Anthropic Cybersecurity Skills
Library of 800+ cybersecurity skill files for AI coding agents
What it is
A collection of structured skill files in the agentskills.io format that teach AI agents how to carry out security work such as memory forensics, threat hunting, incident response, cloud breach scoping, and penetration testing. Every skill is tagged against six frameworks: MITRE ATT&CK and ATLAS, D3FEND, NIST CSF 2.0, the NIST AI RMF, and MITRE F3. Security analysts and red or blue teams load it into Claude Code, Codex CLI, Cursor, Gemini CLI, or similar agents to get step-by-step guidance during authorized investigations and tests.