Sorry Not Sorry (Pickleball Ledger)

Companion project for an upcoming talk. This app tracks pickleball games, scores, locations, and money won/lost between players.

Why the Name?

In pickleball, it’s common to get hit by the ball or see a ball clip the net and drop in. Players often say “sorry not sorry” in a playful, ironic way: you’re being polite, but secretly not sorry at all.

Overview

What it does

• Tracks singles and doubles matches.
• Records game location and final score.
• Supports live scoring or post-game entry.

Tech stack

• Frontend: Next.js + shadcn-admin UI
• Backend: Supabase

Future ideas

• Optional DUPR integration: https://backend.mydupr.com/swagger-ui/index.html

Getting Started

Prerequisites

• Node.js 18+ (or 20+)
• npm (or your preferred Node package manager)

Setup

1. Install dependencies:
   • npm install
2. Create your environment file:
   • cp .env.local.example .env.local
   • Fill in Supabase keys and any required values.
3. Start the dev server:
   • npm run dev
4. Open http://localhost:3000

Common Scripts

• npm run dev — run the app locally
• npm run build — create a production build
• npm run start — run the production build
• npm run lint — lint the project

Project Structure

• src/ — application code
• public/ — static assets
• supabase/ — migrations and seed data

Talk Steps (from slide deck)

1) Initial research / planning phase

• Treat planning as critical. Maintain an instruction file that includes:
  • High-level project organization
  • External dependencies and constraints
  • Positive and negative examples
• Watch token usage; keep prompts scoped and disciplined.
• Use helpers like /init-project and /create-prd.
• Be judicious about how much code gets generated.

2) Demo flow

• Run /create-prd and walk through the generated documentation.
• Emphasize security and architecture standards.
• Show the live context window.
• Show the validation workflow end-to-end.
• Use hooks to run linting/formatting after each turn.
• Build and reuse prompt/instruction libraries over time.

3) Task planning (part 2)

• Break work into linear, well-scoped tasks.
• Include testing requirements (TDD when possible).
• Require docs and organization updates as part of each task.
• Integrate into existing structure; avoid unnecessary new files.

4) Validation: “trust (kinda) but verify”

• Use /validation:validate or /test.
• For bugs: /debug plus clear repro steps.
• Maintain unit tests plus an e2e test (Playwright or Chrome MCP).

5) Implementation (part 2)

• Combine RAG, memory, prompt engineering, and task management as needed.
• Memory can be hardcoded logs or a tool like claude-mem.
• Task management can be off-the-shelf tools or custom workflows.

6) Validation (agentic + dev)

• Agentic validation: unit/integration tests, plus MCP integrations.
• Chrome MCP is powerful but can be tedious unless using a -yolo mode.
• Dev validation: review code; if it’s too much to review, the task was too big.
• Always document manual validation steps.

7) Example: Pickleball game tracker

• Requirements: basic auth, teams, leagues support.
• /skills-brainstorming can help generate ideas.
• Start with a plan.md outlining stack, user profile, and MVP milestones.
• Use /superpowers:brainstorm or /create-prd to build a detailed PRD.
• Highlight MVP checkboxes, docs, API references, and codebase structure.
• Leave as little ambiguity as possible.

8) Debugging

• Same process, smaller loops.
• Use a debug-focused skill/profile (e.g., “smart debug”).
• TDD preferred; can be skipped for tiny fixes.

9) Common gotchas

• Illusion of productivity: lots of prompts, little progress.
• Tools change rapidly; a working method may break later.
• “Copy-paste bot” loop: fix one error, create another.
• The escape hatch is to zoom out and return to planning.

10) Active development loop

• When validation reveals a missed requirement, update the docs.
• Core loop:
  • /core_piv_loop:prime
  • /core_piv_loop:plan-feature
  • /core_piv_loop:execute
  • /validation:validate
  • /validation:code-review
  • /validation:code-review-fix
  • /validation:execution-report
  • /validation:system-review
• Reference: https://github.com/thedotmack/claude-mem This is a Next.js project bootstrapped with create-next-app.

Getting Started