> For the complete documentation index, see [llms.txt](https://greenhat.gitbook.io/interview-bank/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://greenhat.gitbook.io/interview-bank/web/code-review.md).
# Code Review
So here is a code that uses a Flask web server with routes to set and get values stored in a Redis database. The values retrieved from Redis are unserialized using Python `pickle` module.
**Spot the vulnerability in this code snippet!**
Code Review Challenge
**Are you able to find the vulnerability?**
The vulnerability code: `data = pickle.loads(value)`
`Pickle` is a powerful module for serializing and deserializing objects in Python. **However**, it is infamously insecure when used with untrusted data! The vulnerability lies in `pickle.loads()` which can execute arbitrary code contained within the data it is attempting to unserialize.
This is known as **Insecure Deserialization** where user-controlled data is deserialized by a website. **Deserialization** is the process of restoring this byte stream to a fully functional replica of the original object.
Process of Serialization and Deserialization. Credits: Portswigger
If an attacker can store malicious data in the Redis cache, he/she could craft a payload that leads to remote code execution.
**So, how can we fix this vulnerability?**
For starters, we should not use `pickle` to unserialize data coming from an untrusted or controllable source. We can use a safer format like JSON which does not allow code execution
