Ai on QuickRef - Personal Developer Reference

Claude CLI

Mon, 18 Aug 2025 00:00:00 +0000

Overview

@anthropic-ai/claude-code is the official command-line interface (CLI) for Claude, developed by Anthropic. It allows developers to interact with Claude models directly from the terminal, making it easy to integrate AI into development workflows, analyze code, and generate content without leaving the command line.

Installation

You must have Node.js (version 18 or higher) and npm installed.

To install the Claude CLI globally on your system, run the following command:

npm install -g @anthropic-ai/claude-code

Configuration

The first time you run the claude command, it will guide you through an authentication process in your web browser.

Gemini CLI

Mon, 18 Aug 2025 00:00:00 +0000

Overview

@google/gemini-cli is the official, open-source command-line interface (CLI) for Gemini, developed by Google. It brings the power of Gemini models to your terminal, allowing you to ask questions, understand and generate code, and integrate AI into your development workflows.

Installation

You must have Node.js (version 20 or higher) and npm installed.

You can either run the CLI without installation using npx or install it globally for regular use.

Run with NPX (no install):

Managing AI Agents with AGENT.MD

Mon, 18 Aug 2025 00:00:00 +0000

AGENT.MD: Persistent Guidance for AI Agents

For ongoing projects, especially when collaborating with an AI agent, creating an AGENT.MD file in the project’s root directory is a powerful technique. This file acts as a persistent, high-level prompt that guides the agent’s behavior, ensuring consistency and adherence to project standards over multiple interactions.

Why is `AGENT.MD` Important?

Consistency: Ensures the agent follows the same style, conventions, and architectural patterns every time it works on the codebase.
Efficiency: Reduces the need to repeat instructions in every prompt. The agent can refer to the AGENT.MD for standing orders.
Safety: Allows you to define clear safety rules, forbidden commands, or sensitive areas of the codebase to avoid accidental changes.
Onboarding: Helps the agent (and new human developers) quickly understand the project’s specific operational procedures and conventions.

How to Design an Effective `AGENT.MD`

An effective AGENT.MD should be structured like a set of clear instructions. Use Markdown to create a readable and well-organized document.

Prompt Engineering

Sun, 01 Jan 2023 00:00:00 +0000

Overview

Prompt engineering is the art and science of crafting effective prompts for Large Language Models (LLMs) to achieve desired outcomes. It involves understanding how to structure, phrase, and format inputs to get the best possible responses from AI systems.

Core Principles

Clarity and Specificity

❌ Poor: "Write about dogs"
✅ Good: "Write a 300-word informative article about Golden Retriever training techniques for first-time dog owners"

Context Setting

❌ Poor: "Explain recursion"
✅ Good: "Explain recursion to a computer science student who understands basic programming concepts but struggles with advanced algorithms. Use simple examples and analogies."

Role Definition

❌ Poor: "Help with my code"
✅ Good: "As an experienced Python developer, help me debug this function and explain the error in detail"

Framework

Getting the most out of generative AI is all about understanding inputs and outputs. What are you hoping to get from the tool? That’s the output. What kind of prompt is going to help you get it? That’s the input. And there’s a simple framework that helps you provide clear and specific directions to any gen AI tool.

Ai on QuickRef - Personal Developer Reference

Claude CLI

Overview

Installation

Configuration

Gemini CLI

Overview

Installation

Managing AI Agents with AGENT.MD

AGENT.MD: Persistent Guidance for AI Agents

Why is AGENT.MD Important?

How to Design an Effective AGENT.MD

Prompt Engineering

Overview

Core Principles

Clarity and Specificity

Context Setting

Role Definition

Framework

Why is `AGENT.MD` Important?

How to Design an Effective `AGENT.MD`