Impact
In mitmproxy 7.0.4 and below, a malicious client or server is able to perform HTTP request smuggling attacks through mitmproxy. This means that a malicious client/server could smuggle a request/response through mitmproxy as part of another request/response’s HTTP message body. While mitmproxy would only see one request, the target server would see multiple requests. A smuggled request is still captured as part of another request’s body, but it does not appear in the request list and does not go through the usual mitmproxy event hooks, where users may have implemented custom access control checks or input sanitization.
Unless you use mitmproxy to protect an HTTP/1 service, no action is required.
Patches
The vulnerability has been fixed in mitmproxy 8.0.0 and above.
Acknowledgements
We thank Zeyu Zhang (@zeyu2001) for responsibly disclosing this vulnerability to the mitmproxy team.
Timeline
- 2022-03-15: Received initial report.
- 2022-03-15: Verified report and confirmed receipt.
- 2022-03-16: Shared patch with researcher.
- 2022-03-16: Received confirmation that patch is working.
- 2022-03-19: Published patched release and advisory.
References
- https://github.com/mitmproxy/mitmproxy/security/advisories/GHSA-gcx2-gvj7-pxv3
- https://nvd.nist.gov/vuln/detail/CVE-2022-24766
- https://github.com/mitmproxy/mitmproxy/commit/b06fb6d157087d526bd02e7aadbe37c56865c71b
- https://mitmproxy.org/posts/releases/mitmproxy8/
- https://github.com/advisories/GHSA-gcx2-gvj7-pxv3