Install Anthropic Cybersecurity Skills for agent security workflows

Use this skill when an operator wants an AI agent to follow structured cybersecurity playbooks instead of improvising commands from a prompt. The operator installs the Anthropic Cybersecurity Skills library, points the agent runtime at the skills, selects the relevant domain workflow, and has the agent execute the steps with the repository’s MITRE ATT&CK, NIST CSF, MITRE ATLAS, D3FEND, AI RMF, and MITRE F3 mappings available as context.

Invoke this for authorized security work where repeatable skill selection matters: threat hunting, DFIR, malware analysis, cloud security review, SOC triage, penetration testing, AI security assessment, or fraud-investigation workflows. The source library is explicitly for lawful, authorized security work and includes dual-use/offensive techniques, so the operator must apply normal rules of engagement and permission checks before use.

The scope is installing and using the structured cybersecurity skill library as agent context for a bounded security workflow. It is not a generic cybersecurity product card, a broad “security automation” listing, or a claim that every included playbook has been independently security reviewed by ASE.

Source-backed quick start from the upstream README:

“`bash
npx skills add mukul975/Anthropic-Cybersecurity-Skills
“`

Alternative source-backed install path:

“`bash
git clone https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git
cd Anthropic-Cybersecurity-Skills
“`

After installation, have the agent scan frontmatter for the relevant security domain, load the top matching skills, execute the workflow steps, and record findings with the included verification and framework mappings.