afd-developer

Guide Agent-First Development for building software with AI agents as primary users.

4|1|Updated Dec 30, 2025
One-click install
npx skills add https://github.com/lushly-dev/afd --skill afd-developer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: afd-developer
Source: https://github.com/lushly-dev/afd/tree/main/.claude/skills/afd-developer
Command: npx skills add https://github.com/lushly-dev/afd --skill afd-developer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides expert guidance on the Agent-First Development (AFD) methodology, enabling the creation of software where AI agents are first-class users, ensuring commands are the product and UI is a swappable surface.

Core Features & Use Cases

  • Command Design: Learn best practices for naming conventions and structuring commands for clarity and agent compatibility.
  • CLI Validation: Understand the importance of testing commands via the CLI before building UI, ensuring a robust foundation.
  • CommandResult Schema: Implement structured outputs that include UX-enabling metadata like confidence, reasoning, and suggestions for AI-driven interactions.
  • Use Case: When designing a new feature for your application, use this Skill to understand how to define it as a command first, ensuring it's accessible to both human users and AI agents from the outset.

Quick Start

Use the afd-developer skill to learn about the Agent-First Development methodology and its core principles.

Frequently Asked Questions about afd-developer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is Agent-First Development and how does it structure software for AI agents?

Agent-First Development (AFD) is a methodology for building software where AI agents are first-class users, treating commands as the core product while the UI remains a swappable surface.

How do I design CLI commands that are optimized for AI agent compatibility?

To design commands for AI agents, follow AFD best practices for naming conventions and command structuring to ensure clarity, then validate command behavior via the CLI before building any UI layer.

What is a CommandResult schema and why do I need it for agent-driven interactions?

A CommandResult schema provides structured outputs containing UX-enabling metadata like confidence, reasoning, and suggestions, enabling AI agents to process command results and drive interactions effectively.

How do I integrate an MCP server into a command-first development workflow?

You integrate an MCP server by defining application features as commands first, ensuring they are accessible via CLI, and then connecting the MCP server to expose those validated commands to AI agents.

Do I need to validate commands via CLI before building the UI for my application?

Yes, CLI validation is a core principle of Agent-First Development, ensuring you test commands thoroughly through the command line to build a robust foundation before constructing any swappable UI surfaces.

When should I not use the command-first workflow for building features?

The command-first workflow may not suit projects where AI agents are not intended users or where immediate visual interfaces are strictly required before validating underlying command structures and outputs.