The web application framework we use has built-in support for handling Cross-site Request Forgery. This works well when data is posted with a browser to our webserver.
Currently we are developing an API in which an uploaded XML file is processed by the same application framework. Our API requires a unique token in the uploaded XML file for authentication. Since CSRF detection is enabled by default and the XML file does not contain a CSRF token we currently can not upload any data through this API.
However, we can quite easily disable CSRF detection, but is this safe?
A post here states -- quite boldly -- the following.
It is safe to remove csrf for API calls as the particular vulnerability can only be executed through a web browser.
Is this true? Can nothing similar to a CSRF attack happen through an API?
That depends on how you use the API. Say if the website using the API is vulnerable to CSRF, that means the API is also vulnerable.
Wekipedia says that
CSRF exploits the trust that a site has in a user's browser.
To support API calls the server requires that the credentials be sent along with every request (or some equivalent like digest, security handle, hash). If the credentials are stored in application memory (like mobile app) API is not vulnerable to CSRF. But if the credentials are saved in a session or cookie the API is exposed to CSRF
It depends on what you mean by "disable CSRF detection".
Some pointers:
As long as you do validate the unique authentication token without fail, then there is no way the attacker can spoof a valid request without a valid token. This is straightforward.
"unique authentication token" here refers to something that is not sent by browsers automatically. (So no using stuff like HTTP Basic / Digest, Cookie header and etc.) It must be something unique that you (the API creator) came up with. This can be as easy as an additional Foobar:the_unique_token header.
Note that it is perfectly fine to identify the client based on the Cookie (or other tokens that browsers automatically send), but you must only allow entry when the unique token is provided.
As long as the attacker can spoof a valid request as long as he is able to guess/obtain the token (the_unique_token). So the token needs to be long, random, and single-use to be secure.
Related
When building SPA style applications using frameworks like Angular, Ember, React, etc. what do people believe to be some best practices for authentication and session management? I can think of a couple of ways of considering approaching the problem.
Treat it no differently than authentication with a regular web application assuming the API and and UI have the same origin domain.
This would likely involve having a session cookie, server side session storage and probably some session API endpoint that the authenticated web UI can hit to get current user information to help with personalization or possibly even determining roles/abilities on the client side. The server would still enforce rules protecting access to data of course, the UI would just use this information to customize the experience.
Treat it like any third-party client using a public API and authenticate with some sort of token system similar to OAuth. This token mechanism would used by the client UI to authenticate each and every request made to the server API.
I'm not really much of an expert here but #1 seems to be completely sufficient for the vast majority of cases, but I'd really like to hear some more experienced opinions.
This question has been addressed, in a slightly different form, at length, here:
RESTful Authentication
But this addresses it from the server-side. Let's look at this from the client-side. Before we do that, though, there's an important prelude:
Javascript Crypto is Hopeless
Matasano's article on this is famous, but the lessons contained therein are pretty important:
https://www.nccgroup.trust/us/about-us/newsroom-and-events/blog/2011/august/javascript-cryptography-considered-harmful/
To summarize:
A man-in-the-middle attack can trivially replace your crypto code with <script> function hash_algorithm(password){ lol_nope_send_it_to_me_instead(password); }</script>
A man-in-the-middle attack is trivial against a page that serves any resource over a non-SSL connection.
Once you have SSL, you're using real crypto anyways.
And to add a corollary of my own:
A successful XSS attack can result in an attacker executing code on your client's browser, even if you're using SSL - so even if you've got every hatch battened down, your browser crypto can still fail if your attacker finds a way to execute any javascript code on someone else's browser.
This renders a lot of RESTful authentication schemes impossible or silly if you're intending to use a JavaScript client. Let's look!
HTTP Basic Auth
First and foremost, HTTP Basic Auth. The simplest of schemes: simply pass a name and password with every request.
This, of course, absolutely requires SSL, because you're passing a Base64 (reversibly) encoded name and password with every request. Anybody listening on the line could extract username and password trivially. Most of the "Basic Auth is insecure" arguments come from a place of "Basic Auth over HTTP" which is an awful idea.
The browser provides baked-in HTTP Basic Auth support, but it is ugly as sin and you probably shouldn't use it for your app. The alternative, though, is to stash username and password in JavaScript.
This is the most RESTful solution. The server requires no knowledge of state whatsoever and authenticates every individual interaction with the user. Some REST enthusiasts (mostly strawmen) insist that maintaining any sort of state is heresy and will froth at the mouth if you think of any other authentication method. There are theoretical benefits to this sort of standards-compliance - it's supported by Apache out of the box - you could store your objects as files in folders protected by .htaccess files if your heart desired!
The problem? You are caching on the client-side a username and password. This gives evil.ru a better crack at it - even the most basic of XSS vulnerabilities could result in the client beaming his username and password to an evil server. You could try to alleviate this risk by hashing and salting the password, but remember: JavaScript Crypto is Hopeless. You could alleviate this risk by leaving it up to the Browser's Basic Auth support, but.. ugly as sin, as mentioned earlier.
HTTP Digest Auth
Is Digest authentication possible with jQuery?
A more "secure" auth, this is a request/response hash challenge. Except JavaScript Crypto is Hopeless, so it only works over SSL and you still have to cache the username and password on the client side, making it more complicated than HTTP Basic Auth but no more secure.
Query Authentication with Additional Signature Parameters.
Another more "secure" auth, where you encrypt your parameters with nonce and timing data (to protect against repeat and timing attacks) and send the. One of the best examples of this is the OAuth 1.0 protocol, which is, as far as I know, a pretty stonking way to implement authentication on a REST server.
https://www.rfc-editor.org/rfc/rfc5849
Oh, but there aren't any OAuth 1.0 clients for JavaScript. Why?
JavaScript Crypto is Hopeless, remember. JavaScript can't participate in OAuth 1.0 without SSL, and you still have to store the client's username and password locally - which puts this in the same category as Digest Auth - it's more complicated than HTTP Basic Auth but it's no more secure.
Token
The user sends a username and password, and in exchange gets a token that can be used to authenticate requests.
This is marginally more secure than HTTP Basic Auth, because as soon as the username/password transaction is complete you can discard the sensitive data. It's also less RESTful, as tokens constitute "state" and make the server implementation more complicated.
SSL Still
The rub though, is that you still have to send that initial username and password to get a token. Sensitive information still touches your compromisable JavaScript.
To protect your user's credentials, you still need to keep attackers out of your JavaScript, and you still need to send a username and password over the wire. SSL Required.
Token Expiry
It's common to enforce token policies like "hey, when this token has been around too long, discard it and make the user authenticate again." or "I'm pretty sure that the only IP address allowed to use this token is XXX.XXX.XXX.XXX". Many of these policies are pretty good ideas.
Firesheeping
However, using a token Without SSL is still vulnerable to an attack called 'sidejacking': http://codebutler.github.io/firesheep/
The attacker doesn't get your user's credentials, but they can still pretend to be your user, which can be pretty bad.
tl;dr: Sending unencrypted tokens over the wire means that attackers can easily nab those tokens and pretend to be your user. FireSheep is a program that makes this very easy.
A Separate, More Secure Zone
The larger the application that you're running, the harder it is to absolutely ensure that they won't be able to inject some code that changes how you process sensitive data. Do you absolutely trust your CDN? Your advertisers? Your own code base?
Common for credit card details and less common for username and password - some implementers keep 'sensitive data entry' on a separate page from the rest of their application, a page that can be tightly controlled and locked down as best as possible, preferably one that is difficult to phish users with.
Cookie (just means Token)
It is possible (and common) to put the authentication token in a cookie. This doesn't change any of the properties of auth with the token, it's more of a convenience thing. All of the previous arguments still apply.
Session (still just means Token)
Session Auth is just Token authentication, but with a few differences that make it seem like a slightly different thing:
Users start with an unauthenticated token.
The backend maintains a 'state' object that is tied to a user's token.
The token is provided in a cookie.
The application environment abstracts the details away from you.
Aside from that, though, it's no different from Token Auth, really.
This wanders even further from a RESTful implementation - with state objects you're going further and further down the path of plain ol' RPC on a stateful server.
OAuth 2.0
OAuth 2.0 looks at the problem of "How does Software A give Software B access to User X's data without Software B having access to User X's login credentials."
The implementation is very much just a standard way for a user to get a token, and then for a third party service to go "yep, this user and this token match, and you can get some of their data from us now."
Fundamentally, though, OAuth 2.0 is just a token protocol. It exhibits the same properties as other token protocols - you still need SSL to protect those tokens - it just changes up how those tokens are generated.
There are two ways that OAuth 2.0 can help you:
Providing Authentication/Information to Others
Getting Authentication/Information from Others
But when it comes down to it, you're just... using tokens.
Back to your question
So, the question that you're asking is "should I store my token in a cookie and have my environment's automatic session management take care of the details, or should I store my token in Javascript and handle those details myself?"
And the answer is: do whatever makes you happy.
The thing about automatic session management, though, is that there's a lot of magic happening behind the scenes for you. Often it's nicer to be in control of those details yourself.
I am 21 so SSL is yes
The other answer is: Use https for everything or brigands will steal your users' passwords and tokens.
You can increase security in authentication process by using JWT (JSON Web Tokens) and SSL/HTTPS.
The Basic Auth / Session ID can be stolen via:
MITM attack (Man-In-The-Middle) - without SSL/HTTPS
An intruder gaining access to a user's computer
XSS
By using JWT you're encrypting the user's authentication details and storing in the client, and sending it along with every request to the API, where the server/API validates the token. It can't be decrypted/read without the private key (which the server/API stores secretly) Read update.
The new (more secure) flow would be:
Login
User logs in and sends login credentials to API (over SSL/HTTPS)
API receives login credentials
If valid:
Register a new session in the database Read update
Encrypt User ID, Session ID, IP address, timestamp, etc. in a JWT with a private key.
API sends the JWT token back to the client (over SSL/HTTPS)
Client receives the JWT token and stores in localStorage/cookie
Every request to API
User sends a HTTP request to API (over SSL/HTTPS) with the stored JWT token in the HTTP header
API reads HTTP header and decrypts JWT token with its private key
API validates the JWT token, matches the IP address from the HTTP request with the one in the JWT token and checks if session has expired
If valid:
Return response with requested content
If invalid:
Throw exception (403 / 401)
Flag intrusion in the system
Send a warning email to the user.
Updated 30.07.15:
JWT payload/claims can actually be read without the private key (secret) and it's not secure to store it in localStorage. I'm sorry about these false statements. However they seem to be working on a JWE standard (JSON Web Encryption).
I implemented this by storing claims (userID, exp) in a JWT, signed it with a private key (secret) the API/backend only knows about and stored it as a secure HttpOnly cookie on the client. That way it cannot be read via XSS and cannot be manipulated, otherwise the JWT fails signature verification. Also by using a secure HttpOnly cookie, you're making sure that the cookie is sent only via HTTP requests (not accessible to script) and only sent via secure connection (HTTPS).
Updated 17.07.16:
JWTs are by nature stateless. That means they invalidate/expire themselves. By adding the SessionID in the token's claims you're making it stateful, because its validity doesn't now only depend on signature verification and expiry date, it also depends on the session state on the server. However the upside is you can invalidate tokens/sessions easily, which you couldn't before with stateless JWTs.
I would go for the second, the token system.
Did you know about ember-auth or ember-simple-auth? They both use the token based system, like ember-simple-auth states:
A lightweight and unobtrusive library for implementing token based
authentication in Ember.js applications.
http://ember-simple-auth.simplabs.com
They have session management, and are easy to plug into existing projects too.
There is also an Ember App Kit example version of ember-simple-auth: Working example of ember-app-kit using ember-simple-auth for OAuth2 authentication.
The Setup
We’re building a PWA (progressive web app). The main components are the app shell (SPA) and the API. The REST API will supply the data needed for the app, while the SPA will handle the rest (as per Google recommendation).
The Problem
Authentication of the end-user seems problematic because the web browser needs to be accounted for. We want the user login to persist through closing down the browser.
We’ve done the research about the possible ways of going about it, however we’d like to ensure that we’re not going in the wrong direction.
Solutions we’ve considered
Session based authentication - the user sends username and password to /accounts/auth and receives a HTTP only cookie with the session ID. The session needs to be stored in a database or Redis. The issue with this option is that cookies are automatically sent by the browser therefore we need a CSRF protection in place. Using the Synchronizer Token Pattern a new token would be generated every time a state changing request has been made e.g. POST. This means that the application needs to supply a CSRF token with every request so that the PWA can send it via AJAX. We determined that it’s not ideal as the user can send multiple post requests in a quick succession making some of them fail and resulting in a bad user experience.
We could also use this method without the CSRF by limiting the CORS policy to same domain and adding a header requirement which technically should stop all CSRF, however we're unsure how secure it would be.
JWT token based authentication - the user sends username and password to /accounts/auth and a new JWT token is issued. The JWT then needs to be stored in localstorage or a cookie. Using localstorage means that JWT is XSS vulnerable and if the token is stolen, an attacker can impersonate the user completely. Using cookies we will still have a CSRF issue to resolve. We considered a double submit cookie method but the CSRF would only refresh every time the JWT is reissued which creates a window for the attacker to find out what the CSRF is. It is not clear which method is best to use.
Session based authentication + JWT token authentication - the user sends username and password to /accounts/auth, a session is created, a HTTP only cookie is set in the browser and a JWT token is sent back to the user. The PWA can authenticate requests with the JWT and whenever the JWT expires the app calls /accounts/auth again to acquire a new one. The /accounts/auth endpoint would still need to be CSRF protected, however the impact of it on usability would be minimised.
There seems to be a large amount of articles claiming that localStorage is insecure and shouldn't be used so why are high profile organisations like Amazon still recommending it? https://github.com/aws/amazon-cognito-auth-js - this SDK uses localStorage to store the token.
You don't need to generate new CSRF token each time a client make a request. It's much easier to use a scheme like token = hash(id + secret + current_day). You only need to update it once a day, or even employ mixed scheme (if the token is invalid today, but is okay for the previous day, the server accepts the operation and returns new token in a predefined header for client to renew it). You may also use the cookie as an id, making the token totally stateless and much easier to check, no need to store them in the database.
Here is how I look at it.
JWT token authentication : with this approach, you can always use a time-bound token with its expiration set to say 2 hours or something?
Or another approach would also be to try and see how you could use some of the approaches the Credentials Management API suggests for example, auto-sign-in of users whenever they come back.
Stuff like 2-step verification with OTPs for instance; for very important features in your web app can be a choice. In this case basic stuff are tied to whichever one time authentication method you have.
Actually, you can also use user-defined pins or short codes (seen a lot in banking apps) to grant access to some features in your web app.
Hope this helps, or sparks some ideation.
I'm planning an API and I haven't found any information about where the access token should be placed in the response. I'm interested in the case where I'm providing the access token. What is the best practice to send the token to the clients (from the backend service)?
Should it be sent to clients in the Header or in the Body of the answer?
You should look at examples and see what others do. For example, Facebook, Twitter, Google, Amazon... and all those popular PaaS services exposing a REST API all use OAuth2 as authentication mechanism:
In OAuth2 you'll see the specification requires to send the generated token to the client in the body of a json response: https://www.rfc-editor.org/rfc/rfc6749#section-5 . But you can also see how Google and other vendors extend this mechanism so it can be sent as a callback url param, for example (Check out https://developers.google.com/identity/protocols/OAuth2UserAgent).
Once you get the authorization token you put it on the Authorization: HTTP header you send on your requests for accessing protected resources. They have to support this way of doing it, because it is how the OAuth2 standard specifies it https://www.rfc-editor.org/rfc/rfc6749#section-7
If you want to tinker a little bit more with oauth, check out https://developers.google.com/oauthplayground
(OAuth is pretty much the same thing.)
They usually also extend the Authorization header mechanism to allow the token to be provided as a GET/POST parameter in the url or the body of the response, respectively (For example, Facebook's Graph API also supports passing an access_token= parameter on your HTTP POST request body or GET URI). There is no way to manipulate or even read HTTP headers on a javascript browser application (see the modern fetch API and other proposals on Accessing the web page's HTTP Headers in JavaScript), so providing this functionality makes life easier for many app developers.
Another popular authentication scheme is SOAP authentication. It doesn't support tokens but it supports digest authentication, which is a similar thing. The interesting part of it is that it is not HTTP/Web based (although it is primarily used that way), so you can use it over other application protocols. It's a little more cumbersome, but you can find ready to use implementations for both server and client.
You can also use digest authentication over HTTP without SOAP. It is also based on Authorization: headers and every browser supports it. Check out https://en.wikipedia.org/wiki/Digest_access_authentication to see how the authorization headers are formed in different ways depending on the level of security you want to reach.
Some services, like redmine, allow you to use an API token (API key) instead of a session token. Then you can make basic http auth on your requests like https://whatever:yourapikey#example.com/protectedMethod, although passing of auth data on URLs is currently deprecated in favor of basic auth header, and plain passwords / static API keys should only be sent over secured SSL connections. In this case the client can have the server generate an api key using a web interface or a rest api (so the generated key will be passed as a JSON response). This method may not be such a good idea, though: Check http://talks.codegram.com/http-authentication-methods#/intro if you want to know why, and also this question discussing where to put them: Where should I place API keys in REST API calls?
I was reading articles about token based authentication.
And when I was going through this article "Cookies Vs Tokens", I did not understand the following four points (I have asked my doubts below each of the quoted points)
Cross-domain / CORS: cookies + CORS don't play well across different
domains. A token-based approach allows you to make AJAX calls to any
server, on any domain because you use an HTTP header to transmit the
user information.
Why now there will not be cors problem ? What if someone one from the neighboring tab in the browser steals your token(because in javascript of that domain there is code which can clearly tell where the token is stored and how it is retrieved) and starts making
requests ?
And more over it is said "cookies + cors" . why is it said that cookie and cors don't play well ?
CDN: you can serve all the assets of your app from a CDN (e.g.
javascript, HTML, images, etc.), and your server side is just the API.
Why is this an advantage in token based auth systems ? We were using cdn's even when we were doing cookie based authentication right ?
This was because script tags can call scripts from other domains any way right ?
CSRF: since you are not relying on cookies, you don't need to protect
against cross site requests (e.g. it would not be possible to
your site, generate a POST request and re-use the existing
authentication cookie because there will be none).
What if someone one from the neighboring tab in the browser(from some other domain) steals your token and starts making
requests ? I cannot understand this either.
Mobile ready: when you start working on a native platform (iOS,
Android, Windows 8, etc.) cookies are not ideal when consuming a
secure API (you have to deal with cookie containers). Adopting a
token-based approach simplifies this a lot.
Did not understand it completely . Can someone explain , how?
The linked article is a bit sloppily written and does not properly define and distinguish between the various concepts.
First, you have to define what exactly you mean by "token based". The article does not do that, but it seems it has OAuth 2.0 JWT access tokens with the Bearer Token Usage (http://self-issued.info/docs/draft-ietf-oauth-v2-bearer.html) in mind. This means that:
Bearer access tokens are used. The upside of bearer tokens (as opposed to MAC token usage) is simplicity of usage -- all a client needs to pass to a server is a token. The downside is that whoever gets hold of such a token is considered its rightful owner, so he can do whatever he wants with it.
Tokens are passed in the Authorization HTTP header, following a specific syntax.
Then you have to define what you really mean by "cookie based" authentication -- which again, the article does not do. A cookie, essentially, is just a transport mechanism to pass data back and forth between client and servers of a specific domain, without the client having to do anything explicitly. So you can really stuff pretty much anything into a cookie -- even OAuth 2 access tokens.
What the author seems to have in mind is server side-managed sessions where a session id is passed in a cookie.
With these definitions in mind, it is possible to answer your questions:
Cross-domain / CORS: cookies + CORS don't play well across different domains. A token-based approach allows you to make AJAX calls to any server, on any domain because you use an HTTP header to transmit the user information.
This is true, but for different reasons: Cookies are scoped to a specific domain, so as long as you use cookies to pass a session id between clients and servers, you are bound to having to put all servers under one domain. This looks like a limitation, but the true limitation is somewhere else -- when using server side-managed sessions, a session id will only make sense to those servers who have access to the server side session management. Any server outside the domain most likely cannot make any sense of a session id.
A JWT access token, however, is self contained and can be interpreted by other parties as well. This is why access tokens are well suited for cross-domain/cross-service calls.
CDN: you can serve all the assets of your app from a CDN (e.g. javascript, HTML, images, etc.), and your server side is just the API.
Yes. But this does not really matter in practice because anyting you put into a CDN will most likely not require authentication in the first place -- so this is kind of a silly argument.
CSRF: since you are not relying on cookies, you don't need to protect against cross site requests (e.g. it would not be possible to your site, generate a POST request and re-use the existing authentication cookie because there will be none).
The danger of cookies w.r.t. CSRF is that they are passed implicitly to the server -- the browser adds a cookie to the request, regardless of whether it was good of rogue code that triggered the request.
When tokens are passed in a cookie, this danger applies to tokens we well. However, when they are passed in a header (as the Bearer Token Usage spec defines), this danger does not exist as the header needs to be added to teh request explicitly, every single time. This makes this approach more CSRF-safe than cookies.
However, when the rogue code can get hold of the token some other way (because it was passed as a URL parameter, for example (which the spec allows), it is still possible to make CSRF attacks with tokens.
Mobile ready: when you start working on a native platform (iOS, Android, Windows 8, etc.) cookies are not ideal when consuming a secure API (you have to deal with cookie containers). Adopting a token-based approach simplifies this a lot.
Dealing with cookie containers is usually easier than setting up OAuth, so I do not buy into this one.
Alright, I'll try to go through these in as structured a way as possible:
CORS + Cookies: Cookies were never meant to travel across domains. A cookie must be submitted to the server from its origin, cross-origin cookies don't exist. There is no need to worry about another application accessing the token if it's stored in localstorage since that one is site-specific, e.g every site gets it own allotted space of localstorage.
I'm not certain about this one. I'd assume they mean an authenticated CDN, e.g you provide a token to get access to the resources
Again, the token won't be stolen by a neighbour, because of how localstorage works. If you're worried you could actually issue a new token from the server with every response, effectively making one-time use tokens. That way, even if your token is compromised, it is far less likely to work.
Cookies are fairly complex structures, and if you one day decide you want to consume your API on a mobile device, it will be much easier to use a token, which is just a string (compared to a structured object)
The most important difference is that tokens are stateless, while cookies are not. Cookies will contain information about the authenticated user and context-specific data. A token, on the other hand, is just a string that an API consumer assumes can be used to access restricted data.
I'm having some trouble deciding how to implement authentication for a RESTful API that will be secure for consumption by both a web app and a mobile app.
Firstly, I thought to investigate HTTP Basic Authentication over HTTPS as an option. It would work well for a mobile app, where the username and password could be stored in the OS keychain securely and couldn't be intercepted in transit since the request would be over HTTPS. It's also elegant for the API since it'll be completely stateless. The problem with this is for the web app. There won't be access to such a keychain for storing the username and password, so I would need to use a cookie or localStorage, but then I'm storing the user's private details in a readily accessible place.
After more research, I found a lot of talk about HMAC authentication. The problem I see with this approach is there needs to be a shared secret that only the client and server knows. How can I get this per-user secret to a particular user in the web app, unless I have an api/login endpoint which takes username/password and gives the secret back to store in a cookie? to use in future requests. This is introducing state to the API however.
To throw another spanner into the works, I'd like to be able to restrict the API to certain applications (or, to be able to block certain apps from using the API). I can't see how this would be possible with the web app being completely public.
I don't really want to implement OAuth. It's probably overkill for my needs.
I feel as though I might not be understanding HMAC fully, so I'd welcome an explanation and how I could implement it securely with a web app and a mobile app.
Update
I ended up using HTTP Basic Auth, however instead of providing the actual username and password every request, an endpoint was implemented to exchange the username and password for an access key which is then provided for every authenticated request. Eliminates the problem of storing the username and password in the browser, but of course you could still fish out the token if you had access to the machine and use it. In hindsight, I would probably have looked at OAuth further, but it's pretty complicated for beginners.
You should use OAuth2. Here is how:
1) Mobile App
The mobile app store client credentials as you state yourself. It then uses "Resource Owner Password Credentials Grant" (see https://www.rfc-editor.org/rfc/rfc6749#section-4.3) to send those credentials. In turn it gets a (bearer) token it can use in the following requests.
2) Web site
The website uses "Authorization Code Grant" (see https://www.rfc-editor.org/rfc/rfc6749#section-4.1):
Website sees unauthorized request and redirects browser to HTML-enabled autorization endpoint in the REST api.
User authenticates with REST service
REST site redirects user back to website with access token in URL.
Website calls REST site and swaps access token to authorization token.
Here after the website uses the authorization token for accessing the REST service (on behalf of the end-user) - usually by including the token as a "bearer" token in the HTTP Authorization header.
It is not rocket science but it does take some time to understand completely.
3) Restricting API access for certain applications
In OAuth2 each client is issued a client ID and client secret (here "client" is your mobile app or website). The client must send these credentials when authorizing. Your REST service can use this to validate the calling client
I resolved this for my own API quite easily and securely without the need to expose any client credentials.
I also split the problem into 2 parts. API authentication - is this a valid request from a recognised entity (website or native app). API authorisation, is that entity allowed to use this particular endpoint and HTTP verb.
Authorisation is coded into the API using an access control list and user permissions and settings that are set up within the API code, configuration and database as required. A simple if statement in the API can test for authorisation and return the appropriate response (not authorised or the results of processing the API call).
Authentication is now just about checking to see if the call is genuine. To do this I issue self signed certificates to clients. A call to the API is made from their server whenever they want - typically when they generate their first page (or when they are performing their own app login checks). This call uses the certificates I have previously provided. If on my side I am happy the certificate is valid I can return a nonce and a time limited generated API key. This key is used in all subsequent calls to other API endpoints, in the bearer header for example, and it can be stored quite openly in an HTML form field or javascript variable or a variable within an app.
The nonce will prevent replay attacks and the API key can be stolen if someone wants - they will not be able to continue using after it expires or if the nonce changes before they make the next call.
Each API response will contain the next nonce of if the nonce doesn't match it will return an authentication error. In fact of the nonce doesn't match I kill the API key too. This will then force a genuine API user to reauthenticate using the certificates.
As long as the end user keeps those certificates safe and doesn't expose the method they use to make the initial authentication call (like making it an ajax request that can be replayed) then the API's are nice and secure.
One way of addressing the issue of user authentication to the API is by requesting an authentication token from the API when the user logs in. This token can then be used for subsequent requests. You've already touched on this approach - it's pretty sound.
With respect to restricting certain web apps. You'll want to have each web app identify itself with each request and have this authentication carried out inside your API implementation. Pretty straight forward.