Projects
AI Agent Kit

AI Agent Kit

Stop wasting hours on repetitive prompts. Instantly inject expert skills, workflows, and infrastructure directly into your codebase to standardize your AI agent's output.

The Problem

Standard AI coding assistants (Cursor, Copilot, Windsurf) are powerful, but they lack your project's specific architectural context. Without standard instructions, you are forced to constantly explain your tech stack, rewrite bad AI code, and repeatedly enforce best practices.

The Solution

AI Agent Kit is an open-source CLI toolkit that installs a curated, battle-tested collection of specialized skills, workflows, and MCP configs directly into your project's .agents/ directory. With one simple command, your AI instantly learns how to work inside your repository.

The Outcome

  • Faster Execution: Write expert-level Web3, Backend, and Frontend code without repetitive explanations.
  • Standardized Architecture: Ensure every chunk of generated code precisely follows your team's internal conventions.
  • Zero Configuration: Seamless, out-of-the-box integration with Claude, Cursor, Copilot, Windsurf, and Gemini.

How It Works

1. One-Command Initialization

Prepare your workspace instantly.

$npx @ivannikov-pro/ai-agent-kit@latest init

2. Inject Specialized Skills

Over 15+ curated skills available. Teach your AI how to handle Express v5, deploy Solidity contracts, implement Figma API extractions, or use pnpm workspaces efficiently.

$npx @ivannikov-pro/ai-agent-kit@latest add solidity-contracts

3. Run Automated Workflows

Transform your AI into specific team roles. Execute ready-to-use workflows, prompting your assistant to act as a CTO (for code reviews), a Security Engineer (for vulnerability audits), or a Blockchain Dev.