jwt-attack-methodology

No executable code is present—only explicit instructions for conducting JWT attacks (None-alg testing, weak-key brute force, RS256/HS256 confusion, and kid injection). As a standalone snippet it does not demonstrate malware behavior, but its offensive nature is a notable supply-chain concern if such content is shipped as part of a dependency. Additional surrounding code/artifacts are needed to determine whether any runtime component actually performs these actions.

This skill is not internally inconsistent; it is a coherent offensive JWT exploitation playbook. However, because it equips an AI agent to conduct credential abuse, auth bypass, cracking, injection, SSRF, and privilege escalation against targets, it is high-risk exploit tooling and should be classified as suspicious/high-risk rather than benign.