Command Claude Code
Check application errors, slow endpoints, database performance, and request volumes, then create tasks.
Command Claude Code
Check application errors, slow endpoints, database performance, and request volumes, then create tasks.
Command Claude Code
Full security review — GitHub alerts, static code analysis, live app check — then a task per finding.
Command Claude Code
Update Dark Flow to the latest version.
Command Claude Code
Full SEO check — technical + on-page audit plus Google Search Console data, then create tasks with concrete fixes.
Command Claude Code
Submit the product to the directory catalog in /.darkflow/directories.csv through a real browser, and track what has already been submitted.
Command Claude Code
Update Dark Flow project settings: language and/or main branch.
Command Claude Code
Check the site is really up (DNS, HTTP status, page body); file an auto-approved critical task when it is not.
Command Claude Code
Measure Core Web Vitals with a local Lighthouse run; file a task when a metric lands in the 'poor' band.
Skill Claude CodeCodex
Query self-hosted OpenPanel analytics — funnel counts, unique users, breakdowns by os/country/path, raw events. Use whenever analytics numbers are needed: funnel conversion, visitor counts, event counts over a window, period-over-period deltas, "how many signups/checkouts/sessions", or any /darkflow:analytics-review…
Plugin Claude Code
A marketplace plugin for launching a product through the AI Garage method. It covers a sequence from validating an idea and preparing the product to deployment, traffic, analytics, launch metrics, and payments.
Plugin Claude Code
A workflow plugin for launching a product with the AI Garage method, from an initial idea through payments. Each team reads project documents, completes one step, records the result, and suggests the next command.
Agent
A research agent for finding competing products that show signs of customer demand and checking whether they are credible opportunities.
Agent
A deployment agent for an AI Garage stack: Next.js, GitHub, Vercel, Neon, and Clerk. It handles the technical setup while requiring confirmation before external or irreversible actions.
Command
A website analytics setup that uses Plausible, PostHog, or both to track visitors, registrations, and purchases. Plausible focuses on simple privacy-friendly traffic data, while PostHog also tracks product actions and user journeys.
Command
A guided command for choosing one marketing channel using the target audience, buying intent, available time, and budget.
Command
A deployment command for connecting a web app to Neon, a hosted PostgreSQL database, and Clerk, a service for user sign-in and account management.
Command
A deployment command for publishing a website through GitHub and Vercel, then connecting a custom domain. GitHub stores the code, while Vercel builds and hosts the site.
Command
A launch checklist for taking the first real step in one chosen traffic channel, such as search ads, social media, or SEO.
Command
A command for examining a product funnel and unit economics. It covers CAC, the cost to acquire a customer; LTV, the revenue expected from a customer; retention; and UTM campaign tracking.
Command
A payment setup for a web product using Polar, a service that handles checkout, taxes, and subscription management for sellers.
Command
A command that creates a Next.js application, a React web framework, with TypeScript, Tailwind CSS, and shadcn/ui. It builds an initial product landing page from the project's brief and records the setup in documentation.
Command
A command that collects the basic product brief before coding: the target user, their problem, one feature, its value, possible names, and landing-page copy.
Command
The first command in an AI Garage product-research workflow, where you study existing products instead of inventing a market from scratch. It searches for competitors with signs of demand and creates a shortlist in the project's docs folder.
Command
A progress-check command for a project pipeline. It reads docs/STATE.md and compares its checklist with the project’s documented output files.