post-exploit-windows

该技能目的与能力一致，但其目的本身就是 Windows 后渗透与凭据获取。它为 AI 代理提供高危攻击、绕过和远程执行流程，并引入 attacker-hosted 脚本与非官方 PoC，整体应归类为高风险、具有明显滥用价值的攻击性技能，而非良性开发辅助技能。

该片段不是恶意程序实现，而是高度可操作的 Windows/AD 渗透与提权指南。它直接展示凭据窃取、AMSI/日志绕过、RBCD、SYSTEM 提权和反向 Shell 技术；若其中命令被自动执行或打包进安装流程，将构成严重安全风险。仅依据所给内容，不能确认 npm 包存在实际恶意安装行为，但文档的攻击性和潜在滥用风险明确。

该内容不是软件包源代码，而是高度可操作的 PrintNightmare 漏洞利用指南。它明确包含恶意 DLL 生成、远程 SMB 加载、反向 Shell、管理员账户创建和域环境攻击步骤，具有显著攻击性和高滥用风险。文档本身未显示自动执行、隐蔽持久化或数据外传代码，因此恶意软件概率低；但其安全风险很高，应仅在获得授权的隔离测试环境中使用。

No direct malware code or data theft/exfiltration behavior exists in this JSON fragment. However, it is explicitly structured to drive an LLM/agent retrieval workflow toward Windows post-exploitation and privilege escalation content, including credential-dumping-related keywords and a SYSTEM escalation scenario. As a supply-chain dependency in an agent/skill retrieval pipeline, it represents a high-risk capability-enabling configuration artifact rather than a benign library component.

The provided fragment is an explicit offensive Windows privilege-escalation playbook (Potato/UAC bypass/AlwaysInstallElevated/service/DLL hijacking/WinPEAS+PowerUp automation) and includes reverse-shell payload guidance. This strongly indicates malicious intent and would be considered extremely dangerous if included in a software supply chain.

This module is highly risky because it is an attacker playbook dataset that provides scenario-based instructions for Windows privilege escalation (including UAC bypass) and credential theft (LSASS dumping). Even though it contains no executable logic in this fragment, the embedded 'expected_output' content is directly operational for harmful compromise and should be treated as malicious instructional material in a supply-chain context.

This fragment is a high-confidence offensive playbook for Windows credential theft (SAM/SYSTEM hive dumping, LSASS dumping and logon-password extraction, browser/credential-manager/PowerShell-history/WiFi secret harvesting) combined with Active Directory/domain reconnaissance. It strongly indicates malicious intent or extreme misuse risk; it should be treated as dangerous and not incorporated into any trusted software supply chain. Obfuscation is not evident, but the actions themselves are directly credential-compromising.