hypnguyen1209/offensive-claude

This repository is a configuration bundle for Claude Code aimed at security researchers, red teamers, and vulnerability analysts. The author describes it as a comprehensive setup that turns Claude Code into a specialized assistant for offensive security work, with 25 skills, 6 sub-agents, and 46 vulnerability reference files. Installation is a one-line curl-and-bash from the repo, or alternatively a git clone plus an install script, or a manual copy of the skills, agents, and CLAUDE.md folders into the user's ~/.claude directory. After install, the skills activate on their own based on context, and the CLAUDE.md file acts as the system prompt that sets the security-focused persona. The skill list spans the lifecycle the README describes for this kind of work: reconnaissance and open-source intelligence, vulnerability analysis, exploit development, reverse engineering, web pentesting, network and Active Directory attacks, red-team operations, cloud security, malware analysis, AI model security, threat hunting, Linux and Windows privilege escalation, cryptography review, incident response, EDR evasion, initial access, shellcode authoring, Windows mitigation bypass, mobile pentesting, and command-and-control infrastructure planning. The 6 agents are described as roles that Claude can spawn for parallel or specialist work, including a red-team planner, an exploit researcher, a security reviewer, a reverse engineer, an AI researcher, and a network analyst. The 47 reference files contain vulnerable and secure code patterns grouped under topics like taint analysis, memory safety, injection, authentication, cryptography, concurrency, web and API issues, supply chain, and Active Directory. Three MCP servers are listed for web search, IDA Pro, and JADX integration.