By Panaversity -- Learn to Plan Before You Build
Welcome! This package contains all the practice materials, scenario files, and instructions for 24 exercises + 3 capstone projects designed to teach you Spec-Driven Development (SDD) using Claude Code and Cowork. No coding required -- these exercises use real-world planning and problem-solving scenarios.
SDD is a four-phase workflow that transforms vague ideas into clear, executable plans:
- Research -- Gather context from multiple angles before deciding anything
- Specification -- Write a clear document describing what you're building and why
- Refinement -- Interview stakeholders to find gaps and ambiguities
- Implementation -- Break the spec into delegatable tasks and execute
You'll practice each phase individually, then combine them in full end-to-end cycles.
sdd-exercises/
├── README.md <- You are here
├── EXERCISE-GUIDE.md <- Full pedagogical guide
├── module-1/ -- Spec vs. Vibe: Seeing the Difference
│ ├── exercise-1.1-event-gone-wrong/ (botched event scenario)
│ ├── exercise-1.2-side-by-side-test/ (vague vs. spec comparison)
│ └── exercise-1.3-vibe-to-spec-rescue/ (rescue a failed plan)
│
├── module-2/ -- Research Phase: Gathering Context
│ ├── exercise-2.1-multi-angle-investigation/ (3-perspective research)
│ ├── exercise-2.2-source-reliability-audit/ (evaluate source quality)
│ └── exercise-2.3-research-summary/ (raw notes to structured doc)
│
├── module-3/ -- Writing Your First Spec
│ ├── exercise-3.1-home-renovation-spec/ (kitchen renovation)
│ ├── exercise-3.2-charity-fundraiser-spec/ (fundraiser planning)
│ └── exercise-3.3-spec-from-chaos/ (voice memos to spec)
│
├── module-4/ -- Constraints & Success Criteria
│ ├── exercise-4.1-missing-guardrails/ (find missing constraints)
│ ├── exercise-4.2-measurable-vs-vague/ (rewrite vague criteria)
│ └── exercise-4.3-constraint-stress-test/ (adversarial scenario testing)
│
├── module-5/ -- Refinement: Finding What's Missing
│ ├── exercise-5.1-interview-challenge/ (Claude interviews you)
│ ├── exercise-5.2-ambiguity-hunter/ (find more gaps than Claude)
│ ├── exercise-5.3-stakeholder-perspectives/ (3 viewpoints, 3 gap sets)
│
├── module-6/ -- Task Breakdown & Delegation
│ ├── exercise-6.1-dependency-map/ (tasks + dependencies)
│ ├── exercise-6.2-atomic-task-writer/ (rewrite vague tasks)
│ └── exercise-6.3-delegation-simulation/ (assign to 3 team members)
│
├── module-7/ -- Full SDD Cycle: End to End
│ ├── exercise-7.1-community-newsletter/ (full 4-phase cycle)
│ ├── exercise-7.2-office-move-planner/ (full cycle, complex)
│ └── exercise-7.3-product-launch-playbook/ (full cycle, most complex)
│
└── module-8/ -- Capstone Projects
├── capstone-A-wedding-planner/ (multi-stakeholder SDD)
├── capstone-B-small-business-launch/ (business operations SDD)
└── capstone-C-personal-goal-system/ (apply SDD to your own life)
- Open Claude Desktop app on macOS
- Switch to the Cowork tab
- Point Cowork at the exercise folder you want to work on
- Read the INSTRUCTIONS.md file in the exercise folder
- Start describing your task to Claude
- Open your terminal
- Navigate to the exercise folder:
cd module-1/exercise-1.1-event-gone-wrong - Launch Claude Code
- Read the INSTRUCTIONS.md and start working
Week 1 (Why Specs Matter): Exercises 1.1, 1.2, 1.3 Week 2 (Research): Exercises 2.1, 2.2, 2.3 Week 3 (Specifications): Exercises 3.1, 3.2, 3.3 Week 4 (Constraints): Exercises 4.1, 4.2, 4.3 Week 5 (Refinement): Exercises 5.1, 5.2, 5.3 Week 6 (Delegation): Exercises 6.1, 6.2, 6.3 Week 7 (Full Cycle): Exercises 7.1, 7.2, 7.3 Week 8 (Capstone): Choose one capstone project
Use this for every exercise:
- Research -- What context do I need before deciding anything?
- Specify -- What exactly am I building, for whom, and with what constraints?
- Refine -- What did I miss? What's ambiguous? What would stakeholders question?
- Delegate -- How do I break this into clear, independent tasks?
- Execute -- Run the tasks and verify each output
- Reflect -- What would I change about my spec or process?
| Criteria | 1 (Beginner) | 2 (Developing) | 3 (Proficient) | 4 (Advanced) |
|---|---|---|---|---|
| Research Quality | Skipped research | Surface-level search | Multi-angle investigation | Identified conflicting sources |
| Spec Completeness | Missing sections | Has basics, gaps remain | All sections, clear criteria | Anticipates edge cases |
| Constraint Clarity | No constraints listed | Some constraints | Measurable constraints | Stress-tested constraints |
| Refinement Depth | No refinement | Found 1-2 gaps | Found 5+ ambiguities | Stakeholder-tested |
| Task Breakdown | Monolithic tasks | Some breakdown | Atomic, delegatable tasks | Dependencies mapped |
| Iteration | Single attempt | One revision | Multiple refinements | Systematic improvement |
Built for Panaversity's AI-Native Development Curriculum No programming knowledge required -- just clear thinking and structured planning.