[Add] RS256 - HS256 Type Mismatch Attack

[sadkris]

Hey, wanted to just suggest an addition to this project of implementing the algorithm confusion attack as detailed here. Essentially, given any two JWT hashes signed with RS256, you can derive the public key that was used to sign them. Based on that, if the code isn't verifying the set algorithm, you can arbitrarily sign your keys as HS256 by using the public key as the secret since that's what would be used as a key for verification of signatures.

Secondarily, I would prefer that bruteforce processes would be run from an internal instanced docker container when requested, as the current system isn't efficient enough to provide consistent results as opposed to using something such as gojwtcrack, or just hashcat.

[dr34mhacks]

Hi Sadkris,

Fantastic suggestion! Algorithm confusion attacks are definitely on our roadmap, as this is a serious vulnerability affecting many JWT implementations. We’re planning to implement:

1. Detection Module – Automatically check if the server accepts multiple algorithms
2. Public Key Extraction – Derive RSA public keys from RS256-signed tokens
3. Attack Generator – Convert RS256 tokens to HS256 using the extracted public key as the HMAC secret
4. Educational Content – Provide comprehensive documentation on this attack vector as part of our Documentation here at https://jwtauditor.com/docs/jwt-vulnerabilities-guide.html Although we have added a sneak peak of the attack but will expand this further.

This aligns perfectly with our goal of delivering a complete JWT security testing platform. We’ll prioritize this for the next major release. Thanks for the detailed reference, PortSwigger’s documentation is excellent!

Additionally, we appreciate the performance feedback this echoes similar requests we’ve received. While we maintain our 100% client-side philosophy for privacy-critical scenarios (many pentesters can’t send tokens to external processes due to NDAs or compliance), we’re exploring a hybrid approach:

• Short-term: Further optimizing our JavaScript implementation with Web Workers and smarter algorithms
• Long-term: Developing an optional Docker companion tool for speed-focused use cases where privacy isn’t the primary concern (CTF Focused)

The web version will always remain our flagship for universal access and strong data privacy, but we recognize the need for high-performance options.