Cursor for Professional Developers

Class Duration

14 hours of live training delivered over 2-3 days to accommodate your scheduling needs.

Student Prerequisites

Professional software development experience
Cursor installed (Pro or Business plan recommended for full feature access)

Target Audience

Software engineers and tech leads who are adopting Cursor as a primary development environment and want to use its most powerful features—Composer 2.0, Background Agents, Cloud Agents, multi-file refactoring, rules files, and context management—at a professional level, not just for basic autocomplete. Especially relevant for teams evaluating Cursor 3.0's multi-repo workspace and self-hosted cloud agents.

Description

Cursor is built from the ground up as an AI-first code editor, and the 3.0 release turned it into a unified workspace for human-and-agent software development—shipping Cursor's in-house frontier coding model, Composer 2 (built on Moonshot's Kimi K2.5 with continued pretraining and reinforcement learning, 200+ tok/s, scoring 61.3 on CursorBench). This course covers the professional-grade features that distinguish it from add-on assistants. We cover Composer 2 for multi-file generation and refactoring, agent mode for longer-horizon autonomous tasks, the new Background Agents that work in parallel locally while you keep typing, Cloud Agents (and self-hosted cloud agents) for off-machine task execution, Design Mode for visual UI iteration, the multi-repo workspace, rules files (.cursorrules and project-scoped rules) for standardizing AI behavior, and advanced context management using @codebase, @docs, @web, and custom context sources. Labs use realistic multi-file scenarios and include team-configuration best practices.

Learning Outcomes

Navigate Cursor 3.0's feature set: chat, inline edit, Composer 2, agent mode, Design Mode, and the multi-repo workspace.
Use Composer 2 (Cursor's in-house frontier coding model, built on Moonshot's Kimi K2.5) effectively for multi-file generation, refactoring, and scaffolding tasks.
Configure agent mode for autonomous multi-step coding tasks with appropriate guardrails.
Run Background Agents locally and Cloud Agents (Cursor-hosted or self-hosted) for off-thread, multi-tasking development.
Hand off work seamlessly between local and cloud agents and back to the IDE.
Use Design Mode for visual UI iteration on real tasks.
Write and maintain project-level and user-level rules files to standardize AI behavior.
Manage context explicitly: @codebase, @file, @docs, @web, and pinned context.
Integrate Cursor into team workflows with shared configuration and version-controlled rules.
Apply privacy and security best practices for codebases with sensitive data, including self-hosted cloud agent deployment.

Training Materials

Comprehensive courseware is distributed online at the start of class. All students receive a downloadable MP4 recording of the training.

Software Requirements

Cursor (Pro or Business plan), Git, and the student's preferred language runtime for labs.

Training Topics

Cursor Architecture and Philosophy

Cursor vs. VS Code Copilot: key design differences
Cursor 3.0's unified human-and-agent workspace
Model configuration: choosing Claude, GPT, and Gemini backends
Running the OpenAI Codex IDE extension inside Cursor as a complementary surface
Privacy modes and codebase indexing controls

Chat and Inline Edit

Inline code generation and edit shortcuts
Chat with @ context references
Apply vs. preview edit workflows

Composer 2: Multi-File Generation

Composer 2 as Cursor's in-house frontier coding model (built on Moonshot's Kimi K2.5 with continued pretraining + RL; 200+ tok/s with custom GPU kernels)
Opening and driving Composer sessions
Scoping Composer to a feature, module, or refactor
Reviewing and accepting multi-file diffs
Iterating with follow-up prompts
Auto mode: when Composer 2 picks the model for you

Agent Mode

Enabling and configuring agent mode
Autonomous task execution and step-by-step review
Interruption, rollback, and scope limits
Combining agent mode with tests for validation

Background Agents and Cloud Agents

Background Agents: parallel local agent sessions while you keep coding
Cloud Agents: isolated VMs with full development environments
Self-hosted cloud agents for sensitive codebases
Multi-repo workspace and seamless local/cloud handoff

Design Mode

Design Mode: iterate on UI visually with the agent driving rendered changes
Pairing Design Mode with rules files for consistent styling
When Design Mode replaces hand-coded UI scaffolding

Rules Files and Project Configuration

User-level rules for personal preferences
Project-level .cursorrules: format, conventions, and constraints
Version-controlling rules alongside code
Combining rules with system prompts
Pairing Cursor rules with a GitHub Spec Kit constitution and spec-driven workflow (covered in depth in the dedicated Spec Kit course)

Context Management

@codebase semantic search and relevance tuning
@docs for third-party library documentation
@web for up-to-date reference material
Pinning critical context files

Team Workflows and Security

Shared rules configuration across a team
Sensitive file exclusion and .cursorignore
Self-hosted cloud agent deployment patterns
Audit trail and team policy considerations

Workshop

Multi-file Composer lab
Background and Cloud Agent task lab
Rules file authoring exercise
Q&A session