lumina-studio

Cloudflare platform hub — Workers, Pages, Wrangler CLI, D1/R2/KV, Durable Objects, Workers AI, Agents SDK, AI Gateway, Vectorize, Workflows, Tunnel, WAF, Terraform/Pulumi, and production Workers best practices. Use whenever the user mentions Cloudflare, Workers, wrangler, edge functions, Durable Objects, AI agents on Workers, `@cloudflare/workers-types`, `wrangler.jsonc`, Hyperdrive, or iPix Cloudflare tasks under `tasks/cloudflare/` — even if they do not say "Cloudflare" explicitly. NOT for Cloudinary media (→ cloudinary) or Supabase Postgres/RLS (→ ipix-supabase).

iPix / FashionOS domain hub. Routes to the fashion shoot-production toolkit and operator platform skills. Use for garment/product shoots, shot lists, lighting/call sheets, or iPix-specific platform workflows.

Five-phase orchestrator for iPix / FashionOS Linear team IPI — plan, research, implement, test, ship (worktrees, verify matrix, PR workflow, Linear state). Use whenever implementing or shipping IPI-NNN, wiring plans in docs/linear/issues, forensic verify before Done, MVP/P0 queue from tasks/plan/todo.md, multi-file platform work across app/supabase/edge, or user says "implement IPI-", "ship IPI-", "/task IPI", "close out Linear", "next P0 task", "forensic verify", "wiring plan", "open PR with verify". Always use for multi-step iPix delivery. Do NOT use for one-line typo fixes, explain-only questions, isolated copilotkit/ supabase/migration/lean/release-notes tasks without full lifecycle, or non-iPix repos.

Draft a changelog.md entry from git log + closed Linear IPI issues since the last entry. Use after a batch of commits lands or before tagging a release, when you need human-readable release notes rather than the /release gate checklist (which checks readiness, not prose). Triggers: "write release notes", "update the changelog", "draft changelog entry", "summarize what shipped". Do NOT use for the pre-merge gate check — that's `/release`.

Set up and operate git worktrees for isolated, parallel development — running multiple branches or Claude sessions at once without stashing or branch-switching, reviewing a PR while developing, comparing implementations side by side, or giving a risky change its own clean checkout. Use this whenever the user mentions worktrees, "isolated workspace", "parallel branches", "work on two things at once", `git worktree`, `--worktree`, or before executing a multi-step implementation plan that should not touch the current working tree. Also use when you (the agent) need an isolated checkout to make file changes that must not collide with the user's working tree.

Mastra framework: docs lookup (links.md, mastraDocs MCP, embedded docs), agents, workflows, streaming, browser, tools, memory, RAG, processors, CopilotKit guide. Verify from installed docs — never trust training-data APIs. NOT for: product intent routing (mastra-routing), Managed Agents API harness (mde-agents), non-Mastra frameworks. Load when editing src/mastra/**, Mastra tools/workflows, memory, streaming/events, browser automation, or Mastra doc URLs.

Turn a big, open-ended "build X" ask into a single scoped architecture brief before any code gets written. Use whenever the user hands you a goal-first, role-based prompt like "act as a senior software architect, build a production-ready X" or asks for an executive summary + architecture + database + frontend + backend + APIs + roadmap + risks all in one shot — this includes a pasted prompt template with a Requirements/Tasks/Output list, not just a plain-English ask. Forces explicit MVP-now-vs-later scoping and a risk section before routing each part of the brief to the skill that already owns it (graphify for analysis, ipix-supabase for schema, mermaid-diagrams for diagrams, frontend-design for UI, mastra for agents, writing-plans for the task breakdown) instead of freelancing a monolithic doc from scratch. Do NOT use for single-file fixes, for iterative back-and-forth design dialogue where the user wants to be asked questions one at a time (use brainstorming), or once a Linear IPI issue already has an appro

Use when building UI components, creating pages, or before launch for WCAG 2.1 AA compliance. Triggers on "accessibility", "a11y", "screen reader", "keyboard nav", "WCAG", "ARIA", "contrast".

Scrape e-commerce data for pricing intelligence, customer reviews, and seller discovery across Amazon, Walmart, eBay, IKEA, and 50+ marketplaces. Use when user asks to monitor prices, track competitors, analyze reviews, research products, or find sellers.

Pre-build exploration for iPix — clarify intent, propose approaches, get design approval before code. Use before new features, components, API routes, or behavior changes. HARD GATE: no implementation until design is approved. Triggers: "let's design", "how should we build", greenfield ideas, ambiguous scope. Terminal state: writing-plans. Do NOT use for trivial one-file fixes or when Linear issue already has A–E + wiring.

Browser automation with Claude in Chrome. Use this skill for web app testing, debugging with console logs, form automation, data extraction, multi-site workflows, and recording browser interactions as GIFs. Requires Chrome extension.

Local mirror of Anthropic Claude Code product documentation (code.claude.com/docs): CLI, hooks, skills, plugins, MCP, subagents, AGENTS.md, slash commands, output styles, sandboxing, permissions, Agent SDK (TypeScript and Python), Claude Desktop, Slack, Chrome, Web, monitoring and usage, the Claude Code changelog, and weekly "What's New" digests. Use whenever the user asks how Claude Code behaves, how to install or configure Claude Code, or what a Claude Code flag, hook, slash command, environment variable, or feature does. Read this skill's references/ before generic web search for Claude Code product questions. Do NOT use for OpenAI Codex, Cursor, general Anthropic API, Claude.ai consumer, or other coding agents.

Audit and improve CLAUDE.md files in repositories. Use when user asks to check, audit, update, improve, or fix CLAUDE.md files. Scans for all CLAUDE.md files, evaluates quality against templates, outputs quality report, then makes targeted updates. Also use when the user mentions "CLAUDE.md maintenance" or "project memory optimization".

Canonical UI/UX contract for iPix — read before Claude Design sessions, wireframes, Figma handoff, or operator screen implementation. Routes to design.md (how it looks) between PRD (what) and tasks/plan/todo.md (how to build). Use for screen specs, design review, 3-panel layout, HITL states, component reuse, motion, and accessibility gates.

Multi-platform e-commerce competitor analysis skill that automatically scrapes product data from Amazon, Temu, Shopee and generates comprehensive analysis reports using AI. Use when you need to analyze competitor products, extract product insights, or batch analyze multiple product listings. Supports bulk processing with structured outputs including title, price, rating, reviews, and strategic analysis.

Web/UI theming guide for a women's-fashion boutique **website** — a frontend design system applied phase by phase (typography, color, layout, components, animation, polish) for a modern editorial aesthetic. Use when styling or restyling a fashion storefront/landing UI: type scale, palette, grid/spacing, component look (cards, nav, product displays, CTAs), and motion. This is **CSS/UI theming, NOT shoot/wardrobe styling** — for model/garment styling on real shoots use `fashion-production` (styling phase). Pairs with `frontend-design` and `tailwind-design-system`.

Guided multi-file feature development with codebase exploration, architecture design, and quality review. Use whenever the user says build a feature, implement functionality, add a module, /feature-dev, architectural decisions, multi-file changes, complex integration, or greenfield code that spans UI + API + database. For iPix Linear tasks (IPI-/PLT-/UI-), prefer ipix-task-lifecycle hub first — this skill is the Phase 3 deep-dive when scope is large or architecture is unclear.

Medusa commerce development hub — backend (custom modules, workflows, API routes, module links, subscribers/events, scheduled jobs, data models), storefront (SDK, React Query, calling custom API routes, frontend integration), admin dashboard customizations (widgets, custom pages, forms, tables, navigation, data loading), Mercur marketplace (CLI, blocks, dashboard/vendor UI, @medusajs/ui conformance, 1.x→2.0 migration), DB CLI (`db:generate` migrations, `db:migrate` apply, `user` create admin), and an interactive step-by-step Medusa learning track — consolidated into one skill with on-demand references. Use for ANY Medusa or Medusa-backed storefront/marketplace work: custom module, workflow, API route, db:generate/db:migrate, create admin user, cart/checkout/PLP/PDP wiring, admin widget/page, Mercur seller/vendor/blocks, or learning Medusa architecture. Do NOT use for the iPix operator app (Vite/React, no Medusa runtime) — that's `ipix-supabase` / `frontend-design`.

Two-tier memory system that makes Claude a true workplace collaborator. Decodes shorthand, acronyms, nicknames, and internal language so Claude understands requests like a colleague would. CLAUDE.md for working memory, memory/ directory for the full knowledge base.

Color grading direction, retouching guidelines, and export specifications per brand DNA and channel target. Bridges the gap between RAW upload and DNA-scored final. Use when preparing post-production briefs for editors/retouchers.

Best practices and comprehensive guide for Remotion - programmatic video creation in React with animations, compositions, and media handling

Use this skill when creating short-form social media content for LinkedIn, Twitter/X, or other platforms

Create engaging content for social media platforms: X posts and threads, LinkedIn, Threads, Instagram, Facebook. Includes writing styles, AI-avoidance patterns, hooks catalog, and macOS Notes integration. Keywords: social media, X, X, LinkedIn, threads, posts, hooks, content strategy, AI writing avoidance.

Generate beat-by-beat storyboards for fashion video content (reels, TikTok, product videos, lookbook films). Includes frame descriptions, timing, camera movement, audio cues, and text overlays.

Design-to-code handoff for iPix/FashionOS: converts Claude Design exports (HTML dc-import libraries, *.v2*.dc.html screens), screenshots, mockups, and design-system docs into implementation-ready tasks and production code. Visual system: v3 Zeely Editorial (white/grey/black, Inter, black actions, image-first). Use when the user says "implement this design", "import Claude Design", "handoff from design", "convert mockup to code", "design to code", "map design to iPix stack", or shows design files from Universal design prompt/. Covers intake, gap analysis, DC library mapping, Phase A/B/C plan, Linear tasks, code generation, and QA. Does NOT handle Figma imports (Figma MCP) or wireframe creation from scratch (mde-wireframe).

Cloudinary media hub — transformation/delivery URLs, Next.js SDK (next-cloudinary, CldImage), Node.js SDK (server upload/signatures), React SDK, and API/docs lookup, consolidated into one skill with on-demand references. Use whenever the user mentions Cloudinary, image/video upload, CDN delivery, transformation URLs, f_auto/q_auto, resize/crop/overlay, signed uploads, CldImage, CldUploadWidget, next-cloudinary, @cloudinary/react, AdvancedImage, Upload Widget, video player, named/responsive transformations, transformation costs, webhooks/DAM/MediaFlows, or optimizing fashion/product photography assets — even if they do not say "Cloudinary" explicitly.

CopilotKit v2 hub — operator chat runtime, AG-UI/SSE streaming, A2UI surfaces, Mastra/LangGraph integrations, setup, develop, debug, upgrade, OSS contribute. Use whenever the user touches CopilotKit, CopilotChat/CopilotSidebar/Popup, /api/copilotkit (including 401 on /info), useAgent/useFrontendTool/useHumanInTheLoop, agent not found / agentId mismatch, frontend tools not executing, SSE stuck after RunStarted, AG-UI, A2UI createSurface, license token / publicLicenseKey, v1→v2 migration, or wiring Mastra agents in app/ — even if they never say "CopilotKit". Load references/ on demand. iPix prod: references/ipix-production.md; debug: references/debug/debug.md. Do NOT use for Supabase-only migrations/RLS, Mercur checkout, shadcn-only UI tweaks, Gemini edge prompts, or Linear issue updates without CopilotKit code.

Convert Claude Design HTML (*.v2.image-first.dc.html) into production Next.js React screens with verified visual parity. Use whenever the user says "design to production", "DC parity", "convert HTML to React", "match the design file", "image-first parity", references Universal design prompt/*.dc.html (including Shoots List), or asks to port Command Center / Brand List / Brand Detail / Shoots List into app/. Load before line 1 on any DESIGN-* UI task. HTML design wins for layout; React wins for data/auth/shell. Enforces RSC + CSS modules + tokens + state parity + side-by-side verification. Does NOT replace claude-design-handoff for full program planning — use this skill for execute-and-verify on a single screen or small route bundle.

Fashion shoot-production toolkit — the real-shoot pipeline from creative direction to delivery, consolidated into one skill with phase-organized references. Use for: creative direction / concept / moodboard rationale; shot lists (brief → shot-by-shot table with timing/crew/setup groups/channels); crew call sheets; model direction / casting / pose cards; garment lifestyle & location shoot briefs (scouting, props, talent); product, macro, Amazon, and Shopify product-photography specs (shot types, camera, lighting, QC); lighting plans (key/fill/accent per shot type & brand DNA); live shoot-day coordination (shot counter, schedule, issue log, reshuffling); post-shoot asset triage (auto-sort by shot/setup/channel, roughs vs finals, EXIF match); and AI ad generation via ComfyDeploy Morpheus. Triggers on shoot, shot list, call sheet, lighting plan, retouch/QC brief, lookbook, editorial, product shoot, reshoot, or any fashion production-planning task.

Integrate Firecrawl into product code and agent workflows — onboarding, /search, /scrape, /interact, and research paper retrieval. Use whenever adding Firecrawl to an app, wiring FIRECRAWL_API_KEY, building query-to-URL pipelines, single-page extraction, browser interactions (clicks, forms, pagination, auth flows), literature or paper search, or any web discovery + extraction feature — even if the user does not say "Firecrawl".

Frontend & UI design hub — create distinctive, production-grade interfaces, with on-demand references for a UI pattern catalog and brand color systems. Use when building or styling web components, pages, dashboards, or apps; choosing a visual style / font pairing / layout / chart; building a design system (Tailwind tokens, component libraries, responsive patterns); or generating an accessible brand color palette (11-shade scale, semantic tokens, dark mode, WCAG contrast). Generates creative, polished code that avoids generic AI aesthetics. Pairs with the `shadcn` MCP skill, `accessibility`, and `ipix-wireframe`.

Use when integrating Gemini AI models in Supabase Edge Functions or Deno — Gemini 3.1 flash-lite, thinking levels, URL Context, Google Search grounding, structured output, function calling, image generation via generateContent. Use for AI enrichment, DNA scoring, product linking, and analysis tasks. Do NOT use for CopilotKit/Mastra client-side AI (use mastra/copilotkit), OpenAI/Anthropic direct integration, or Interactions API through MVP.

Generates Vitest unit tests for iPix using @testing-library/react in app/. Use when adding tests for React components, hooks, services, or utilities — required by ipix-task-lifecycle per-task testing. Do NOT use for Playwright/E2E tests, Supabase RLS Docker tests, edge function Deno tests, or legacy root src/ Vitest (retiring) — use the area-specific verify scripts instead.

Query the iPix codebase and docs knowledge graph instead of grepping. Use for IPI dependency analysis, PRD traceability, Supabase schema impact, Brand Intelligence dependencies, Asset DNA dependencies, CopilotKit/Mastra workflow mapping, or Medusa commerce graph queries. Also use when the user asks "what connects X to Y", "find all dependencies of Z", "what files does feature W touch", "trace the impact of changing A", "map the data flow for B", or "which skills are relevant for C".

Groq API + ultra-fast LLM inference for chat, vision, audio STT/TTS, tool use, structured outputs, and Mastra agents. Use whenever the user mentions Groq, GroqCloud, GROQ_API_KEY, fast inference, low latency, Whisper on Groq, PlayAI TTS, Llama on Groq, compound models, groq hello world, groq quick start, groq-sdk, API reference, production checklist, model pricing RPM, or migrating off Gemini to Groq — even if they don't name this skill. Replaces legacy groq-api and groq-hello-world skills.

Infisical secret-management hub — agent daemon, REST API, dynamic/ephemeral credentials, secret syncs, and project setup (CLI, Docker, CI/CD, K8s, SDKs), consolidated into one skill with on-demand references. Use for: injecting secrets, infisical run / infisical init, machine identity auth (Universal Auth, K8s, AWS, OIDC), the Infisical Agent daemon + Go-template config + sidecar injection, the Infisical REST API (secrets CRUD, projects, identities, pagination/rate limits), dynamic/short-lived credentials for PostgreSQL/MySQL/Redis/MongoDB/cloud IAM/SSH/K8s, pushing secrets to AWS/GCP/Azure/GitHub/Vercel/Cloudflare/Vault via Secret Syncs, and SDK setup (Node, Python, Go, Java, .NET, Ruby). The one skill for any Infisical task.

iPix Supabase hub (project nvdlhrodvevgwdsneplk): schema, RLS, migrations, auth, edge functions, storage, Realtime, verify-rls. Consolidates edge-functions, generic Supabase SQL/RLS/migrations, the Supabase CLI, and Postgres performance best-practices into one skill with on-demand references. Use for ANY Supabase work in this repo. NOT for Mercur commerce tables or legacy Medellín/FashionOS edge functions.

Wireframes, lo-fi clickable prototypes, Wire DSL mockups, and sketch-to-spec handoff for iPix / Lumina Studio operator UI. Use for wireframe, mockup, lo-fi UI, interactive prototype, early feedback, stakeholder alignment, operator dashboard, brand intake, DNA scoring UI, shoot wizard, or ideating UI before production code.

Development speed optimizer for Claude Code workflows. Audits a repository and development environment to maximize velocity, shrink feedback loops, and cut AI context waste. Use this skill whenever the user mentions slow builds, slow CI, slow tests, slow Claude responses, "why is this taking so long", repo getting bloated, too many files, Claude reading too much, wasted tokens, or any request to "speed things up", "optimize the dev loop", "audit the repo", or "run /lean". Also trigger proactively when you notice: more than 3 CI roundtrips on a single PR, Claude reading >10 files to answer a simple question, builds taking >2min, or test suites taking >60s locally. The output is a scored report with ranked fixes — always produce it, never just answer in prose.

Consolidated Linear hub for linear.app issue/project/cycle/initiative/milestone work, Linear Method planning, issue generation, search, sync, and iPix IPI-### workflows. Use whenever the user mentions Linear, linear.app, IPI-###, PLT-###, COM-###, UI-###, AI-###, DNA-###, issue, project, cycle, roadmap, milestone, backlog, or asks to create/update/search/sync Linear work. Prefer this skill over legacy linear-* skills.