---
id: pattern:installing-skills
type: pattern
title: Installing the Skill Library on an AI Agent
status: active
confidence: 0.7
sources:
  - 2026-05-09-cybersec-library-overview.md
created: 2026-05-09
updated: 2026-05-09
updated_log:
  - 2026-05-09: created
tiers: semantic
half_life_days: 180
tags: [installation, agent-integration, deployment]
---

# Installing the Skill Library on an AI Agent

## Summary

Two commands turn any [[concept:agentskills-io-standard]]-compatible AI agent into a security analyst: `npx skills add mukul975/Anthropic-Cybersecurity-Skills` (recommended), or a plain `git clone`. The library works zero-config across 26+ AI platforms — all the major code assistants, CLI agents, autonomous agents, and agent frameworks. Once installed, the agent scans skill frontmatters during prompt processing and loads only the skills relevant to the current task.

## Claims

- The recommended install is `npx skills add mukul975/Anthropic-Cybersecurity-Skills`. The fallback is `git clone https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git && cd Anthropic-Cybersecurity-Skills`. `[src: raw/2026-05-09-cybersec-library-overview.md] {conf: 0.85}`
- The library "works immediately" with: Claude Code (Anthropic), GitHub Copilot (Microsoft), OpenAI Codex CLI, Cursor, Gemini CLI (Google), and any agentskills.io-compatible platform. `[src: raw/2026-05-09-cybersec-library-overview.md] {conf: 0.85}`
- Compatible AI code assistants (12 listed): Claude Code, GitHub Copilot, Cursor, Windsurf, Cline, Aider, Continue, Roo Code, Amazon Q Developer, Tabnine, Sourcegraph Cody, JetBrains AI. `[src: raw/2026-05-09-cybersec-library-overview.md] {conf: 0.8}`
- Compatible CLI agents: OpenAI Codex CLI, Gemini CLI. Compatible autonomous agents: Devin, Replit Agent, SWE-agent, OpenHands. Compatible frameworks/SDKs: LangChain, CrewAI, AutoGen, Semantic Kernel, Haystack, Vercel AI SDK, any MCP-compatible agent. `[src: raw/2026-05-09-cybersec-library-overview.md] {conf: 0.8}`
- The runtime model: an agent prompt triggers a sweep of 754 skill frontmatters (~30 tokens each), top matches load fully (500-2,000 tokens each), then the agent executes the structured `Workflow` section step-by-step and validates via the `Verification` section — IOCs map back to ATT&CK techniques. `[src: raw/2026-05-09-cybersec-library-overview.md] {conf: 0.75}`
- "All platforms that support the agentskills.io standard can load these skills with zero configuration" — portability flows through the open standard, not platform-specific glue. `[src: raw/2026-05-09-cybersec-library-overview.md] {conf: 0.7}`

## Relationships

- uses → [[concept:cybersec-skill-library]] `{conf: 0.85}`
- depends-on → [[concept:agentskills-io-standard]] `{conf: 0.8}`
- complements → [[pattern:skill-anatomy]] `{conf: 0.75}`

## Open questions

- [ ] Concrete platform-specific install instructions (where do skills land for Claude Code vs. Cursor vs. Copilot)? README treats them as interchangeable.

## Changelog

- 2026-05-09 — created