trilwu/secskills ? reverse-engineered prompt

Build me a Claude Code skill pack for security work called SecSkills.

I want it to feel like a real personal security engineer inside Claude, not just a bunch of command snippets. It should be split into three installable parts, one for offensive security, one for defensive work, and one shared core for reverse engineering, code auditing, crypto review, AI and MCP security, and reporting. The skills should be written like practical playbooks with clear when to use this, when not to use this, and what bad assumptions to avoid.

Please make it so the skills can reference each other cleanly across the three packs, and so installing only one side still works without breaking anything. Keep everything fact checked against primary sources, and include a way to validate that the skills are structured correctly. If you need to look up current docs online while building it, do that.

I want the result to support red team, AppSec, DFIR, SOC, detection engineering, cloud and Kubernetes work, with the judgment and workflow discipline professionals actually use.