tencent-pentesting

SUSPICIOUS/HIGH-RISK skill. Its actual footprint is an offensive cloud exploitation runbook for an AI agent: harvesting credentials from metadata and secret stores, escalating privileges, executing commands, extracting code/data, and changing network controls. Official Tencent CLIs are consistent with the target platform, but the overall capability set is inherently dangerous for an agent skill and is amplified by the third-party `cf@latest` install. Not confirmed malware, but clearly a high-risk offensive-security skill.

该文件不是软件供应链代码，而是高度可操作的腾讯云渗透与凭据窃取指南。内容涵盖容器逃逸、元数据凭据获取、日志和密钥批量读取、云函数后门及潜在数据勒索。文档自身没有执行或外传能力，但其中命令被未经授权执行时可造成严重安全影响。

该输入不是软件供应链代码，而是一份高度攻击导向的腾讯云渗透与云资源滥用指南。内容本身不会自动执行操作，但所示命令可在具备相应权限或存在 SSRF、公开存储等条件时导致远程命令执行、临时凭据泄露、数据外泄、数据库持久化账号和负载均衡流量劫持。未发现包内恶意载荷、自动外传或持久化机制。