Launch three SKUs. Scale Peak Flow to 2,000 units a month.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

Extract the AI-suggested "Looking for specific info?" / "Ask Alexa" (Rufus) questions from Amazon product pages for one or many ASINs/URLs, and save them to a timestamped JSONL + CSV. Use this whenever the user wants to scrape, collect, or extract the suggested / Alexa / Rufus questions, the "Looking for specific info" chips, or the per-product question prompts from Amazon listings. Trigger even if the user just pastes a list of Amazon ASINs or /dp/ URLs and asks for "the questions" — that is this skill. Drives the user's own logged-in browser through the Playwright MCP extension and is built to stay human-paced so Amazon does not flag it as a bot.

Scrape Amazon customer reviews for an ASIN or a /product-reviews/ URL through the user's own logged-in browser (Playwright MCP extension), newest first, and save them to a timestamped JSONL + CSV. Use whenever the user wants to scrape, collect, pull, track, or monitor Amazon reviews for a product / ASIN, or pastes a product-reviews URL and asks for "the reviews" or "recent reviews". Drives the real Chrome Beta session and stays human-paced so Amazon does not flag it as a bot. Checking which reviews Amazon removed is the review-audit skill.

Check which top-100 Amazon bestsellers in a niche use the new two-part title (main title + second "Item Highlight" line) and push the result to the Titles Google Sheet. Use when the user says "/amazon-titles", "amazon titles", "check the split titles", "who has the new title format", "pull competitor titles", or wants main + second titles for a bestseller niche. Interactive - asks which niches, then whether to reuse or re-pull the bestseller list and the Apify product data before spending credits.

Pull an Amazon Best Sellers top-100 category list from Apify (junglee amazon-bestsellers actor) into data/apify/bestsellers/<niche>.csv. Use when the user says "/apify-bestsellers", "pull the bestseller list", "top 100 for <niche>", "refresh the bestsellers", "get a new best seller list", or gives an Amazon Best Sellers category URL. This is the single source of bestseller lists - asin-reviews, amazon-titles and comps work read from it.

Pull full Amazon product-page data for a list of ASINs from Apify (junglee Amazon-crawler - title, Item Highlight, bullets, price, rating, monthly volume, images, A+ content, attributes) into data/apify/product-details/<label>/. Use when the user says "/apify-product-details", "pull product details", "get the product data for these ASINs", "scrape these ASINs", "pull the products for the top 100", or asks how fresh the product data is. This is the single source of product data - comps-listing and amazon-titles read from it.

Pull Amazon reviews for arbitrary ASINs (Apify axesso actor), push to a Google Sheet, and cross-check reviewer profiles between competitor ASINs and STRIV reference ASINs. Use for reviewer-overlap analysis or any ad-hoc "get me reviews for ASIN X" request. Bestseller lists come from the apify-bestsellers skill.

Use when someone asks for an AIOS audit, asks to score their setup against the Four Cs, or says "is my AIOS working" / "audit my setup" / "find gaps in my AIOS". Produces a Four-Cs scoreboard with top-3 fixes ranked by leverage.

Use when Norbert wants his daily briefing — "what's for today", "morning brief", "daily brief", "what's on today", "start my day", or "/briefing". Merges today's Google Calendar with Todoist into one prioritized rundown.

Restate the last message in plain human language, with no jargon.

Add word-highlight captions to a video in the STRIV/CapCut style — Proxima Nova Black, all caps, dark rounded box that snaps to each word as it is spoken. Free and fully local: mlx-whisper transcribes, libass renders. Use when the user says "/captions", "add subtitles", "caption this video", "auto subtitles", or drops a video path and asks for captions. Replaces the paid CapCut auto-caption feature.

Documentation reference for using Browser Use Cloud — the hosted API and SDK for browser automation. Use this skill whenever the user needs help with the Cloud REST API (v2 or v3), browser-use-sdk (Python or TypeScript), X-Browser-Use-API-Key authentication, cloud sessions, browser profiles, profile sync, CDP WebSocket connections, stealth browsers, residential proxies, CAPTCHA handling, webhooks, workspaces, skills marketplace, liveUrl streaming, pricing, or integration patterns (chat UI, subagent, adding browser tools to existing agents). Also trigger for questions about n8n/Make/Zapier integration, Playwright/ Puppeteer/Selenium on cloud infrastructure, or 1Password vault integration. Do NOT use this for the open-source Python library (Agent, Browser, Tools config) — use the open-source skill instead.

End-user control of cmux topology and routing (windows, workspaces, panes/surfaces, focus, moves, reorder, identify, trigger flash). Use when automation needs deterministic placement and navigation in a multi-pane cmux layout.

cmux package architecture, refactor layering, dependency inversion, file organization, DocC documentation, package design discipline, testability, and Swift 6 concurrency rules. Use before adding or meaningfully rewriting Swift files, Swift packages, coordinators, services, repositories, or public package APIs.

Backend TypeScript and Cloud VM development rules for cmux. Use when editing web/app/api, web/services, backend scripts, Cloud VM lifecycle, provider integrations, Postgres, Stack Auth pricing gates, migrations, or provider image build scripts.

Stripe checkout, pricing, subscription, Pro plan, webhook, and entitlement runbook for cmux billing work. Use when editing or debugging billing, pricing, Stripe Checkout, subscription recording, Pro plan status, webhooks, entitlement metadata, or pricing dev/prod tooling.

End-user browser automation with cmux. Use when you need to open sites, interact with pages, wait for state changes, and extract data from cmux browser surfaces.

Build a custom cmux sidebar from a plain-language request. Use when the user asks for a custom sidebar, a sidebar that shows their workspaces/tabs/PRs/clock, a vibe-coded sidebar, or anything involving files in ~/.config/cmux/sidebars/. Covers authoring the interpreted SwiftUI-style file, enabling the beta flag, selecting it, and iterating with hot reload.

Customize cmux for an end user. Use when changing cmux.json actions, custom commands, workspace layouts, plus-button behavior, surface tab bar buttons, Command Palette entries, Dock controls, sidebar and app settings, shortcuts, notifications, browser routing, examples-library presets, or Ghostty-backed terminal preferences.

Debug logging, Debug menu, runtime pitfalls, typing-latency-sensitive paths, SwiftUI list snapshot boundaries, OS-version repros, and local visual iteration for cmux. Use when adding debug probes, diagnosing UI/runtime issues, touching terminal rendering, tab/sidebar list views, drag/drop UTTypes, or using the Debug menu.

Contributor workflow rules for cmux setup, Xcode project normalization, tagged sidebar ExtensionKit development, and dev builds. Use when setting up the cmux repo, changing Xcode project files, adding sidebar extensions, or working with tagged debug builds.

Run end-user cmux diagnostics. Use when cmux hooks, notifications, session restore, settings, browser automation, socket access, CLI control, or agent resume behavior is not working, or when the user asks for a cmux health check, doctor report, or support-safe debug summary.

Ghostty submodule and GhosttyKit workflow rules for cmux. Use when modifying the ghostty submodule, rebuilding GhosttyKit.xcframework, updating the parent submodule pointer, or documenting fork conflict notes.

Guide and apply cmux keyboard shortcut customization. Use when the user asks to customize, rebind, unbind, reset, audit, or create shortcut templates for cmux, including tmux-style, Vim-style, terminal-first, browser-heavy, iTerm/Terminal-like, or agent-triage layouts.

Localization rules and audit workflow for cmux UI strings, settings rows, menus, shortcuts, schema/config text, docs, command/help text, alerts, tooltips, and web messages. Use whenever changing user-facing text.

Open markdown files in a formatted viewer panel with live reload. Use when you need to display plans, documentation, or notes alongside the terminal with rich rendering (headings, code blocks, tables, lists).

cmux release workflow, version bumping, changelog updates, pretag guard, release tags, and release asset expectations. Use when preparing or troubleshooting a cmux release.

View and edit cmux settings in ~/.config/cmux/cmux.json. Use when the user wants to change cmux preferences (appearance, sidebar, notifications, automation, browser, shortcuts), set a value by JSON path, validate the file, open it in an editor, or look up which keys cmux recognizes. Triggers on '/cmux-settings', 'change cmux setting', 'set <something> in cmux', 'cmux config', 'cmux.json', or 'rebind a cmux shortcut'.

Shared behavior and mutation-path rules for cmux. Use when a behavior is exposed through multiple entrypoints such as keyboard shortcuts, command palette, context menu, CLI, settings, debug menu, optimistic UI, or tests that previously missed a bug.

Socket command threading and focus policy for cmux CLI/socket work. Use when adding or changing socket commands, CLI commands, telemetry commands, focus/select/open/close/send-key behavior, or automation that could steal app focus.

cmux testing rules for Swift Testing, test target compilation, and package/refactor validation. Use when adding or changing tests, touching package/refactor code, or deciding whether reload.sh is enough validation.

Work inside the current cmux workspace and terminal. Use for cmux workspace, current workspace, caller surface, panes, surfaces, socket targeting, and non-interfering cmux automation.

Publish a competitor set's gallery images and A+ content to the self-hosted tldraw comps board (Cloudflare Pages + R2, Access-gated) for side-by-side visual comparison. Use when the user says "comps board", "/comps-board <label>", "put <label> on the board", "visualize comps", "upload the images to the board", or wants the Miro-style view without Miro. Consumes output/comps/<label>/images/ and output/comps/<label>/aplus/ from the comps-listing pipeline. Free replacement for miro-upload.

Build a competitor "comps listing" CSV (one row per Amazon ASIN: brand, title, bullets, formula, price, rating, monthly volume, per-gallery-image OCR text, and a messaging-angle summary) from a list of ASINs/URLs. Use when the user says "comps listing", "/comps-listing", "build comps for <product>", "competitor comparison CSV", "comp sheet for these ASINs", or pastes a batch of Amazon ASINs to compare. Scrapes via Apify + enriches with Gemini vision. Output lands in data/comps/<label>/comps-listing.csv.

Builds the customer value page (LTV, customer lifetime value) from real Amazon.com orders. What a new customer spends on the first order and after 30, 90, 180 days and a year, who buys again and when, Subscribe & Save starters vs one-time starters, cohorts by first month, customers who buy 2 products, and ad spend per new customer. One page for all products with a product switch. Use when the user says "/customer-value", "refresh the customer value page", "what is a customer worth", "LTV", "lifetime value", "repeat rate", "CAC", or asks how much we can pay for a new customer.

Rebuild and publish the AIOS mission-control dashboard (dashboard/ → Cloudflare Pages). Use when the user says "update the dashboard", "deploy the dashboard", "/dashboard", "refresh mission control", or after changes that belong on it (new skill, closed launch step, new deployment, new decision, closed loop).

Explain any topic to a total beginner as a visual HTML artifact — big pictures, very few words, one idea per screen. Use when the user says "/eli5", "explain like I'm five", "explain this like I know nothing", "ELI5 this", "make me a visual explainer", "dumb this down with pictures", or asks to understand an unfamiliar topic and wants to look at it rather than read it.

Expand or outpaint an image so it fills a larger canvas, using fal.ai Bria Expand (subject-safe) or nano-banana-pro (prettier, redraws). Use this skill whenever the user asks to "expand this image", "fill out the rest of the area", "outpaint", "extend the background", "fill the white space", "make this photo fill the frame", "widen this photo", "uncrop", "reframe to 1:1 / 9:16", or drops a Figma frame export where a photo sits in a corner with flat white padding around it. Also triggers for "extend the image beyond its borders" or "generative fill".

Choose the right fal.ai endpoint for a given task. Modality-organized catalog of production endpoint defaults, text-to-image, image-to-image, text-to-video, image-to-video, and more. Use when the user has not named a specific model, or asks "which model for X", "best endpoint for Y", "what should I use for Z".

Model-family-specific prompt craft for fal.ai endpoints. Trigger when the user mentions a specific model family by name and asks how to prompt it ("how do I prompt Kling", "GPT Image 2 prompt structure", "Happy Horse tips"), or when prompts to a routed endpoint keep coming back generic and the family's known nuances should be applied. For endpoint selection ("which model for X"), use `fal-models-catalog` instead. This skill is about how to talk to a model once it has been chosen.

Open the local image gallery (gallery/serve.py, port 5180) to browse, filter, star and trash generated images under output/, with each image's prompt and model shown. Use when the user says "/gallery", "open the gallery", "show me the images", "let me pick", "show me the takes", or "use my picks". Also offer it at the end of an image-generation session.

Generate a verified batch of human-sounding Shopify reviews for a STRIV supplement, then auto-check them for weird/false/off-avatar content. Use when the user says "generate reviews", "/generate-reviews", "make N reviews for <product>", "verify these reviews", or wants a clean/verified review CSV. Product configs live in data/generate-reviews/<slug>/ (peak-flow, peak-potential); run outputs land in output/generate-reviews/<slug>/.

Use the genmedia CLI to search, inspect, run, and manage 1200+ fal.ai model endpoints. Trigger when the user mentions "genmedia", "fal CLI", or asks to "search models", "run a model", "fetch schema", "check pricing", "upload to fal", "queue async job", "track request", or any direct interaction with the fal.ai endpoint catalog. This is the foundational skill. Every other fal.ai-related skill in this repo executes its work through genmedia commands. Use `--json` whenever the output will be parsed by an agent.

Interview the user relentlessly about a plan, design, or topic, checkpointing every answer to a brainstorm file so nothing is lost. Use when the user wants to stress-test a plan, get grilled on a design, run a brainstorm or discovery session, extract what's in their head into a doc, or says "grill me".

Read and write Google Docs.

Google Docs: Append text to a document.

gws CLI: Shared patterns for authentication, global flags, and output formatting.

Google Sheets: Read and write spreadsheets.

Google Sheets: Append a row to a spreadsheet.

Google Sheets: Read values from a spreadsheet.

Compact the current conversation into a handoff document for another agent to pick up.

Pull Helium 10 Cerebro reverse-ASIN keyword data (one new search per ASIN, no filters, Export Data > CSV) and append it to BigQuery `amazon.h10_cerebro`. Also builds the ASIN matrix page from a multi-ASIN Cerebro export (our ASIN + competitors, organic rank per ASIN per keyword). Use when the user says "/helium10-cerebro", "pull Cerebro", "run Cerebro for <ASIN>", "reverse ASIN keywords", "what keywords does <ASIN> rank for", wants Cerebro data in BigQuery, or says "ASIN matrix", "refresh the ASIN matrix", "competitor rank matrix", or drops a US_AMAZON_cerebro_<ASIN>_<date>.xlsx. On demand only - every search costs one of the plan's monthly Cerebro uses.

Pull daily Amazon keyword ranks from Helium 10 Keyword Tracker (Export > History, every tracked product) and append them to BigQuery `amazon.h10_keyword_ranks` for Connected Sheets. Use when the user says "/helium10-ranks", "pull the Helium 10 ranks", "refresh keyword ranks", "update rank tracking", or when the daily scheduled run fires. Also use to diagnose a failed run.

Control herdr from inside it. Manage workspaces and tabs, split panes, spawn agents, read output, and wait for state changes — all via CLI commands that talk to the running herdr instance over a local unix socket. Use when running inside herdr (HERDR_ENV=1).

Use weekly to find and ship one new automation. Walks the 3Ms interview — Mindset (find the candidate) → Method (scope one) → Machine (build it). Trigger on "let's level up", "what should I automate next", "find me leverage this week", or as a Friday ritual. One run = one shipped artifact.

MANUAL ONLY. Run this skill only when Norbert types /listing-composite or names it explicitly. Never auto-trigger it because a text PNG and a photo appear together, or because a request sounds like a slide composite; for open-ended composition or design ideas, call the image model directly instead. What it does: turn a licensed lifestyle photo plus a Figma text layer into finished Amazon gallery slides. Generates N background variations with GPT Image 2.5 (Sunburst/Flare) that keep the text areas dark, then adds a gradient and places the exact Figma text on top in code, so every take ships as a matched with-text / no-text pair at 3000px.

Keep Norbert's two Macs' shell + tooling setup in sync via a git-tracked install log and a copy of ~/.zshrc in this repo. Use when Norbert installs or removes a CLI tool / zsh plugin / font / shell-config change and wants it recorded, when setting up or catching up the "other" laptop, or says "log this install", "sync my other laptop", "what's different on this machine", "update my setup notes", "compare my zshrc", "did I install X here".

Generate an Amazon main product image for a STRIV product by relabeling the proven Peak Potential studio photo via fal.ai nano-banana (recolor + text-swap). Use when the user asks for a "main image", "hero image", or "listing image" for a product, wants to "relabel the tub" for a new product (Peak Vitality, Root Ritual, ...), or wants to continue the Peak Renewal / Peak Flow main-image work.

Put a competitor set's downloaded gallery images and A+ content onto a Miro board for side-by-side visual comparison. Use when the user says "upload to miro", "/miro-upload", "put <label> images on a board", "visualize comps on miro", "add <product> to the miro board", or gives a Miro board URL plus a run label. Consumes the output/comps/<label>/images/ and output/comps/<label>/aplus/ outputs produced by the comps-listing pipeline (one product per gallery ROW + one product per A+ COLUMN, ordered by bestseller rank).

Spawn 5+ Claude instances into a shared conversation room where they debate, disagree, and converge on solutions. Uses round-robin turns with parallel execution within each round. Triggers on "model chat", "multi-model debate", "agent debate", "spawn a chat room", or /model-chat. Pass a topic as the argument.

Generate and edit images using fal.ai (nano-banana). Use this skill whenever the user asks to "generate an image", "create a picture", "make artwork", "create a sprite", "edit this image", "make a variation of this", "generate a hero image", "create an icon", "make a mockup", or any task involving AI image generation or image editing/transformation. Also triggers for requests mentioning "nano-banana" directly.

Evidence-backed deep research on any question: plan search routes, retrieve in parallel, keep a per-claim evidence ledger, verify and critique (deep mode), and write a cited report to disk. Manual only: run with /nr-research <question>. Spawns ~8 to 20 agents, so it never auto-triggers.

Use on Day 1 of an AIS-OS install, when someone says "set me up", "onboard me", "let's get started", "fill in my AIOS", or has just cloned the kit. Combined wizard — runs the 7-question intake AND scaffolds the Day-1 file set at the end. Idempotent — re-run any time after editing aios-intake.md.

Remove the background from one or more images, producing transparent PNG cutouts via fal.ai BiRefNet v2. Use this skill whenever the user asks to "remove background", "remove the background", "cut out", "knock out the white", "make this transparent", "get a transparent PNG", "isolate the subject", "extract the product", or batch-remove backgrounds from a folder of images (e.g. ingredient cutouts, product shots, hero images). Also triggers for requests mentioning "rembg", "remove.bg", or "background remover".

Check which Amazon reviews for a STRIV product are still live and which Amazon removed, by opening each review's own page in the logged-in Chrome Beta. Use when the user says "/review-audit", "check removed reviews", "did Amazon delete reviews", "some reviews got removed", "run the deletion check", or "check each review" for Peak Flow / Peak Potential. Updates review_status in data/amazon-reviews/reviews.csv and the Sheets dashboard. Scraping new reviews is the amazon-reviews skill, not this one.

Build a product's ingredient science page from real, verifiable PubMed studies. Pulls studies per ingredient from Europe PMC (PubMed), scores and shortlists them, has Claude curate the picks against the product's claim vocabulary, then emits a canonical honest dataset (JSON + CSV) and a team-review workbook (the v5 page copy). Use when the user says "science page", "/science-page", "pull the studies for <product>", "PubMed studies for <ingredients>", "build the ingredient evidence page", or wants to add a new product's studies. Per-product config lives in data/science-page/<slug>/; outputs land in output/science-page/<slug>/.

Keyword research for Amazon ranking campaigns from SQP + Cerebro data. Owns the "SQP Claude Test" Google Sheet (Candidates / Shortlist / Picks tabs over a Connected Sheets extract of striv_views.sqp_cerebro_matrix), the sync that adds new keywords to it, and the loader for the hand-downloaded Brand View SQP quarterly CSV. Also builds the SQP trend page (market vs our share), the keyword watch page (keywords losing rank or share) and the brand halo page (Amazon searches for our name vs TikTok). Use when the user says "/sqp-keywords", "refresh the trend page", "refresh the keyword watch", "refresh the brand halo page", "sync the keyword sheet", "sync the Peak Flow sheet", "load the brand SQP report", "which keywords can we rank on", or wants the keyword sheet built for another product.

Spawn N agents with the same prompt (slight framing variations) to independently analyze a problem, then aggregate results by consensus. Use for decision-making, ranking options, strategic analysis, or any problem where you want to filter hallucinations and surface high-variance ideas. Triggers on "consensus", "poll agents", "stochastic consensus", "spawn N agents to analyze", "multi-agent vote", or /stochastic-multi-agent-consensus. Also triggers on phrases like "what do 10 agents think", "get multiple opinions", or "poll on [topic]".

Use when someone asks to run Storm Research, use the storm-research skill, run the STORM method on a topic, says "storm research this" / "storm report on X" / "give me a STORM briefing on X", or wants a multi-perspective, citation-verified HTML research briefing on a topic. Runs a 4-phase pipeline: five expert lenses (Practitioner, Academic, Skeptic, Economist, Historian) -> contradiction map -> synthesized HTML report -> adversarial peer review + primary-source verification. Best for topics where multiple viewpoints and fact-checked claims matter; overkill for a simple factual lookup.

Pull Amazon Subscribe & Save data from the SP-API Replenishment endpoints and push it to the S&S Google Sheet (raw tabs + a BigQuery-fed retention history + a glossary). Use when the user says "/subscribe-and-save", "pull S&S data", "refresh the S&S sheet", "subscribe and save numbers", asks about S&S retention, active subscribers, subscription units/revenue, coupon share, or the seller-funded discount tier. Also use for questions about which S&S metrics the API can and cannot give (retention history, weekly forecast). Read-only against Amazon. Data lands in data/subscribe-and-save/<date>/.

Generate the quarterly Toggl Track import CSV for STRIV Labs cofounder hours. Use when Norbert asks about time tracking, Toggl, cofounder hours, or the quarterly hours import.