security · Found in 2 repositories

exploiting-jwt-algorithm-confusion-attack

Exploits JWT algorithm confusion vulnerabilities where the server's token verification library accepts the algorithm specified in the JWT header rather than enforcing a fixed algorithm. The tester manipulates the alg header to switch from RS256 to HS256 (using the RSA public key as the HMAC secret), sets alg to none to bypass signature verification, or exploits kid/jku/x5u header injection to supply attacker-controlled keys. Activates for requests involving JWT algorithm confusion, alg none a...

Install

Install with the skills CLI. It asks which coding agents to install into:

npx skills add 'yanacuti1121/Yana-AI/core/skills/exploiting-jwt-algorithm-confusion-attack/SKILL.md#main'

Or install globally for one agent:

# Claude Code → ~/.claude/skills/
npx skills add 'yanacuti1121/Yana-AI/core/skills/exploiting-jwt-algorithm-confusion-attack/SKILL.md#main' -g -a claude-code
# Codex → ~/.agents/skills/
npx skills add 'yanacuti1121/Yana-AI/core/skills/exploiting-jwt-algorithm-confusion-attack/SKILL.md#main' -g -a codex
# Cursor → ~/.cursor/skills/
npx skills add 'yanacuti1121/Yana-AI/core/skills/exploiting-jwt-algorithm-confusion-attack/SKILL.md#main' -g -a cursor

Manual install locations: Claude Code ~/.claude/skills/, Codex ~/.agents/skills/, Cursor ~/.cursor/skills/. Source-path status is inferred from metadata; it does not verify a live download. Scan and quality scores describe registry checks and are not a guarantee of safety.

What the registry checked

Scanned 2026-11-16 by registry scanner 1.1.7 (ruleset 371c677df515, content f2f129e15205).

  • File size limit
  • Frontmatter and schema
  • Dangerous code patterns
  • Sensitive file paths
  • Bundled scripts and support files
  • Blocked source repositories
  • Prompt injection phrases
  • Hardcoded credentials
  • Obfuscated execution chains

Findings: 5 warnings: dangerous pattern (1), schema validation (1), sensitive path (3).

Checks are pattern-based and cover the archived copy; they do not guarantee the skill is safe.

Content history

  1. 2026-11-16 Several upstream SKILL.md files match this contentsha256:6199600a15b0 · 1 file

Each entry is a period in which the archived files (SHA-256 per file) and their match against the author's repository stayed the same. Upstream checks run at most weekly, so detection can lag.

Attribution and permissions

Check the original repository for permission and license terms before reuse. Registry metadata does not grant a license.

This guide links to the author’s instructions and includes metadata only.

Request removal or correct attribution

For the skill author

Add this badge to your README to link readers to the registry home:

Listed on Claude Skills Registry

[![Listed on Claude Skills Registry](https://majiayu000.github.io/claude-skill-registry/assets/listed-badge.svg)](https://majiayu000.github.io/claude-skill-registry/)

Copies with matching content

Exact Markdown body copies across 2 repositories. This count does not identify the original author.

More in security