Custom password hashing function for Cerberus SFTP server - encryption

I'm trying to migrate users from our legacy Cerberus SFTP server to a new AWS Transfer Family SFTP server. The problem is most of our Cerberus users have password based authentication and all I have access to is their one-way hashed password. Thus I'm trying to reverse engineer how Cerberus hashes it's password so I don't have to ask our 100+ customers to submit a new password to use or switch to a public key based authentication.
I came across this blog post that I think details how to do it but I can't seem to get it to work for some reason - https://support.cerberusftp.com/hc/en-us/articles/360000040039-Securely-Storing-User-Passwords.
Here are the steps I've taken so far -
created a user in Cerberus with a password of "asdf"
exported my collection of users to a CSV file
identified the hashed password from the export as follows -
{{PBKDF2 HMAC SHA256}}:5000:42ED67592D7D80F03BF3E2413EB80718C5DAFEB5237FC4E5E309C2940DF1DBB2A4ABD9BB63B8AD285858B532A573D9DE
attempted to write a Python script that could hash "asdf" to the same hash as shown above
Here is what my script looks like so far - https://replit.com/#ryangrush/sample.
import hashlib
import base64
password = b'asdf'
salt = b'sample_salt'
combined = salt + password
first = hashlib.pbkdf2_hmac('sha256', combined, b'', 5000)
combined = salt + first
second = hashlib.pbkdf2_hmac('sha256', combined, b'', 5000)
base_16 = base64.b16encode(second)
print(second.hex())
print(base_16)

The documentation must've been written before the v7.0 PBKDF2 HMAC functions were adopted. The salt and the password are now used just as described in the documentation for PBKDF2.
import hashlib
import base64
hashed_password_entry = '{{PBKDF2 HMAC SHA256}}:5000:42ED67592D7D80F03BF3E2413EB80718C5DAFEB5237FC4E5E309C2940DF1DBB2A4ABD9BB63B8AD285858B532A573D9DE'
entry_components_strings = hashed_password_entry.split(':')
password = b'asdf'
iterations = int(entry_components_strings[1])
salt_plus_hashvalue = base64.b16decode(entry_components_strings[2])
hash_len = 256 // 8
salt, hashvalue = salt_plus_hashvalue[:-hash_len], salt_plus_hashvalue[-hash_len:]
hashvalue_test = hashlib.pbkdf2_hmac('SHA256', password, salt, iterations)
print(hashvalue_test.hex())
the output is c5dafeb5237fc4e5e309c2940df1dbb2a4abd9bb63b8ad285858b532a573d9de which you can see matches the hashed value that is at the end of entry.

Related

Glassfish Change Admin Password

How can I change the admin password for a Glassfish Domain using a password file? I know the conventional method of manually typing the password upon prompt.
However I want to change the admin password using a script where in I do not have to manually type the password.
This is possible, but you will need 2 password files if you want to script this fully in the easiest way.
Create a temporary file (tmpfile in my example) which will hold the current password (blank by default) and the desired new password:
AS_ADMIN_PASSWORD=
AS_ADMIN_NEWPASSWORD=myNewPassword
Now create a password (pwdfile in my example) file which will contain the changed admin password:
AS_ADMIN_PASSWORD=myNewPassword
You can then use the files to change the password using the commands below, making sure to use tmpfile when changing the password, then pwdfile afterwards
$PAYARA_PATH/bin/asadmin start-domain
$PAYARA_PATH/bin/asadmin --user $ADMIN_USER --passwordfile=/opt/tmpfile change-admin-password
$PAYARA_PATH/bin/asadmin --user $ADMIN_USER --passwordfile=/opt/pwdfile enable-secure-admin
$PAYARA_PATH/bin/asadmin restart-domain
This example was adapted from the way the Payara Server dockerfile works
For anyone still interested in manually setting the admin account password:
I tried to generate the contents of the "admin-keyfile" located in "glassfish/domains/{ACTIVE_DOMAIN_NAME}/config/admin-keyfile" based on the current implementation of the Payara Repo. This file (as the data source for the FileRealm) is used to authenticate the admin user when accessing the admin interface under port 4848.
Each line of this text file represents an account and is structured as
USERNAME;PASSWORD;GROUPS
The field "PASSWORD" is prefixed with a hash algorithm keyword (wrapped in curly braces, e.g. "SSHA" or "SSHA256") followed by a BASE64 encoded hash of the concatenated salted hash and the salt value itself (some random bytes):
{SSHA}BASE64(SHA(password,salt),salt)
Long story short: If you want to generate user accounts manually you could for example use the following Python script:
import hashlib
from base64 import b64encode
from secrets import token_bytes
from getpass import getpass
username = 'admin' # input('Username: ')
plainTextPassword = getpass()
randomSalt = token_bytes(8)
passwordHash = hashlib.sha256()
passwordHash.update(plainTextPassword.encode('utf-8'))
passwordHash.update(randomSalt)
passwordDigest = passwordHash.digest()
# cryptic range reflects the strange implementation... feel free to change it to "range(98)"
# https://github.com/payara/Payara/blob/6488cbdc90fd0f6c42de6a42affcd09f697be715/nucleus/common/common-util/src/main/java/org/glassfish/security/common/SSHA.java#L108
for run in range(2, 101):
passwordHash = hashlib.sha256()
passwordHash.update(passwordDigest)
passwordDigest = passwordHash.digest()
saltedHashAndSalt = b64encode(passwordDigest + randomSalt).decode('utf-8')
result = '{0};{{SSHA256}}{1};asadmin'.format(username, saltedHashAndSalt)
print(result)
Insert the console output into the "admin-keyfile" and (re)start your server.
As far as I know, it is impossible to change it via a file as a parameter for security reasons.
You can consider an alternative solution (pipe) but the confirmation of the password is always necessary. https://docs.oracle.com/cd/E19798-01/821-1758/change-admin-password-1/index.html

Can I save passwords securely and retrieve them without asking for a master password?

I'm trying to understand how to save passwords used on my app securely, so the user doesn't have to remember them but at the same time nobody can get them looking at the data from my app.
I imagine the files containing the passwords should be encrypted, my doub is whether the user would have to input a "master password" to retrieve stored passwords or if there is any way so that only my app can retrieve them without any input from the user.
Can my app retrieve the passwords without the user needing to write a master password? How is this done?
On Windows your best solution is to use the Data Protection API, used by Chrome, IE, Remote Desktop Connection, and dozens of other technologies, to encrypt data.
The virtue is that the data is encrypted (in a round-about way) with the user's own Windows password. When the user types their password into Windows, it makes all the "protected" data available.
Features:
the data is encrypted
the user doesn't have to enter their password to encrypt data
only the user can ever decrypt it
the user does not have to enter their password to decrypt data
Sample pseudo-code
The API you want is CryptProtectData and CryptUnprotectData:
public bytes[] ProtectBytes(bytes[] plaintext)
{
DATA_BLOB dataIn;
dataIn.cbData = plaintext.Length;
dataIn.pbData = Addr(plaintext[0]);
DATA_BLOB dataOut;
BOOL bRes = CryptProtectData(
dataIn,
null, //data description (optional PWideChar)
null, //optional entropy (PDATA_BLOB)
null, //reserved
null, //prompt struct
CRYPTPROTECT_UI_FORBIDDEN,
ref dataOut);
if (!bRes) then
{
DWORD le = GetLastError();
throw new Win32Error(le, "Error calling CryptProtectData");
}
//Copy ciphertext from dataOut blob into an actual array
bytes[] result;
SetLength(result, dataOut.cbData);
CopyMemory(dataOut.pbData, Addr(result[0]), dataOut.cbData);
//When you have finished using the DATA_BLOB structure, free its pbData member by calling the LocalFree function
LocalFree(HANDLE(dataOut.pbData)); //LocalFree takes a handle, not a pointer. But that's what the SDK says.
}
Later, when you need to decrypt the blob, you use CryptProtectData.
The data is (effectively) encrypted with the user's Windows password; and only the person with their Windows password can decrypt it.
Note: Any code released into public domain. No attribution required.

Storing user password in Keystore

I am working on an application in which I would like to create a
keystore with a master password and store user passwords after
hashing inside it. Is it possible to do so?
I am working in Java, so I would appreciate Java based solutions. I
am also checking how to create keystore in Java code itself.
Currently I have :
KeyStore ks = KeyStore.getInstance(KeyStore.getDefaultType());
char[] password = "some password".toCharArray();
ks.load(null, password);
FileOutputStream fos = new FileOutputStream("newKeyStoreFileName");
ks.store(fos, password);
fos.close();
I don't know if it is possible to store Strings in keystore, if not then I am looking to someway store those hashed passwords but encrypted again with a master-password. Any help would be nice. Thanks a lot. :-)
No. KeyStores can contains keys and certificates, nothing else. What you're looking for is LDAP or a database with column-hashing features.

wordpress password encryption in NodeJS

How to do user authentication from wordpress database in nodeJS.
I need to validate user if username/password is correct, using wordpress database. Wordpress is using PHPass PHP library to encrypt passwords. But I need to match password in NodeJS.
Edit: Today there is an implementation which supports Wordpress portable hashes: wordpress-hash-node.
Previous reply:
Sigh... I took an interest in this, and spent half an hour pouring through PHPass source code. Then I googled for node phpass.
Edit: On closer inspection, this seems to only implement bcrypt encryption, while the PHPass default (which I think Wordpress uses) is something they call "Portable Hashes". node-phpass throws 'Portable hashes are not implemented' when you ask for Portable Hashes. I suggest you implement that for node-phpass and send a pull request.
For Wordpress 4.9.5, in NodeJS after
npm i wordpress-hash-node
var hasher = require('wordpress-hash-node');
let wordpressHashPass = "$P$BzPE3JGpq4CUpvpMHhtPh3lZmIoG.s1";
let wordpressPlainTextPass = '(&#fZsImcKq7K3Lmd&qBe!Jx';
var checked = hasher.CheckPassword(wordpressPlainTextPass, wordpressHashPass); //This will return true
console.log(checked); // returns true
var hasher = require('wordpress-hash-node');
let wordpressHashPass = "$P$BzPE3JGpq4CUpvpMHhtPh3lZmIoG.s1";
let wordpressPlainTextPass = 'goodday';
var checked = hasher.CheckPassword(wordpressPlainTextPass, wordpressHashPass); //This will return false
console.log(checked); // returns false
wordpressHashPass is the MD5 hashed password that you can find in the wp_users table of Wordpress for a user.
wordpressPlainTextPass is the plain text password that the user types in the password field.
The method CheckPassword compares the plain text password and the hash password. It returns true if it coincides and false if it does not coincides.

advantage of md5+salt in asp.net application

I have a question about how to use md5 and a salt to secure a password, I have already made many searches for answers to my questions.
An article I saw was using c# to convert password to md5 string, something like this:
public static string md5(string <b>sPassword</b>)
{
System.Security.Cryptography.MD5CryptoServiceProvider x = new System.Security.Cryptography.MD5CryptoServiceProvider();
byte[] bs = System.Text.Encoding.UTF8.GetBytes(sPassword);
bs = x.ComputeHash(bs);
System.Text.StringBuilder s = new System.Text.StringBuilder();
foreach (byte b in bs)
{
s.Append(b.ToString("x2").ToLower());
}
return s.ToString();
}
}
My question is the above code seems server side is its mean password traveling over internet in plain text doesn't it create any security issue or it does not matter i don't know (may be i am getting it wrong way or i am not clear about password security concept) ?
What i have done in my project is i am creating a secure password at client side with java script md5.js file and with user's entered password before posting login.aspx form back to server then at server side i am fetching hashed password of user from database(which was stored at the time of registration of user with same technique) and match both client side and server side hashed passwords if they match user authenticated.
i don't know weather i am doing it right way or not please let me know right way if i am wrong .
Now the problem is i want to use SALT with the md5 (md5+salt) to make password more secure with Randomly generated salt string. how to do this should i make a random salt string at server side while page_load of login page and then send it to client side and at client mix this salt with user password before posting form. after post again mix the password(fetched from database) with same random string and match both password to authenticate.
One more question, at the time of registration of a new user, where should originally user entered password convert in md5 at client side or server side if at server side then password should post to server as it is means original password.(like "MyPassword")
Firstly you should be aware that SHA1 is now industry standard, but it's still fine to use Md5 for most things.
Secondly to stop plain text transmitting over the public network, use an HTTPS connection (you may need to purchase a certificate from a recognised vendor).
Also if this is for a user system, consider using ASP.net's membership system. It does this all for you and has been extensively reviewed.
The basic flow of what you describe anyway would be:
User enters password
Server generates random salt
Hashed password = md5(salt + raw password)
Store hashed password and salt along side username, dispose of raw
When user logs in, find the associated salt with the username login is being attempted for.
Is password valid = does md5(salt + entered password) = store hash?
If they do, login
Once they have logged in, it might be a good idea also to regenerate a new salt and hash. Also the md5() should be applied to the password thousands of times before storing to make a dictionary attack uneconomical.
There are plenty of resources out there that go into this in more detail.
Good luck!

Resources