Anthropic-Cybersecurity-Skills

Why we included this project

Security teams running AI coding agents will find a lot to like here: 817 structured skills spanning 29 security domains, from memory forensics and malware traffic analysis to cloud breach scoping and fraud detection. Each skill follows the agentskills.io open standard, so Claude Code, Copilot, Codex, Cursor, and other tools can load them directly. The framework mapping is the real draw: every skill is tagged against the MITRE and NIST taxonomies that apply to it, so a forensics skill carries ATT&CK and CSF references while an AI-security skill adds ATLAS and AI RMF, which beats generic advice when you point an agent at a task. This is a curated skill collection rather than a standalone application, so plan to clone it and wire it into your existing agent setup, but the coverage makes it a solid starting point for automating incident response, threat hunting, and red-team work.