I am currently writing an HTML5 web app with a Sails.js (node framework) backend. Right now, most of my APIs are secured against the user authentication system I'm using with PassportJS. Unauthorized users trying to use my APIs will get a 401 error.
However, there's one hole in the system, which is the sign up API itself. I obviously can't secure my sign up API with user authentication (because the user wouldn't have had an account to sign in with yet), therefore anyone could easily spam the API with many fake accounts. On my sign up page, I have a small verification question on the lines of "What is 2+2?" (it is generated randomly) and it is checked on the client and if the answer is correct, the client sends a request to my sign up API route with all the necessary parameters like name, birthday and username. How can I secure this API to ensure that people must go through my sign up page, and cannot simply bypass this security measure and call the API directly?
Just as a note, my APIs are not RESTFul.
There are multiple possible ways.
First of all, you may consider adding a rate-limit to IP addresses. This is NOT 100% effective, but will certainly slow down some spam attempts. For example, you can limit the number of accounts created by the same IP address to 5 every 5 minutes.
Secondly, if you want to use some sort of captcha, consider reCAPTCHA. Among the different captcha services, this is particularly effective against bots, as they especially use words that fail OCR recognition.
Eventually, to make sure that people actually visit your signup page before calling the API, you can use a "security token". This is the same technique that is used for example to protect against CSRF (Cross-Site Request Forgery) attacks.
When the server generates the signup page, it also passes to the client an hidden field (for example "token") that contains a uniquely-generated value. The client will submit this value along with the form back to the API server when it requests the creation of a user, and the server uses the token to validate the request.
There are basically two approaches to generate these tokens.
First method
The signup page creates a random string/number and stores it in the database to be used as token. When the user submits the form, the server searches for that token into the database: if it's present, then the submission is valid; otherwise it fails. The token is then removed from the database.
Additional security can be obtained by storing into the database, along with the token, an expiration date and the client's user-agent (unlike IP's, user-agents are unlikely to change during the same session).
Pros: each token can be used only once.
Cons: the app needs a database, and it will be queried 3 times just for the token insertion, validation and deletion (requiring time and adding load to the database). You also should to purge regularly expired tokens from the database.
Second method
The signup page creates a token by digitally signing a plain-text string containing all the validation information. For example, suppose that you want to create a token that expires on 1411660627 (UNIX timestamp) and it's associated with the user-agent "Mozilla/5.0 ...". The server also possesses a secret salt (for example "123456abcde") that needs to be unique for the application and kept secret.
The signup page generates the token in a way similar to:
Create a plain-text string to be signed, by concatenating all the information. For example, if the expiration is 1411660627 and the MD5-hashed user agent is 0f7aee3e0a65ff9440d2a0183b4b1f49, your base signature would be something similar to: 1411660627-0f7aee3e0a65ff9440d2a0183b4b1f49.
Append the secret salt to that string: 1411660627-0f7aee3e0a65ff9440d2a0183b4b1f49_ 123456abcde.
Hash that string, using any hashing algorithm (for example MD5). The result is your signature: 0742d84065cb9497c1ba4c1d33190a93.
Concatenate your signature to the plain-text string to obtain your security token: 1411660627-0f7aee3e0a65ff9440d2a0183b4b1f49-0742d84065cb9497c1ba4c1d33190a93. This is what the user received and has to submit back.
To verify the token, then, a similar operation is done. When the server receives the token 1411660627-0f7aee3e0a65ff9440d2a0183b4b1f49-0742d84065cb9497c1ba4c1d33190a93, it performs these steps:
Extract the expiration from the token and check if it's still valid. If 1411660627 is smaller than the current timestamp, then it's still valid.
Compute the hash of the user agent and check if it matches the one on the token: 0f7aee3e0a65ff9440d2a0183b4b1f49.
Re-generate the signature as before: expiration-useragent_secretsalt using the data from the security token received from the user. In our example: 1411660627-0f7aee3e0a65ff9440d2a0183b4b1f49_ 123456abcde.
Compute the hash of the string as before. If it matches the third parameter from the token (0742d84065cb9497c1ba4c1d33190a93), then the security token is valid.
Pros: this solution does not require a database, but it's equally safe (as long as the salt is kept secret into the server).
Cons: the same security token can be used more than once until it expires.
It sounds like you're not making full use of your math-problem countermeasure--you're using it as a client-side barrier, but as you observed, a bot could just skip the client side and call your API. To make it more effective, instead of generating the question randomly on the client, you would generate it on the server, save the answer in the session (req.session.mathProblemAnswer = 4) and then send the user's answer with the API call so that it can be checked against the answer in the session.
Using a token as #Qualcuno describes in his answer will be effective against bots spamming your API endpoint directly, but there are bots that are smart enough to load your signup page, scan for hidden fields and submit the form (including the token). It's still a good idea to use CSRF protection though in general, which Sails has built-in support for.
Related
I have a REST API (.net) which for the initial login requires the password to be encrypted (RSA with OAEP). The reason was to obscure the users passwords from local logs.
Performing this encryption with javascript is complicated and I would need to let the client know the public key. The end user would be able to reverse engineer the encryption method then could use it to brute-force (or worse) access. So why bother, right (it is SSL Secure).
Thing is, I still need that password to be encrypted. Should I have some sort of encryption service at the server side that gives me the password to throw at the token endpoint? Why not have the service just log in for me then return the token. How should I proceed?
Thanks ^_^
This seems like a general authentication question. You can solve it like you would solve user authentication. Think of it this way:
When a user signs-in into your app, they provide their data on the client, and then it is validated on the server. In order for them to stay logged in, they get some sort of token, either via a Cookie session, JWT or whatever. Which is then saved on the client and sent on each request to the server in order to verify they are authenticated.
The above is how websites can show "registered users only" content. By validating a previously given token on each new request.
Now, applying this method to your REST Api. A user needs to request a token (which should not be your master password, but a uniquely generated one, in a per-user basis), and then save it locally for X amount of time. Every time the user makes a request to the API they send that token, which is validated.
This is also how normal APIs do it. You will need a token or some sort either way. If it's really sensitive information you're showing, the token should update every now and then (from minutes to days depending on how sensitive). You keep a record of valid tokens in your server. That way, if any token is "stolen", then it will only be valid for a small amount of time.
I'm developing a hybrid app where the user has the possibility to click "remember me" when logging in with username and password. In case user has only 1 "stored" account it automatically logs him in, but in case he has more than 1 "stored" account, the app shows him the list of the available accounts (like the one when logging into Gmail).
To implement the above behaviour, I have come up with this procedure:
At the first login the username and password are sent to server via HTTPS
If the credentials are correct, the server generates a token with such procedure:
merge username and password hash into a string
hash the string again with SHA and a server secret
substitute the chars in the string
create a N-char string (token) from the string
This token is then sent back to the device and the username and this token are stored to LocalStorage
From now on the user logs in with the username and this token (automatically or when clicking the account he wants to login into)
Would this be secure enough or should I improve something? I'm a bit worried though about storing usernames into LS, but that's the only information I have when showing the user what account he's logging into.
Edit: There can be several different people (for instance family members) logged in the account, because the app controls a device.
For the part about generating tokens you can look into something called JWT. As said on the page JWT is a "method for representing claims securely between two parties", which means you can use it to verify that the user using your page is in fact who he states to be. For the other parts, what you came up with is a preety standard strategy (user signs in, gets token, uses this token to use the app without needing to sign in again).
Simple explaination about JWT since you had a lot of questions:
JWT consists of three parts Header, Payload and Signature. Header and Payload are public (ie. user having the token can read them, they are only Base64 encoded), so don't store secret data inside them (althrough username and password hashed with salt should be fine). When you generate jwt, server calculates hash of header+payload+secret (secret known only to server) and puts it in the signature. Then when user tries to authenticate the signature must match with the data (since server again hashes header+payload+secret and compares it with signature) and only then it is accepted by server. This way without knowing the secret user can't change the data by himself.
JWT also implement "out of the box" one additional feature you might be interested in - expiration time. This way you can automatically logout users if they haven't used the page for certain periods of time. As to refreshing tokens there are a couple of ways and you need to deicide yourself whats the right way for you, Link
I am using ember to write a web ui for a site that requires user to log in. Suppose the browser has stored some cookie from last login of a user. Now the user visits the site again. So, is it a secure and common way for ember to log the user in automatically based on the cookie from the last visit? If so, what are the common ways to implement this? (I can't find anything from Google.) Furthermore, how do I create the cookie upon login? Is it a common way to just put a user id, password hash, and expiration in the cookie?
Additionally, any references related to this subject are greatly appreciated.
Edit 1
In light of Vohuman's answer, I think I can make my question a little more specific. Basically, what I want to know is a common and secure implementation to keep a user logged in, even when they close and reopen the browser. Namely, the life time is beyond the session scope. Take linkedin for example. If you are logged in and exit the browser. Then next time you revisit linkedin, you are still logged in automatically. Right now, what I can picture is a solution like the following.
When you first log in to the site, the server will return a cookie which includes an authentication hash token. Then next time when you revisit the site, the server will receive the hash token and thus authenticate your session.
So, is above flow basically what people usually do to keep a user logged in? If so, is the JSON Web Token (JWT) basically one way to construct the hash token I mentioned above? Additionally, assuming the connection is HTTPS, this approach seems secure to me. Is it not?
Edit 2
This article gives an interesting discussion regarding where to store the access token.
is it a secure and common way for ember to log the user in automatically based on the cookie from the last visit?
Yes and no. Security is a complex topic. Usually session cookies are used for authorizing users. This is actually the most used method of keeping the users logged in. If the user can't keep his credentials secure then any layers of security can be vulnerable.
For Single-page applications usually access tokens are used instead of cookies and sessions. The client sends the user credentials and server returns an access token. The token is encrypted and expirable and can be stored in localStorage or sessionStorage. Using JSON Web Tokens (JWT) standard is a popular method for implementing user authentication and authorization in web services. As an example, the Facebook Open Graph API uses access tokens.
JSON Web Token (JWT) is a compact, URL-safe means of representing
claims to be transferred between two parties. The claims in a JWT
are encoded as a JSON object that is used as the payload of a JSON
Web Signature (JWS) structure or as the plaintext of a JSON Web
Encryption (JWE) structure, enabling the claims to be digitally
signed or integrity protected with a Message Authentication Code
(MAC) and/or encrypted.
edit:
So, is above flow basically what people usually do to keep a user logged in?
For traditional websites, yes.
The whole point of using access tokens is keeping the web service/API stateless. This means that server doesn't have to store any cookies/sessions for authenticating and authorizing users. The stateless is one of the key factors of implementing web services that follow the REST paradigm. It's client that has to store the token and send it to the server (via the Authorization header or query parameters). The server doesn't store the token. Of course, you can store the tokens on the server if you want to add another layer of security, but it's not so common and not necessary. Storing the tokens on the server can also make your application vulnerable to database attacks and is not recommended.
If you want to make the process more secure you can decrease the validity time of access tokens (1 hour, 1 day or 1 week, it's up to you).
As for localStorage, is it secure?
localStorage data are stored separately for each origin (domain). A malicious user can only read the data if he/she has access to the user browser. You should make sure that your app doesn't have any XSS vulnerabilities so malicious users can't inject any scripts to your application. This is actually a different topic.
So here is my scenario , I generated a jwt token and stored that token in redis with 1 hour TTL.
Now I see most of tutorials use jwt.verify to verify the token..
I know they are verifying the token is authentic or not
Why I need to use jwt.verify.. Why can't I use redis.exists to check the token is authentic or not..
Most of them say, we can use jwt main feature is no need to use db to check the user and expiration..
But in my scenario I cant store everything in token.. So I am using redis to store the token with session information.
Questions are
1. So I should not use jwt for this kind of scenario.
2. Can I skip jwt.verify?
I am a node newbie..
JWTs can help you quickly retrieve information about the caller, without hitting a database (redis is also a database).
When using JWTs used by client applications/external services you must always verify them to make sure that you are the one that generated them and they have not been tampered with.
Common info stored in the JWT are things like username, real name, group etc. In your scenario, you could use the JWT to store a redis key that holds the info that you want. It might be the case that you always will hit redis to get the info you want, so JWTs don't add a lot of value to your case, but it might be so that you could use JWTs to write smarter code that will only hit redis under certain circumstances eg. if the user has this right, or if we have stored something in redis about this user or not (missing redis key from the JWT token)
You are the only one that can evaluate your scenario and the usefulness of JWTs but don't be hasty to dismiss them, as they provide a nice perfomance/security improvement out of the box.
Assuming that the Redis server is secured and you generate the JWT yourself (as seems to be the case here), you don't need to verify it. Once created, stored it in the cache and retrieve it later you don't need to verify it again because you know it could not have been tampered with in the Redis cache.
Only when receiving JWTs that are generated by 3rd parties you would need to verify that they are authentic.
If on the other hand you are distributing JWTs to 3rd-party applications and clients that you don't control then you will have to make sure that once you they are replayed back to you, they are untampered with by verifying the signature (or do a binary compare against the one stored in the Redis cache) and (when in use) checking the expiry timestamp in the exp claim.
Without any verification in place, it will be possible for a 3rd party to send requests to your API and in most cases the requests will likely turn into a man-in-the-middle attack. It's good security practice to keep a record of all the tokens generated on the server and then authenticate against them with each incoming request.
I'm trying to find a way to interact securely with an OAuth (draft v2-23) API completely in the browser. There are obvious security concerns with authenticating in the browser because the auth token and other keys are exposed somewhere in the browser.
Is there a secure way to hide those keys, or am I limited to creating a server side solution that keeps track of those keys?
I came across this very concern when working with the mobile version of my application (since a normal login auth wouldn't work.)
The solution I came up with is to work with 2 keys. When the user logs in, they are initially given their token. The client already has the token saved as a variable in the app (you can make the token whatever string you want, I usually generate a base64 key of 32 length.)
So this is the basic play by play of what happens when they call the server from a GET/PUT/etc..
Send request with USER TOKEN and SERVER TOKEN as parameters.
SERVER TOKEN is checked against the server token (which is salted and hashed). (If SERVER TOKEN is good, proceed to step 3, else deny request here.
USER TOKEN is checked against the user token assigned to said user (which is salted and hashed) and makes sure it exists. If good, proceed to step 4, else deny request here.
Execute request (GET/PUT/DEL/POST etc.)
Long story short, i keep a salted and hashed version of the user tokens (which are unique to the user) and the server tokens (which is the same for every user) on the server and the de-salting and de-hashing is done on the back end.
I hope this kinda makes sense. It might not be the standard or completely right but it's what i understood token auth to be and this works for me.
If others see something wrong with my reasoning please chime in and tell me so.