Welcome to the New Era of Software Development.
AI coding assistants are no longer just for autocompletion. They are active collaborators in the creative process. This guide provides an interactive exploration of two leading platforms, **Claude Code** and **Cursor**, helping you understand their philosophies, master their workflows, and build your first AI-powered agent.
Start ExploringHead-to-Head: Claude Code vs. Cursor
Choosing between Claude Code and Cursor is a choice between two philosophies: the AI as a powerful tool you wield, or as an integrated partner that anticipates your needs. Explore their core differences below.
A Philosophical Divide
This chart visualizes the core design philosophies. Hover over the points to learn more.
Interactive Feature Comparison
Claude Code
Cursor
Mastering Agentic Workflows
Move beyond basic features and unlock a step-change in productivity. Mastering these advanced techniques requires a mental shift from giving instructions to orchestrating processes.
Supercharging Claude Code
Orchestrate with Sub-Agents (New!)
Delegate tasks to specialized AIs. Create sub-agents with specific prompts and tools (e.g., a "Test Writer" or "Docstring Generator") that Claude can invoke, keeping the main context clean and focused.
The `CLAUDE.md` Manifest
Create a permanent "brain" for your project. Use `/init` to generate a file where you define core logic, common commands, and coding styles for Claude to follow in every session.
Strategic Planning with "Plan Mode"
Press `Shift+Tab` twice to enter a read-only mode. Claude will research the codebase and propose a detailed plan for your approval before making any changes, ensuring safety and precision.
Hyper-Productivity in Cursor
Autonomous "YOLO Mode"
Enable this mode to let the AI run tests and build commands in a loop. Give it a goal, like "fix all build errors," and it will work autonomously until the task is complete.
AI-Driven TDD
Prompt the agent to write tests first, then the implementation code to make them pass. Cursor can run the full `code -> test -> fix` cycle from a single high-level instruction.
Project-Specific Guardrails
Use a `.cursorrules` file to enforce team-wide standards, like "Always use Tailwind CSS for styling." This ensures AI-generated code remains consistent with your project's architecture.
Essential GitHub Repositories
microsoft/autogen
A powerful framework for simplifying the orchestration, optimization, and automation of complex LLM workflows.
crewAIInc/crewAI
Designed for orchestrating role-playing, autonomous AI agents that can collaborate to tackle complex tasks.
e2b-dev/awesome-ai-agents
A curated list of the best AI agent frameworks, tools, and resources to accelerate your projects.
PrefectHQ/ControlFlow
A Python framework for building agentic AI workflows with a focus on developer control and transparency.
system-prompts-and-models
Peek behind the curtain. This repo collects the system prompts and models used by popular AI tools like Cursor.
500-AI-Agents-Projects
A massive collection of over 500 AI agent projects and use cases for inspiration and learning.
The Art of Instruction
The quality of AI-generated code is directly proportional to the quality of your prompts. Master these expert-level techniques to gain precise control over the output.
Assignment: Build "Paper Buddy"
Apply what you've learned by building a useful command-line agent. "Paper Buddy" finds the latest academic paper on a topic and emails you a summary. Follow the steps below to build it with an AI assistant.