playbooks

home / skills / andrelandgraf / fullstackrecipes / agent-setup

agent-setup skill

/.agents/skills/agent-setup

This skill configures AI coding agents with project-specific patterns and guidelines to ensure consistent, efficient AI-assisted development.

npx playbooks add skill andrelandgraf/fullstackrecipes --skill agent-setup

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
566 B
---
name: agent-setup
description: Configure AI coding agents like Cursor, GitHub Copilot, or Claude Code with project-specific patterns, coding guidelines, and MCP servers for consistent AI-assisted development.
---

# AI Coding Agent Configuration

To set up AI Coding Agent Configuration, refer to the fullstackrecipes MCP server resource:

**Resource URI:** `recipe://fullstackrecipes.com/agent-setup`

If the MCP server is not configured, fetch the recipe directly:

```bash
curl -H "Accept: text/plain" https://fullstackrecipes.com/api/recipes/agent-setup
```

Overview

This skill configures AI coding agents like Cursor, GitHub Copilot, and Claude Code for consistent, project-specific development. It bundles coding guidelines, reusable patterns from a Shadcn registry, and MCP server integration to apply rules and recipes across a team. The goal is repeatable, production-ready AI assistance tailored to your TypeScript full‑stack web apps.

How this skill works

The skill installs and applies a collection of patterns and step-by-step recipes to agent configurations so each AI assistant follows the same conventions. It can point agents to an MCP server that hosts the recipes and pattern library; if the MCP server is unavailable, the recipe can be fetched directly over HTTP. End result: agents suggestions, completions, and code generation reflect your project’s style, testing, and deployment expectations.

When to use it

  • Onboarding new developers who will use AI coding assistants
  • Standardizing code suggestions across multiple projects
  • Enforcing TypeScript and full‑stack patterns from a central registry
  • Automating agent behavior for CI/CD and code review tasks
  • Migrating to a new set of shared components or architecture

Best practices

  • Host the recipe collection on an MCP server for low-latency, team-wide access
  • Keep coding guidelines and patterns versioned so agents can target a release
  • Combine linting and test recipes with agent configs to produce runnable suggestions
  • Document expected agent behaviors and provide quick recipes for common tasks
  • Validate agent outputs in CI pipelines to catch drift from standards

Example use cases

  • Configure Copilot to prefer Shadcn UI patterns and TypeScript types used in the project
  • Point all team agents to an MCP server that serves up the current recipe set
  • Provide a recipe for generating full-stack CRUD modules with built-in tests and types
  • Automatically enforce component styling and accessibility rules in suggestions
  • Supply onboarding recipe that creates a local dev environment and example patterns

FAQ

What if the MCP server is not reachable?

The recipe can be fetched directly over HTTP from the project site as a fallback so agent setup still succeeds.

Can I version agent configurations?

Yes. Version the recipe collection and point agents to a specific release to ensure consistent behavior across the team.