ssh-pentesting

MALICIOUS。该技能并非普通开发或运维集成，而是专门赋予 AI 代理 SSH 渗透、爆破、漏洞利用、凭据窃取、隧道转发、MITM 截获和横向移动能力，实际足迹与高风险攻击行动高度一致。虽未见预执行指令或内置第三方窃密网关，但作为 AI Agent Skill，其用途本身即是 offensive security/real-world attack enablement，应判定为高风险恶意技能。

This fragment is not code implementing a library; it is an offensive SSH/pentesting/compromise playbook. It provides clear operational guidance to enable tunneling-based access, SSH key/agent discovery and reuse, SFTP abuse concepts, and SSH MitM credential interception, plus references to exploitation/testing workflows. No obfuscation is present, and there is no evidence of a hidden executable payload in the provided text alone. In a supply-chain scenario, however, publishing or bundling this material is highly suspicious from a misuse standpoint and warrants deeper inspection of surrounding package behavior (e.g., install-time scripts, runtime execution, and whether any automation is triggered).