huawei-pentesting

SUSPICIOUS: the skill is internally consistent as an offensive Huawei Cloud pentesting playbook, but that stated purpose is itself high risk for an AI agent. Install sources are mostly official and not strongly malicious, yet the skill materially enables credential theft, privilege escalation, cloud takeover, and sensitive data access on real infrastructure.

该输入内容为针对华为云多服务的可操作攻击手册，包含获取集群管理凭据、特权 Pod/宿主逃逸、从云元数据窃取 IAM 凭据、函数代码持久化后门、以及通过 KMS 解密与 DEK 生成来支撑数据勒索等明确恶意目的。若该内容存在于软件依赖或发布工件中，将属于高风险供应链恶意内容。

This fragment is a high-risk offensive intrusion playbook that explicitly describes harvesting sensitive credentials from OpenStack metadata (securitykey/user_data), authenticating to OBS to enumerate and recursively download confidential objects, forcing data exposure by changing OBS bucket ACL to public-read, establishing persistence via RDS db_user creation and backups, and hijacking traffic by modifying ELB backend pools. While it is not executable code from a dependency, distributing this content as part of a software package/repo artifact would indicate malicious intent and materially elevate the security risk.