermedia applications. While REST is more widely applicable, it is most commonly used within the context of communicating with services via HTTP.

The key abstraction of information in REST is a resource. A REST API resource is identified by a URI, usually a HTTP URL. REST components use connectors to perform actions on a resource by using a representation to capture the current or intended state of the resource and transferring that representation. The primary connector types are client and server, secondary connectors include cache, resolver and tunnel. In order to implement flows with REST APIs, resources are typically created, read, updated and deleted. For example, an ecommerce site may offer methods to create an empty shopping cart, to add items to the cart and to check out the cart.

Another key feature of REST applications is the use of standard HTTP verbs and error codes in the pursuit or removing unnecessary variation among different services.

Another key feature of REST applications is the use of HATEOS or Hypermedia as the Engine of Application State. This provides REST applications a self-documenting nature making it easier for developers to interact with a REST service without a priori knowledge.

HTTPS

Secure REST services must only provide HTTPS endpoints. This protects authentication credentials in transit, for example passwords, API keys or JSON Web Tokens. It also allows clients to authenticate the service and guarantees integrity of the transmitted data.

See the Transport Layer Protection Cheat Sheet for additional information.

Consider the use of mutually authenticated client-side certificates to provide additional protection for highly privileged web services.

Access Control

Non-public REST services must perform access control at each API endpoint. Web services in monolithic applications implement this by means of user authentication, authorisation logic and session management. This has several drawbacks for modern architectures which compose multiple micro services following the RESTful style.

  • in order to minimise latency and reduce coupling between services, the access control decision should be taken locally by REST endpoints
  • user authentication should be centralised in a Identity Provider (IdP), which issues access tokens

JWT

There seems to be a convergence towards using JSON Web Tokens (JWT) as the format for security tokens. JWTs are JSON data structures containing a set of claims that can be used for access control decisions. A cryptographic signature or message authentication code (MAC) can be used to protect the integrity of the JWT.

  • Ensure JWTs are integrity protected by either a signature or a MAC. Do not allow the unsecured JWTs: {"alg":"none"}. See https://tools.ietf.org/html/rfc7519#section-6.1
  • In general, signatures should be preferred over MACs for integrity protection of JWTs.

If MACs are used for integrity protection, every service that is able to validate JWTs can also create new JWTs using the same key. This means that all services using the same key have to mutually trust each other. Another consequence of this is that a compromise of any service also compromises all other services sharing the same key. See https://tools.ietf.org/html/rfc7515#section-10.5 for additional information.

The relying party or token consumer validates a JWT by verifying its integrity and claims contained.

Some claims have been standardised and should be present in JWT used for access controls. At least the following of the standard claims should be verified:

  • 'iss' or issuer - is this a trusted issuer? Is it the expected owner of the signing key?
  • 'aud' or audience - is the relying party in the target audience for this JWT?
  • 'exp' or expiration time - is the current time before the end of the validity period of this token?
  • 'nbf' or not before time - is the current time after the start of the validity period of this token?

API Keys

Public REST services without access control run the risk of being farmed leading to excessive bills for bandwidth or compute cycles. API keys can be used to mitigate this risk. They are also often used by organisation to monetize APIs; instead of blocking high-frequency calls, clients are given access in accordance to a purchased access plan.

API keys can reduce the impact of denial-of-service attacks. However, when they are issued to third-party clients, they are relatively easy to compromise.

  • Require API keys for every request to the protected endpoint.
  • Return 429 "Too Many Requests" HTTP response code if requests are coming in too quickly.
  • Revoke the API key if the client violates the usage agreement.
  • Do not rely exclusively on API keys to protect sensitive, critical or high-value resources.

Restrict HTTP methods

  • Apply a whitelist of permitted HTTP Methods e.g. GET, POST, PUT
  • Reject all requests not matching the whitelist with HTTP response code 405 Method not allowed
  • Make sure the caller is authorised to use the incoming HTTP method on the resource collection, action, and record

In Java EE in particular, this can be difficult to implement properly. See Bypassing Web Authentication and Authorization with HTTP Verb Tampering for an explanation of this common misconfiguration.

Input validation

  • Do not trust input parameters/objects
  • Validate input: length / range / format and type
  • Achieve an implicit input validation by using strong types like numbers, booleans, dates, times or fixed data ranges in API parameters
  • Constrain string inputs with regexps
  • Reject unexpected/illegal content
  • Make use of validation/sanitation libraries or frameworks in your specific language
  • Define an appropriate request size limit and reject requests exceeding the limit with HTTP response status 413 Request Entity Too Large
  • Consider logging input validation failures. Assume that someone who is performing hundreds of failed input validations per second is up to no good.
  • Have a look at input validation cheat sheet for comprehensive explanation
  • Use a secure parser for parsing the incoming messages. If you are using XML, make sure to use a parser that is not vulnerable to XXE and similar attacks.

Validate content types

A REST request or response body should match the intended content type in the header. Otherwise this could cause misinterpretation at the consumer/producer side and lead to code injection/execution.

  • Document all supported content types in your API

Validate request content types

  • Reject requests containing unexpected or missing content type headers with HTTP response status 406 Unacceptable or 415 Unsupported Media Type
  • For XML content types ensure appropriate XML parser hardening, see the cheat sheet
  • Avoid accidentally exposing unintended content types by explicitly defining content types e.g. Jersey (Java) @consumes("application/json"); @produces("application/json"). This avoids XXE-attack vectors for example.

Send safe response content types

It is common for REST services to allow multiple response types (e.g. "application/xml" or "application/json", and the client specifies the preferred order of response types by the Accept header in the request.

  • Do NOT simply copy the Accept header to the Content-type header of the response.
  • Reject the request (ideally with a 406 Not Acceptable response) if the Accept header does not specifically contain one of the allowable types.

Services including script code (e.g. JavaScript) in their responses must be especially careful to defend against header injection attack.

  • ensure sending intended content type headers in your response matching your body content e.g. "application/json" and not "application/javascript"

Management endpoints

  • Avoid exposing management endpoints via Internet.
  • If management endpoints must be accessible via the Internet, make sure that users must use a strong authentication mechanism, e.g. multi-factor.
  • Expose management endpoints via different HTTP ports or hosts preferably on a different NIC and restricted subnet.
  • Restrict access to these endpoints by firewall rules  or use of access control lists.

Error handling

  • Respond with generic error messages - avoid revealing details of the failure unnecessarily
  • Do not pass technical details (e.g. call stacks or other internal hints) to the client

Audit logs

  • Write audit logs before and after security related events
  • Consider logging token validation errors in order to detect attacks
  • Take care of log injection attacks by sanitising log data beforehand

Security headers

To make sure the content of a given resources is interpreted correctly by the browser, the server should always send the Content-Type header with the correct Content-Type, and preferably the Content-Type header should include a charset. The server should also send an X-Content-Type-Options: nosniff to make sure the browser does not try to detect a different Content-Type than what is actually sent (can lead to XSS).

Additionally the server should send an X-Frame-Options: deny to protect against drag'n drop clickjacking attacks in older browsers.

CORS

Cross-Origin Resource Sharing (CORS) is a W3C standard to flexibly specify what cross-domain requests are permitted. By delivering appropriate CORS Headers your REST API signals to the browser which domains, AKA origins, are allowed to make JavaScript calls to the REST service.

  • Disable CORS headers if cross-domain calls are not supported
  • Be as specific as possible and as general as necessary when setting the origins of cross-domain calls
In Spring Boot (Java), for example, CORS support is disabled by default and is only enabled once the endpoints.cors.allowed-origins property has been set. The configuration below permits GET and POST calls from the example.com domain:
endpoints.cors.allowed-origins=https://example.com
endpoints.cors.allowed-methods=GET,POST

Sensitive information in HTTP requests

RESTful web services should be careful to prevent leaking credentials. Passwords, security tokens, and API keys should not appear in the URL, as this can be captured in web server logs, which makes them intrinsically valuable.

  • In POST/PUT requests sensitive data should be transferred in the request body or request headers
  • In GET requests sensitive data should be transferred in an HTTP Header

OK:

  • https://example.com/resourceCollection/<id>/action
  • https://twitter.com/vanderaj/lists

NOT OK:

  • https://example.com/controller/<id>/action?apiKey=a53f435643de32 (API Key in URL)

HTTP Return Code

HTTP defines status codes [1]. When designing REST API, don't just use 200 for success or 404 for error. Always use the semantically appropriate status code for the response.

Here is a non-exhaustive selection of security related REST API status codes. Use it to ensure you return the correct code.

Status code Message Description
200 OK Response to a successful REST API action. The HTTP method can be GET, POST, PUT, PATCH or DELETE
201 Created The request has been fulfilled and resource created. A URI for the created resource is returned in the Location header
202 Accepted The request has been accepted for processing, but processing is not yet complete
400 Bad Request The request is malformed, such as message body format error
401 Unauthorized Wrong or no authentication ID/password provided
403 Forbidden It's used when the authentication succeeded but authenticated user doesn't have permission to the request resource
404 Not Found When a non-existent resource is requested
406 Unacceptable The client presented a content type in the Accept header which is not supported by the server API
405 Method Not Allowed The error for an unexpected HTTP method. For example, the REST API is expecting HTTP GET, but HTTP PUT is used
413 Payload too large Use it to signal that the request size exceeded the given limit e.g. regarding file uploads
415 Unsupported Media Type The requested content type is not supported by the REST service
429 Too Many Requests The error is used when there may be DOS attack detected or the request is rejected due to rate limiting
500 Internal Server Error An unexpected condition prevented the server from fulfilling the request. Be aware that the response should not reveal internal information that helps an attacker, e.g. detailed error messages or stack traces.
501 Not Implemented The REST service does not implement the requested operation yet
503 Service Unavailable The REST service is temporarily unable to process the request. Used to inform the client it should retry at a later time.

Authors and primary editors

Erlend Oftedal - erlend.oftedal@owasp.org
Andrew van der Stock - vanderaj@owasp.org
Tony Hsu Hsiang Chih- Hsiang_chihi@yahoo.com
Johan Peeters - yo@johanpeeters.com
Jan Wolff - jan.wolff@owasp.org
Rocco Gränitz - rocco.graenitz@owasp.org

Other Cheatsheets