dragonked2/alphacode/hunt-jwt
Claude Code and agent skill by dragonked2/alphacode
- community
Checking that the SKILL.md file is still in the repository.
JWT attacks — none algorithm, alg confusion, key brute, claim manipulation, JKU/JWK injection, token substitution. Chains to auth bypass and ATO.
Install
npx vskill@latest install dragonked2/alphacode/hunt-jwtSKILL.md
JWT ATTACKS — 3 BULLETS MAX
Core: JWT flaws = authentication bypass = ATO.
DETECTION
One-call decode (preferred):
ATTACKS
THE ALG=NONE BYPASS — ONE CALL, NO HAND-CRAFTING
This is the highest-yield JWT attack and the one most often missed because it looks like it "should not still work". It does. Many APIs validate the signature only when alg is a signing algorithm, and skip validation entirely when alg: "none".
Forge the token with the `jwt` tool — do not hand-craft base64:
Test it immediately against the endpoint that rejected the original:
If the server accepts it → Critical. The forged token needs no signature, no secret, and no account. It works against any endpoint the original token reached.
alg:none — the real-world chain
This is not just "auth bypass". The pattern that pays out:
When you get an alg:none token, test in this order (stop at first proven boundary):
- Read:
GET /api/admin/users— confirm admin access - Scope: how many records? One endpoint, one record to prove. Do not
bulk-exfiltrate. See control-verification IMPACT BOUNDARY.
- Write (only if explicitly asked):
POST/DELETEon a test object
you create and delete yourself. Never delete production data.
- CORS: check
Access-Control-Allow-Originon the API response. If
* alongside the forged token, note the compound risk — any website can make authenticated admin calls.
The alg:none test is 30 seconds — always run it
Before any other JWT work, on any token-bearing endpoint, run the forge-and-retry. It is the cheapest highest-severity test in the arsenal and it works on a surprising fraction of targets.
CHAINS
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingsrc/alphacode_base/bundled_skills/bugbounty/hunt-jwt/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/dragonked2/alphacode/hunt-jwt)