Skip to content

pranathi-cursor/cursor-sdlc-workshop

 
 

Repository files navigation

Cursor SDLC Workshop

Experience the Software Development Lifecycle Firsthand

A hands-on workshop designed for Cursor's GTM team (AEs, Field Engineering, ADMs) to understand the Software Development Lifecycle by actually living it—not just learning about it.


Purpose

This workshop exists to solve a critical enablement gap: GTM teams often explain developer workflows without ever having experienced them.

By the end of this workshop, participants will:

  1. Understand customer pain points firsthand — not from a slide deck, but from direct experience
  2. Contextualize Cursor's products within the SDLC — know exactly where Plan Mode, BugBot, Hooks, Skills, and Agents fit
  3. Leave with authentic stories to tell — real struggles that translate into compelling customer conversations

Why This Matters

"AEs are less technical than most Cursor customers—that's the point."

When less technical team members struggle with git, debugging, or code review, they surface friction that experienced developers take for granted. These insights are valuable for:

  • Identifying onboarding pain points
  • Understanding where Cursor adds the most value
  • Speaking authentically about developer challenges

Workshop Structure

Section Name Duration Team Size Focus
0 Pre-Work Setup 10 min Individual Install Cursor, CLI, brew, git
1 Greenfield Project 45 min 5 people Build from scratch through full SDLC
2 Legacy Codebase 30 min 5 people Add features using BugBot, Hooks, Skills, Agents
3 Real-World Chaos 35 min 20+ people CursorFlix Clash — Netflix clone competition
Closing 10 min All Connect experience to customer conversations

Total Duration: ~2 hours


The SDLC Journey

Participants experience every stage of the Software Development Lifecycle:

Plan  →  Design  →  Develop  →  Test  →  Review  →  Deploy

Cursor Features by SDLC Stage

Stage What Happens Cursor Feature
Plan Define requirements, break down tasks Plan Mode, multi-model reasoning
Design Architecture decisions, structure Context-aware suggestions
Develop Write code, implement features Tab completion, Cmd+K, Agent mode
Test Validate functionality Test generation, bug detection
Review Code review, quality checks BugBot, PR analysis
Deploy Ship to production CI/CD assistance

The Git Sandwich (Agent-First)

A core teaching of this workshop is the Git Sandwich pattern—the idea that all productive work is wrapped in git operations. But participants don't need to memorize commands. They tell the Cursor agent what they want.

Since participants aren't collaborators on the workshop repo, they fork it first (the agent handles this).

Fork Setup (One Time)

┌─────────────────────────────────────────────────────────┐
│  Ask Cursor:                                            │
│  "Fork this repo [URL], clone my fork,                  │
│   and create a branch called [team]/setup"              │
└─────────────────────────────────────────────────────────┘

The Git Sandwich (Every Task)

┌─────────────────────────────────────────────────────────┐
│  START — Ask Cursor:                                    │
│  "Sync my fork with upstream, pull the latest,          │
│   and create a branch called [name]"                    │
└─────────────────────────────────────────────────────────┘
                          ↓
┌─────────────────────────────────────────────────────────┐
│  WORK                                                   │
│  (Tell Cursor what to build — this is the fun part)     │
└─────────────────────────────────────────────────────────┘
                          ↓
┌─────────────────────────────────────────────────────────┐
│  FINISH — Ask Cursor:                                   │
│  "Commit my changes with message '[description]',       │
│   push to my fork, and open a PR to the original repo"  │
└─────────────────────────────────────────────────────────┘
                          ↓
┌─────────────────────────────────────────────────────────┐
│  SYNC — Ask Cursor:                                     │
│  "Sync my fork with upstream and pull the latest"       │
└─────────────────────────────────────────────────────────┘

This pattern is reinforced throughout the workshop slides.


Repository Contents

cursor-sdlc-workshop/
├── README.md                        # This file
├── slides/
│   └── cursor-sdlc-slides.html      # Interactive HTML slide deck
├── docs/
│   └── cursor-sdlc-workshop.docx    # Comprehensive facilitator guide
├── startup-stress-test/             # Example app: Startup Idea Stress Tester
│   └── src/
│       ├── App.jsx                  # Main app with form/results toggle
│       ├── components/              # Scorecard, RadarChart, ReportSummary
│       └── data/                    # Scoring engine and industry data
└── teams/                           # Created during workshop
    └── [team-name]/
        ├── members/
        │   └── [name].md
        └── prd.md

Slides

cursor-sdlc-slides.html — An interactive HTML/CSS slide deck with:

  • Arrow key navigation (← →) or spacebar
  • Progress bar
  • Color-coded sections
  • Visual "Git Sandwich" diagrams
  • Checklists with code blocks

To present: Open in any modern browser. Works offline.

Facilitator Guide

cursor-sdlc-workshop.docx — A detailed document containing:

  • Executive summary and purpose
  • Audience breakdown and roles
  • Section-by-section instructions
  • SDLC-to-Cursor feature mapping
  • Closing discussion guide

Startup Stress Tester

startup-stress-test/ — A React app demonstrating what can be built with Cursor:

  • Input form for startup metrics (TAM, competition, feasibility, risks, etc.)
  • Scorecard with letter grades (A-F) for each dimension
  • Radar chart (pure SVG) visualizing strengths and weaknesses
  • Report summary with verdict, risk flags, and recommendations

To run: cd startup-stress-test && npm install && npm run dev


Section Details

Section 0: Pre-Work Setup

10 minutes · Individual

Participants install their development environment:

  • Download & install Cursor
  • Install Cursor CLI (agent command)
  • Use Cursor to install brew and git
  • Verify installations

Key insight: Using AI to solve setup problems from the very first step.


Section 1: Greenfield Project

45 minutes · Teams of 5

Teams build a simple application from scratch.

Phase 1 — Plan (10 min)

  • Fork repo, clone your fork, create team folder
  • Write PRD with MVP + 5 tasks (one per person)
  • Practice full git workflow: fork → branch → work → PR → merge

Phase 2 — Design (10 min)

  • ONE person creates base MVP
  • Team watches and agrees on approach
  • Merge foundation for everyone to build on

Phase 3 — Develop (20 min)

  • EACH person implements their assigned task
  • Use Plan Mode to architect their section
  • Create individual branches and PRs

Phase 4 — Test (5 min)

  • Merge all PRs
  • Run complete application
  • Fix integration issues together

Key insight: Cursor assists at every stage of the SDLC, not just code writing.


Section 2: Legacy Codebase

30 minutes · Teams of 5

Teams receive another team's project and must navigate unfamiliar code.

Step 1 — Enable Advanced Features (10 min)

  • Turn on BugBot (using docs)
  • Add a Hook
  • Add a Skill
  • Add an Agent

Participants jot down onboarding feedback for the eng team.

Step 2 — Understand the Code (5 min)

  • Use Ask Mode to explore unfamiliar codebase
  • Figure out how to run the project

Step 3 — Add a Feature (10 min)

  • Implement a small visible feature
  • Watch BugBot review the PR

Key insight: Advanced Cursor features make unfamiliar code dramatically easier.


Section 3: Real-World Chaos

35 minutes · 4 teams of 20+

CursorFlix Clash — Teams compete to transform an ugly movie website into a Netflix clone.

The Rules:

  1. No PR over 500 lines (auto-rejected)
  2. Everyone must contribute 1+ merged PR
  3. Live deploy to big screen

Transformation:

Before After
Times New Roman Netflix dark theme
HTML table with border="1" Card grid with hover
Gray "[no image]" boxes Real movie posters
"Here are some movies" Hero, search, My List

Key insight: SDLC discipline scales. Teams that plan and coordinate outperform those that don't.


Key Takeaways

By workshop end, participants can:

  1. Speak authentically about SDLC pain points because they experienced them
  2. Know where Cursor features fit in the development workflow
  3. Tell the before/after story that resonates with customers
  4. Explain why process matters at scale to engineering leaders

The Story to Tell

"Cursor isn't just a code editor—it's an engineering force multiplier across the entire SDLC."

From planning to deployment. From greenfield to legacy. From small teams to enterprise scale.


For Facilitators

Before the Workshop

  • Ensure all participants have Cursor downloaded
  • Set up the workshop repository
  • Prepare reviewer assignments (FE, Ryan P, Ryan S)
  • Test CursorFlix starter app deployment
  • Set up big screen for Section 3

During the Workshop

  • Circulate during Section 1 to help with git confusion
  • Watch for "aha moments" when people compare GitHub vs local folders
  • Collect onboarding feedback from Section 2
  • Keep energy high during Section 3 competition

After the Workshop

  • Compile onboarding feedback for eng team
  • Share photos/screenshots of CursorFlix transformations
  • Follow up with participants on customer conversation applications

Success Metrics

  • All participants complete at least one merged PR
  • Teams successfully deploy working applications
  • Participants can articulate where each Cursor feature fits in SDLC
  • Onboarding feedback collected and delivered to eng team

Contributing

Found an issue or have an improvement?

  • Open an issue describing the problem
  • Submit a PR with your fix
  • Request review from workshop maintainers

License

Internal Cursor use only. Not for external distribution.


Now go tell the story.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • JavaScript 70.3%
  • HTML 17.7%
  • CSS 11.3%
  • Shell 0.7%