data-broker-removal

name	data-broker-removal
description	Scaffolds and manages a full-stack data broker removal application. Use when building an Incogni-like service with automated opt-out workers, a user dashboard, and a REST API. Supports project initialization, broker catalog management, and Playwright worker generation.
compatibility	Requires Node.js 20+, Docker, and Playwright.

Data Broker Removal Skill

Scaffolds a monorepo with three packages:

• packages/api — Fastify REST API + PostgreSQL
• packages/worker — Playwright automation workers + BullMQ job queue
• packages/dashboard — Next.js user dashboard
• packages/shared — TypeScript types and database schema

Commands

/skill:data-broker-removal init <project-name>

Scaffolds a new project in ./<project-name>/.

./scripts/init.sh <project-name>

Sets up:

• Root package.json with pnpm workspaces + Turborepo
• docker-compose.yml (PostgreSQL 16 + Redis 7)
• Each package with dependencies, TypeScript config, and starter code
• Database migration schema
• A sample broker catalog with 5 entries

/skill:data-broker-removal add-broker

Interactive prompt (or pass JSON) to append a broker to the catalog.

./scripts/add-broker.sh
# or with JSON
./scripts/add-broker.sh '{"name":"Spokeo","domain":"spokeo.com","method":"form","optOutUrl":"https://www.spokeo.com/opt_out","requiredFields":["fullName","email","address"],"selectors":{"name":"[data-testid=optout-name]","email":"[data-testid=optout-email]","submit":"[data-testid=optout-submit]"},"verificationType":"email_link","region":["US"],"category":"people-search"}'

/skill:data-broker-removal generate-worker <broker-id>

Generates a Playwright worker script for a specific broker using its catalog entry.

./scripts/generate-worker.sh <broker-id>

Outputs packages/worker/src/brokers/<broker-id>.ts.

Project Structure After Init

<project-name>/
├── docker-compose.yml
├── turbo.json
├── package.json
├── packages/
│   ├── shared/
│   │   └── src/
│   │       ├── schema.sql
│   │       ├── types.ts
│   │       └── broker-catalog.json
│   ├── api/
│   │   └── src/
│   │       ├── server.ts
│   │       ├── db.ts
│   │       ├── routes/
│   │       │   ├── users.ts
│   │       │   ├── brokers.ts
│   │       │   └── requests.ts
│   │       └── jobs/
│   │           └── scheduler.ts
│   ├── worker/
│   │   └── src/
│   │       ├── worker.ts
│   │       ├── scheduler.ts
│   │       ├── browser-pool.ts
│   │       └── brokers/
│   │           └── index.ts
│   └── dashboard/
│       └── src/
│           ├── app/         # Next.js App Router — note `src/app/` prefix
│           │   ├── page.tsx
│           │   ├── layout.tsx
│           │   └── dashboard/
│           │       └── page.tsx
│           └── lib/
│               └── api.ts

Running the Project

cd <project-name>

# Start infrastructure
pnpm infra:up

# Run migrations
pnpm db:migrate

# Start API
cd packages/api && pnpm dev

# Start worker (in another terminal)
cd packages/worker && pnpm dev

# Start dashboard (in another terminal)
cd packages/dashboard && pnpm dev

Design Decisions

• Monorepo: Separate concerns; workers crash independently from API.
• BullMQ over Redis: Retry with backoff, concurrency controls, dead-letter queues.
• Playwright in Docker: One container per job to isolate sessions, cookies, and fingerprints.
• PostgreSQL: ACID for request state; JSONB for broker configs.
• Next.js: Familiar, Incogni uses it.