$ sudo docker pull blabla1337/owasp-skf-lab:sqli
$ sudo docker run -ti -p 127.0.0.1:5000:5000 blabla1337/owasp-skf-lab:sqli
{% hint style="success" %} Now that the app is running let's go hacking! {% endhint %}
First, make sure python3 and pip are installed on your host machine. After installation, we go to the folder of the lab we want to practise "i.e /skf-labs/XSS/, /skf-labs/jwt-secret/ " and run the following commands:
$ pip3 install -r requirements.txt
$ python3 <labname>
{% hint style="success" %} Now that the app is running let's go hacking! {% endhint %}
The first step is to identify parameters which could be potentially used in an SQL query to communicate with the underlying database. In this example we find that the "/home" method grabs data by pageID and displays the content.
http://127.0.0.1:5000/home/1
Now let's see if we can create an error by injecting a single quote
http://127.0.0.1:5000/home/1'
By doing so the SQL query syntax is now faulty. This is due to the fact that the user supplied input is being directly concatenated into the SQL query.
db.execute('SELECT pageId, title, content FROM pages WHERE pageId='+pageId)Now we can also use logical operators to determine whether we can actually manipulate the SQL statements.
We start with a logical operator which is false (and 1=2). The expected behaviour for injecting a false logical operator would be an error.
http://127.0.0.1:5000/home/1 and 1=2
After that we inject a logical operator which is true (and 1=1). This should result in the application run as intended without errors.
http://127.0.0.1:5000/home/1 and 1=1
Now that we know that the application is vulnerable for SQL injections we are going to use this vulnerability to read sensitive information from the database. This process could be automated with tools such as SQLMAP. However, for this example let's try to exploit the SQL injection manually.
The UNION operator is used in SQL injections to join a query, purposely forged to the original query. This allows to obtain the values of columns of other tables. First we need to determine the number of columns used by the original query. We can do this by trial and error.
http://127.0.0.1:5000/home/1 union select 1
This query results in an error, this is due to the fact that the original query started with 3 columns namely
* pageId
* title
* content
http://127.0.0.1:5000/home/1 union select 1,2,3
Notice how "title" and "content" became placeholders for data we want to retrieve from the database
Now that we determined the number of columns we need to take an educated guess for the table we want to steal sensitive information from. Again we can see if we try to query a non existent table we get an error. For a correct table we see the application function as intended.
http://127.0.0.1:5000/home/1 union select 1,2,3 from user
http://127.0.0.1:5000/home/1 union select 1,2,3 from users
SQL Injection can be prevented by following the methods described below:
Primary Defenses:
First step: White-list Input Validation
Second step: Use of Prepared Statements (Parameterized Queries)\
Additional Defenses:
Also: Enforcing Least Privilege
Also: Performing Allow-list Input Validation as a Secondary Defense
In this case, we have presented a SQLi code fix by using parameterized queries (also known as prepared statements) instead of string concatenation within the query.
The following code is vulnerable to SQL injection as the user input is directly concatenated into query without any form of validation: PATH:/SQLI/models/sqlimodel.py
cur = db.execute('SELECT pageId, title, content FROM pages WHERE pageId='+pageId)
This code can be easily rewritten in a way that prevent the user input from interfering with the query structure:
cur = db.execute('SELECT pageId, title, content FROM pages WHERE pageId=?', (pageId,))
Can you try to implement input validation or escaping?
Please refer to the OWASP testing guide for a full complete description about SQL injection with all the edge cases over different platforms!
https://www.owasp.org/index.php/Testing_for_SQL_Injection_(OTG-INPVAL-005)