SQL injection (or SQLi) is one of the most widespread code vulnerabilities. To perform a SQL injection attack, an attacker inserts or "injects" malicious SQL code via the input data of the application. SQL injection allows the attacker to read, change, or delete sensitive data as well as execute administrative operations on the database.
In this lesson, you will learn how SQL injection works and how to protect your code against it. We will begin by using SQL injection to bypass the login screen of a vulnerable web application. We will then dive deeper into the code of that vulnerable application and explain why the SQL injection attack was effective. Finally, we will teach you how to fix and prevent SQL injection in your applications.
Did you know that, according to Star Trek, SQL injection will still plague humanity 300 years from now? In "If Memory Serves", a 2019 Star Trek Discovery episode, a probe used SQL injection when attacking a datastore on one of the ship's shuttlecraft. The attack was discovered by Commander Airiam. Sadly, in the 24th century, people are still writing insecure SQL queries.
To see how easy it is to perform a SQL injection attack, let's play with a vulnerable web application–the login screen of a fictional company called startup.io. We will attempt to bypass their login screen by supplying malicious input to the password field.
Let's start with a valid email and password.
email@example.com in the email field and
correcthorsebatterystaple as password.
Try it out. After you press the login button, you should see a successful login message.
We used the quote character (
') in the password field to inject an unintended string into the query in order to change the query's logic. Using other metacharacters, such as
;, is also possible.
Let's look at this example in more detail. First, consider the backend logic of startup.io’s login functionality. The following code sample shows a function which aims to verify the user's credentials.
The key here is the construction of the query through string concatenation. Let's try to inject the malicious password which we use to bypass the login into the query. After the string concatenation is done, the
WHERE clause of the query becomes:
WHERE firstname.lastname@example.org'AND password='idontknow' OR 1=1;
If we consider the following:
ANDhas precedence over
OR(true for most SQL implementations)
FALSEfor the given email
we can reduce the
WHERE clause as follows:
-> email@example.com' AND password='idontknow' OR 1=1-> (TRUE AND FALSE) OR TRUE-> FALSE OR TRUE-> TRUE
A SQL injection attack illustration where a malicious input is injected into a login window of a web application
By injecting malicious input into the query, we managed to change the query's logic! The query will evaluate as
TRUE even when the supplied password is invalid. Since the query result is used to decide if a given user should be allowed to log in, we’ve bypassed the application's authentication mechanism.
Now that you know how SQL injection works, imagine the implications. By using SQL injection, an attacker can change the logic of the vulnerable query. They can read or modify any data stored in the database. If the vulnerable query is executed as a privileged user, the attacker can run any administrative operations, including deleting the entire database schema.
SQL injection is widespread because it is easily detected and exploited. Any website or application which uses a database is subject to a SQL injection attack. In the past, almost all popular SQL clients and libraries were vulnerable to SQL injection attacks.
SQL injection has been used by creative people in many unexpected ways. For example, in 2014, an individual in Poland legally renamed his business to "Dariusz Jakubowski x'; DROP TABLE users; SELECT '1" to disrupt the operation of spammers' harvesting bots. Wonder what this peculiar name means? Keep reading to find out!
To protect your app against SQL injection, you need to prevent user-supplied input from affecting the logic of the executed query. One way to achieve that is with a parameterized query, which almost all SQL frameworks support.
For instance, if our example used a parameterized query, and you were to insert
' OR 1=1' as a password input to that query, a database client would not interpret that input as SQL. Instead, it would look for a password string
' OR 1=1'
A SQL injection mitigation illustration where the malicious input injection is prevented by parametrized query
Parameterized queries bind user-supplied input, preventing SQL injection. Under the hood, they typically use the database's prepared statement functionality. Go's
database/sql package automatically creates prepared statements for you with bound parameters whenever additional arguments are passed.
Consider a possible fix to our SQL injection problem, which uses bound parameters. You can see the difference in the diff below.
Most SQL client libraries allow for queries with parameterized values. In our case, this is
conn.QueryRowContext. Notice the usage of placeholders ? as query parameters, and
password as additional arguments. In the broken code example, we inject parameters directly into the query. In the fixed code example, we put placeholders in all places where we previously injected input parameters. This way, we tell the SQL database to expect input parameters in all locations where it sees ?. Our call to
conn.QueryRowContext notices the additional
password arguments, and sends a preparted statement to the database, then runs it with the additional arguments.
When the prepared statement is sent to the database, the statement does not contain any user-supplied data. When the database executes the statement, it uses our additional arguments as bound parameters, replacing the ? placeholders in the query with the values of the additional arguments in a safe manner.
This way we keep the parameters of the query and the query itself cleanly separated. The job of combining the two falls into the database engine, which eliminates the SQL injection vulnerability.
To learn more about SQL injection, check out some other great content produced by Snyk:
The SQL injection cheat sheet which lists other methods of preventing SQL injection in your code.
Our blog post on SQL injection in ORM for more information on a different prevention method (ORM) and why sometimes it is not enough to mitigate SQL injection.
Snyk OWASP Top 10 for an overview of other common vulnerabilities present in modern web applications.
You’ve learned what SQL injection is and how to protect your systems from it. We hope you will apply your new knowledge wisely and make your code much safer.
Feel free to rate how valuable this lesson was for you and provide feedback to make it even better! Also, make sure to check out our lessons on other common vulnerabilities.