jwt-attack

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.95). The skill includes explicit JWTs and a leaked HS256 secret and instructs forming/sending Authorization headers and printing cracked secrets, which requires the model to handle and could output secret values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document is an explicit offensive guide for exploiting JWTs (alg:none, RS256→HS256 confusion, secret brute-force, kid injection, use of hardcoded/leaked tokens) and describes real-world data access, indicating deliberate malicious intent to bypass auth and steal data.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). The document contains one real high-entropy secret and one example JWT:
• The JWT in the potential matches / examples (eyJhbGci...dozjgNry...) is an illustrative/test token used in the decode/how-to examples (payload decodes to {"sub":"1234567890"}). This is a documentation/example token (low-sensitivity demo) so I am ignoring it as a false positive.
• The string b0c1df0e3f9c1e858d3bb0b8d58a119 (shown as JWT_SECRET) is a 32-hex-character, high-entropy literal explicitly labeled as a leaked secret in src/env.ts. That matches the definition of a secret (HS256 HMAC secret) and should be flagged.