jezweb/claude-skills
77 skills76.2k installs64.8k starsGitHub
Install
npx skills add https://github.com/jezweb/claude-skillsSkills in this repo
1Shadcn Uishadcn-ui is a concise catalogue skill for indie builders shipping React and TypeScript interfaces with shadcn/ui, Tailwind, and common form libraries. It documents the roughly fifteen most-used primitives—Button, Input, Label, Card, Form—with install commands via pnpm dlx, variant enums, and TSX snippets agents can paste without hallucinating prop names. The readme calls out integration sharp edges, especially FormField usage with react-hook-form where spreading field objects can break on null values. Cards follow the standard Header, Title, Description, Content, Footer structure so agents generate accessible layouts consistent with the official docs. Use during Build frontend passes when scaffolding settings pages, auth forms, or dashboards. It is reference material, not a design system audit; pair with your brand tokens and accessibility review at Ship. Beginners benefit from the explicit add commands; advanced users still save time on repetitive prop documentation.2.8kinstalls2Tailwind V4 ShadcnTailwind v4 + shadcn/ui is a frontend agent skill for solo builders shipping React apps who want shadcn’s component model on Tailwind’s v4 CSS-first configuration. Use it when initializing a Vite + React project, migrating from Tailwind v3, or debugging symptoms like all-gray UI, missing utilities, broken @apply, or dark mode that never toggles. The skill encodes a four-step pattern—define CSS variables, map them through @theme inline, apply base styles, then wire automatic dark mode—so agents stop guessing import order and deprecated directives. It is tuned to error messages practitioners actually hit (tw-animate-css, typography import syntax, theme provider conflicts). Complexity is intermediate because you should already be comfortable with React tooling, but the checklist keeps first-time v4 adopters out of week-long stylesheet rabbit holes. After setup, your agent can extend shadcn tokens consistently for dashboards, marketing sites, and internal tools alike.2.7kinstalls3Tanstack QueryTanStack Query is a Claude-oriented agent skill for managing server state in React with TanStack Query v5. It is aimed at solo builders shipping SaaS or mobile-web frontends who need reliable caching, mutations, infinite lists, and offline-aware network modes without re-reading the entire migration guide. The skill activates on familiar triggers—useQuery, staleTime, query invalidation, optimistic updates—and on concrete breakage such as missing cacheTime, removed query callbacks, or hydration mismatches with streaming RSC. Documentation emphasizes production-tested TypeScript patterns, DevTools usage, prefetching, and suspense queries so agents propose v5-correct APIs instead of deprecated v4 habits. Use it during Build when wiring QueryClientProvider, refactoring data layers, or debugging SSR hydration. It does not replace backend API design; it ensures the client cache, keys, and mutation lifecycle behave predictably so Ship-phase testing and Grow-phase analytics views stay stable under real network conditions.2.5kinstalls4Tailwind Theme BuilderTailwind Theme Builder is an agent skill package for solo and indie builders who use shadcn/ui with Tailwind CSS v4 and want a correct, copy-paste theme foundation instead of guessing variable structure. The skill encodes the four-part pattern from shadcn’s Tailwind v4 docs: define CSS custom properties at the document root, override them under `.dark`, map them with `@theme inline`, and express every color as hsl(). It includes a components.json scaffold (style, RSC/TSX flags, Tailwind paths, alias map) plus starter `:root` tokens for cards, popovers, primary, secondary, muted, accent, destructive, and success states. Use it when you are starting a new dashboard or marketing app, refactoring from Tailwind v3, or fixing broken dark mode after upgrading shadcn. It does not replace brand exploration in Idea or accessibility audits in Ship—it gives you a standards-aligned baseline so your agent can extend tokens consistently across components.2.5kinstalls5Color PaletteColor Palette is an agent skill packaged as a copy-ready Tailwind v4 theme template for solo builders shipping SaaS or browser extensions. It encodes an 11-step shade ladder from primary-50 through primary-950, maps your brand baseline (typically 500/600), and wires shadcn-style semantic tokens for background, foreground, card, popover, and destructive states in both light and dark modes. Use it when you have a single brand hex but need accessible contrast, consistent hover states, and CSS variables your agent can extend without guessing HSL steps. It matters because ad-hoc one-off colors break design systems and slow every new component; a full scale plus semantics lets Claude Code, Cursor, or Codex scaffold UI faster and keep tokens aligned across routes and marketing pages.2.2kinstalls6FastapiFastAPI is an agent skill for solo and indie builders shipping Python HTTP APIs. It encodes production-tested conventions for spinning up projects with uv, organizing code by domain, and wiring Pydantic v2 for validation that actually matches what clients send. Coverage extends through async SQLAlchemy 2.0 persistence, JWT/OAuth2 bearer auth, async-safe handler patterns, and pytest-based API tests, finishing with ASGI deployment notes for Uvicorn and Gunicorn. The README explicitly maps recurring failures—422 validation errors, CORS browser blocks, fields wrongly treated as required, and hung requests from blocking calls in async code—to preventative patterns. Use it when your agent is scaffolding or hardening a REST backend rather than improvising stack choices from chat. It is a procedural playbook, not a hosted service or MCP connector.2kinstalls7Tanstack Starttanstack-start is a Claude-oriented generator skill for building a complete TanStack Start application deployed on Cloudflare Workers. Solo builders who want SSR, file-based routing, and server functions without downloading a starter repo invoke it when they mention TanStack Start, a Cloudflare full-stack dashboard, or React 19 on Workers. The skill prescribes an end-state tree: root layout, login/register, protected dashboard and items routes, Tailwind v4, shadcn components, Drizzle on D1, and better-auth. That stack suits indie SaaS dashboards, internal tools, and agent-assisted greenfield products where you need one coherent opinion instead of piecing docs together. It is not a migration guide for existing Next or Remix apps. Expect intermediate familiarity with Wrangler, SQL schemas, and OAuth callback URLs. The outcome is a project-shaped artifact your agent can implement file by file until `wrangler dev` runs.1.7kinstalls8Image ProcessingImage Processing is an agent skill for solo web builders who need dependable asset prep without juggling ImageMagick installs. It centers on the shipped img-process command-line tool plus optional custom Python scripts using Pillow when logic exceeds the standard commands. Typical jobs include resizing heroes, converting PNG or JPG to WebP, trimming whitespace from logos, generating square thumbnails, compressing large photos, batch-processing folders, and composing simple OG social cards with title and subtitle text. The skill tells agents to prefer the CLI for standard operations because it is faster than regenerating one-off scripts each session, while still supporting compositing, watermarks, and conditional flows when requirements are nonstandard. It fits the Build frontend track when you are wiring landing pages, app UI, or docs sites that need correct dimensions, modern formats, and share images before deploy. Documented as claude-code-oriented compatibility, it still applies wherever your agent can run local shell commands against a Python environment with Pillow installed.1.7kinstalls9Favicon GenFavicon-gen is an agent skill that walks solo builders through creating a production-ready favicon set from whatever brand inputs they have—full logo, initials only, or colors alone. The workflow branches early: extract an icon from an existing logo, compose a centered monogram in a 32×32 viewBox, or pick a branded shape (circle, rounded square, shield, hexagon) before generating derivatives. It explicitly targets the annoying CMS-default favicon problem and common display failures on iOS and in PWA installs. After the source SVG is settled, the skill drives generation of SVG, ICO, Apple touch icon, standard PWA sizes, and manifest wiring plus HTML integration guidance. ImageMagick is required for bitmap conversion. Best used when you are building or refreshing a site and want one consistent package instead of guessing which sizes browsers and stores expect.1.6kinstalls10Ux Auditux-audit gives solo builders a repeatable, agent-driven protocol to bolt automated accessibility onto a broader UX audit—especially during polish before launch. Manual keyboard scenarios still matter for focus traps and tab order, but axe-core closes the gap on heading hierarchy skips, contrast at hover and focus, missing aria-labels, live-region announcements, and role mismatches that human passes often miss. The skill documents a strict per-page flow: navigate, inject the CDN script if needed, evaluate axe.run(), normalize the violations array, and translate axe impact into your audit severity table so findings land in the same backlog as visual UX issues. It is framed as mandatory in Phase 4 (Polish), which makes it a ship-phase testing habit rather than an optional nicety. Compatible with browser eval, Playwright CLI, or Chrome MCP contexts, it fits indie teams who cannot staff dedicated accessibility engineers but still need WCAG-oriented signal before shipping customer-facing pages.1.6kinstalls11Wordpress ElementorWordPress Elementor is an agent skill for solo builders who maintain sites built with Elementor and need reliable edit paths without breaking serialized page data. It walks you through identifying builder pages with WP-CLI meta filters, choosing the right method for each change type, and applying updates either via safe search-replace and meta updates or through browser automation when structure or styling must change visually. The skill is explicit that low-risk copy swaps belong on the CLI with backups, while section adds, widget styling, and layout moves require the Elementor editor driven by automation. It assumes Elementor is installed and that you already have site access from the wordpress-setup skill. For indie founders updating landing pages, pricing blurbs, or hero images, this reduces the guesswork that corrupts _elementor_data JSON.1.6kinstalls12Elevenlabs AgentsElevenlabs-agents is an agent skill that gives solo builders a structured reference for assembling ElevenLabs voice agents: system prompts, model choice, tool and knowledge-base attachments, text-to-speech and speech-to-text parameters, turn-taking, and conversation duration limits. Builders shipping customer support, onboarding, or hands-free product experiences use it while integrating the Agents API rather than improvising JSON in chat. The skill surfaces patterns like custom LLM endpoints with secret placeholders, turbo TTS models, Deepgram ASR quality settings, and language presets so one agent can greet users in English and Spanish with distinct voices. Workflow nodes hint at multi-step or subagent routing for more complex call flows. Complexity is intermediate because you must align API keys, voice IDs, tool IDs, and latency budgets with your backend. It does not replace ElevenLabs account setup, billing, or telephony/web SDK wiring—pair it with your deployment docs when you move from config draft to production traffic.1.3kinstalls13Zustand State Managementzustand-state-management is a production-oriented agent skill for building global React state with Zustand v5-style APIs, TypeScript inference, and common middleware. Solo builders reach for it when Redux feels heavy, Context causes rerender pain, or localStorage persistence and DevTools need a consistent recipe. The skill documents selector usage, shallow equality, optional Immer integration, and the slices pattern for larger apps. It explicitly addresses Next.js server rendering and hydration mismatches— a frequent footgun when persisting client state. Auto-trigger keywords cover migrations from Redux or Context, persist to localStorage or sessionStorage, and error-led searches like infinite render loops or TypeScript inference gaps. You should use it while implementing features in Build, not as a deployment or monitoring playbook. The author positions it as React 19 compatible and ecosystem-tested as of late 2025, with MIT-licensed reference material in the jezweb skills repo.1.3kinstalls14Icon Set GeneratorIcon Set Generator is an agent skill packaged as an industry icon reference for solo and indie builders shipping business websites. When someone describes a project or vertical, the skill steers the agent toward sensible icon names for navigation, contact, trust, and service storytelling instead of generic or mismatched symbols. It separates nearly universal patterns every site needs from richer sets tailored to trades and home services or professional services, including notes on typical set sizes so scope stays realistic. The workflow is consultative: recommendations are starting points, and the user should confirm before implementation in whatever icon library or SVG set the stack uses. It fits Claude Code, Cursor, and similar agents during frontend and landing-page design when you want faster IA decisions without hiring a visual designer for the first pass.1.3kinstalls15React Hook Form Zodreact-hook-form-zod teaches solo builders how to ship validated React forms without duplicating types or fighting mysterious resolver and uncontrolled-input warnings. It standardizes on one Zod schema that powers client-side UX and server-side validation, using `z.infer` so TypeScript tracks field shapes through submit handlers, API routes, and error display. The skill is aimed at product UIs—single pages, multi-step wizards, dynamic lists via useFieldArray—and integrates with shadcn/ui Form primitives when you want accessible labels and messages fast. Invoke it when you are building forms, tightening error handling, or debugging performance on large field sets. Documented production notes cite roughly sixty percent token savings versus ad-hoc explanations and twelve recurring issues the patterns prevent. Intermediate complexity: you should know basic React hooks and npm installs. It is not a substitute for accessibility audits or backend authorization logic, but it keeps validation coherent across the stack so agents generate consistent, copy-paste-safe form code.1.2kinstalls16Google Apps ScriptGoogle-apps-script is a Claude-oriented skill for solo builders who run parts of their business in Google Sheets and want server-side automation without spinning up a separate backend. The agent walks through understanding the scenario—custom menus, reactive triggers, sidebar data entry, outbound email, PDF generation, or calling external APIs—then emits paste-ready Apps Script with consistent structure and installation instructions via Extensions > Apps Script. It targets the generous Google free tier and fits indie operators who prefer low-ops automation over maintaining VPS cron jobs. Use it while building operational tooling tied to a sheet, and again in Operate when you extend triggers or fix production sheet workflows. It does not replace full Apps Script OAuth review for sensitive scopes; you still validate permissions Google prompts for at install time.1.2kinstalls17D1 Drizzle Schemad1-drizzle-schema is a template skill for solo builders shipping on Cloudflare Workers who want Drizzle ORM aligned with D1 instead of hand-rolling raw SQL. It bundles a drizzle-kit configuration that points schema output at ./src/server/db/schema.ts, writes migrations to ./drizzle, and authenticates remote pushes via d1-http and standard Cloudflare API credentials. The accompanying schema module shows production-minded SQLite patterns on D1: crypto.randomUUID primary keys, role enums on text columns, booleans stored as integers with mode boolean, createdAt and updatedAt as integer timestamps, and preferences as typed JSON blobs. Indexes include a unique constraint on email for a users table, with relations ready to extend to related entities. Use it when you have chosen D1 as your database and need a consistent starting point before API routes or auth layers. It is reference material rather than an automated migrator—you still run drizzle-kit generate and apply against your D1 instance. Pair with your Workers binding setup and environment secrets so accountId, databaseId, and token resolve in CI and locally.1.2kinstalls18Seo Local BusinessSEO Local Business is a Claude skill that gives solo builders a copy-ready HTML head section tuned for brick-and-mortar and service-area businesses. It standardizes page title and description placeholders, canonical URLs, Open Graph and Twitter sharing fields, and geo-oriented meta tags so local landing pages send consistent signals to search and social crawlers. Indie operators launching a studio, clinic, or shop site often skip geo tags and canonical discipline; this template reduces that gap without requiring a full SEO agency stack. Use it when you are shipping location pages, service city pages, or a main business homepage before you turn on ads or directory listings. Pair with real NAP consistency and Google Business Profile work outside the skill.1.2kinstalls19Project HealthProject-health (Context Capture on Commit) is a journey-wide Claude Code hygiene skill for solo builders who lose agent context between sessions. It recommends a hookify rule that watches bash git commit, warns after each commit, and asks whether new patterns, gotchas, or commands should be persisted to CLAUDE.md or .claude/rules before you context-switch. That turns every commit into a lightweight retrospective for procedural knowledge your coding agent will reuse on the next feature, bugfix, or deploy. You can install the hook globally or per repo; without hookify, the same behavior can be inlined as a standing instruction in CLAUDE.md. It complements Mode 1 Session Capture rather than replacing tests or monitoring. Best when you already commit frequently and treat the repo as the source of truth for how this project is actually built.1.2kinstalls20Tailwind PatternsTailwind Patterns is an agent skill that supplies production-ready Tailwind CSS recipes for common website components—responsive layouts, cards, navigation, forms, buttons, typography, and section spacing—with mobile-first breakpoints and dark mode in scope. Solo builders use it when they are implementing UI in React or similar stacks and want consistent utility-class structure instead of inventing padding and grid rules from scratch. The skill enumerates auto-trigger keywords around Tailwind, shadcn/ui, hero sections, and form validation so agents suggest it during landing pages and dashboard chrome work. It emphasizes canonical patterns such as centered containers with px-4 sm:px-6 lg:px-8 and reusable component shapes you can paste into JSX className attributes. It does not replace a full design system or accessibility audit; it accelerates Build-phase frontend iteration when you already chose Tailwind as your styling layer.1.1kinstalls21Hono Api ScaffolderHono API Scaffolder is an agent skill for solo and indie builders who want a consistent Hono backend skeleton instead of improvising routes on every project. It encodes opinionated templates for centralized JSON error handling, optional Bearer authentication middleware, and a full CRUD route file you duplicate per resource with @hono/zod-validator. The skill fits the Build phase when you are standing up an API service—often on edge runtimes—and need predictable 401/500 JSON responses and validation at the boundary before business logic grows messy. Highlights emphasize API-only error semantics, auth context variables, and REST-shaped handlers you customize per domain. Complexity is intermediate because you must wire token verification, persistence, and deployment yourself. Confidence is high for template fidelity; README is code-forward rather than narrative docs.1.1kinstalls22Responsiveness Checkresponsiveness-check gives solo builders a structured breakpoint reference for validating responsive web UI instead of guessing one-off phone widths. It documents eight standard viewport widths from small phones through 4K displays, each with what typically breaks—touch targets at 375px, grid columns near 1024px, hero proportions at 1440px, and max-width containment on ultra-wide screens. A denser fifteen-point sweep every 160 pixels highlights transition zones many teams skip, including phablet widths, the 800–960px navigation switch band, and ultra-wide whitespace above Full HD. The skill is meant to drive systematic visual or automated checks across those widths so regressions surface before launch. Pair it with your component library or Tailwind config so agents know which utility breakpoints should align with real device stress points. It does not replace accessibility audits or performance budgets; it focuses on horizontal layout integrity across the viewport ladder.1.1kinstalls23VitestVitest is an agent skill for solo builders who want working test infrastructure instead of a reference card. It scans package.json, tsconfig, wrangler.toml, and existing Vite or Jest config to classify the project—Cloudflare Workers, React, Node, or library—and then produces a tailored vitest.config.ts, setup files, utilities, a sample test, and npm scripts with TypeScript wiring. The workflow is deliberately generative: detect, configure, scaffold, and wire up, including mocking patterns, coverage, monorepo workspace setup, and migration off Jest when jest.config is present. Use it when you mention adding tests, setting up Vitest, fixing broken test infra, or asking how to test a module in a Vite-based stack. It fits indie SaaS, CLIs, and APIs where fast unit tests unblock Ship without adopting a separate Jest toolchain.1.1kinstalls24Cloudflare Worker BuilderCloudflare Worker Builder is an agent skill for solo developers who want a opinionated edge stack without spelunking Cloudflare docs for every wiring detail. It centers on a private npm package with deploy and dev scripts via Wrangler, TypeScript typing generation, and worker-native tests through the Vitest pool. Hono provides lightweight routing while static assets and run_worker_first configuration demonstrate how to keep APIs on the Worker and still ship a small front-end shell. The bundled HTML harness makes it easy to click-test routes locally before you connect a real domain or CI deploy. Use it when you are committing to Workers as your API or BFF layer rather than a long-lived VPS or container-only backend.1kinstalls25NextjsThe Next.js skill from jezweb/claude-skills is a production-oriented playbook for solo builders shipping App Router sites on Next.js 16. It concentrates on breaking changes—async dynamic params, proxy.ts, parallel routes—and the newer caching surface around Cache Components and tag-based invalidation. Agents invoking it should get opinionated snippets for Server Components data loading, Server Actions with validation, and API route handlers for webhooks or streaming. The README positions measurable value: large token savings versus raw doc dumping and a catalog of errors the patterns intentionally avoid. Reach for it when starting a greenfield SaaS frontend, migrating from Pages Router, or debugging Turbopack and i18n cache quirks during an active build sprint rather than during distribution or analytics work.1kinstalls26Wordpress ContentWordPress Content is a Claude Code agent skill for solo builders who run WordPress sites and want the agent to execute real publishing work—not just draft copy in chat. It walks through choosing WP-CLI versus REST for each task, from single post create/update through bulk operations, media upload, term management, and menu updates. Prerequisites assume a working site connection established with the wordpress-setup skill and config from wordpress.config.json or wp-cli.yml. Use it when you need drafts, scheduled posts, complex HTML via temp files, or agent-driven bulk content changes. It matters because it turns repeatable CMS chores into scripted, auditable commands your coding agent can run while you focus on strategy and writing.1kinstalls27Mcp Buildermcp-builder packages a FastMCP-oriented API client pattern for indie builders shipping agent products: a reusable Python skeleton with configuration from environment variables, a singleton async HTTP client, and hooks for cache TTL and retry limits. It fits when you are past idea validation and need reliable tool endpoints that your coding agent can call without reinventing pooling, timeouts, or header auth on every integration. The skill emphasizes procedural knowledge—how to structure Config, client lifecycle, and FastMCP registration—so small teams can stand up MCP servers that talk to REST backends safely. Use it in the build phase alongside your service contracts; pair with security review before ship when keys and network egress matter.992installs28Shopify ProductsShopify Products is a Claude Code workflow skill for indie ecommerce founders who need real products in a Shopify store—not mock data. It walks agents through gathering titles, HTML descriptions, types, vendors, tags, variant options, SKUs, prices, inventory, image sources, SEO handles, and collection assignments, then choosing between the GraphQL Admin API and CSV import for execution and verification. Prerequisites center on an Admin API access token and store URL/API version from shopify.config.json or .dev.vars, with shopify-setup recommended when auth is missing. The bundled CSV header documents the standard Shopify import column shape for bulk catalogs from spreadsheets or scraped URLs. Ideal when you are migrating a catalog, A/B pricing, or uploading many SKUs without clicking through admin for every row.973installs29Wordpress SetupWordpress-setup is an agent skill that teaches WP-CLI essentials for builders who run many WordPress properties remotely rather than clicking only in wp-admin. It centers SSH connection patterns—from one-off `--ssh` invocations to reusable `@alias` definitions at project and global config levels—so agents emit commands that work on real host paths. A flags table covers structured output for scripting, quiet runs, root allowances common on shared hosting, and skip-plugins or skip-themes modes when plugins break CLI bootstrap. Diagnostic command blocks give a fast health pass: URLs, database footprint, plugin inventory, users, and scheduled events that affect publishing. Hosting notes such as Rocket.net reduce guesswork on where the WordPress root lives. Solo builders use it when fixing client sites, auditing before migrations, or automating checks without writing custom PHP bootstrap scripts.961installs30Github Releasegithub-release is a reference skill for solo builders who ship from GitHub and want repeatable tag-and-release steps instead of improvising each time. It starts with version detection—`package.json` when present, otherwise explicit user input—and enforces conventional `v`-prefixed tags, with monorepo scoped tag names when multiple packages share one repo. The workflow covers conflict handling when a tag already exists, the safe push sequence (branch before tags), and generating bullet release notes from commits since the previous tag. Integration with GitHub CLI (`gh release create`) includes title, notes body, and pre-release variants. Use it when a feature branch is ready to become a public artifact: you need auditable history, a pointer tag, and a Release page your users or package consumers can trust.928installs31Landing PageThe landing-page skill helps solo builders ship credible marketing surfaces fast when they need to validate an offer, announce a launch, or host a coming-soon page without standing up a frontend toolchain. It follows a brief-first workflow: product name, value proposition, audience, primary call to action, and optional brand colors, logo, contact details, and section toggles. The deliverable is a single deployable HTML document using Tailwind through a CDN, responsive breakpoints, dark mode support, accessible semantic structure, and Open Graph metadata so links look professional in social and AI search previews. Sections typically include a hero with CTA, feature grid, social proof, optional pricing, FAQ, and footer—aligned with indie validation and lightweight launch needs. Because there is no compile step, you can drop the file on static hosting, object storage, or even open it locally for stakeholder review. Triggers match natural language like landing page, marketing page, launch page, or one-page site, making it easy to pair with positioning work in validate and distribution work in launch.925installs32Google Chat Messagesgoogle-chat-messages is a compact agent skill for solo builders who need reliable Google Chat incoming-webhook payloads from TypeScript. Instead of pasting API docs into chat, you get opinionated builders for cardsV2 and simple text messages, including optional thread keys for threaded rooms. The skill emphasizes Google Chat’s formatting dialect so alerts and CI notifications render correctly. It suits indie operators wiring deploy bots, incident notifiers, or internal agent workflows into Workspace Chat without building a full Chat app. You still supply the webhook URL and HTTP client; the skill focuses on message shape and copy formatting. Use it when your product or ops stack already posts to Chat and you want your coding agent to generate correct message objects every time.892installs33Aussie Business EnglishAussie-business-english is a journey-wide writing style skill for builders and freelancers who sell or support Australian customers. It encodes EN-AU spelling (colour, organise, centre, travelling) and noun–verb pairs (licence vs license) so agents do not slip into US defaults. The tone model is professional but human: direct without bluntness, warm without forced slang, and credible in emails, chat, proposals, and marketing pages. Solo operators use it whenever they draft or edit business text—whether that is a Validate pricing email, a Ship launch announcement, or Grow lifecycle messages. It does not replace domain skills like copywriting or article-content; it layers locale and register on top so the same structure sounds locally right. Invoke for drafting new messages or tone-checking existing copy before you send to AU clients, investors, or regulators who notice American spellings instantly.891installs34Vite Flare StarterVite Flare Starter is an agent skill that gives solo builders a concrete starter matrix for full-stack apps on Cloudflare: Vite on the client, a Worker entry at `./src/server/index.ts`, static assets from `./dist/client`, and D1 plus R2 bindings with drizzle migrations. The included env block covers better-auth secrets, local BETTER_AUTH_URL, optional Google OAuth, feature flags for email signup, and Vite-facing app identity placeholders so one rename pass aligns branding across shell and UI. Wrangler config documents compatibility_date, nodejs_compat, SPA not_found handling, and API-first routing—patterns indie teams need before first deploy. Use it when you are past idea validation and want an opinionated edge stack instead of hand-rolling wrangler and auth env every project. It does not replace Cloudflare dashboard steps for creating database IDs or OAuth consoles; it accelerates consistent file scaffolding your agent can fill in.878installs35Shopify SetupShopify-setup is a scope reference for solo builders and small teams connecting Claude Code or similar agents to Shopify’s Admin API. Before you generate OAuth apps, custom apps, or automation that lists products or updates inventory, you need the right permission bundle—over-scoping creates audit risk and under-scoping breaks scripts mid-run. The skill organizes required scopes for product management (read/write products, listings, inventory), content management (read/write content), and files (read/write files), plus optional scopes when you edit themes, translations, navigation, or custom metaobject types. A scope-selection-by-task section helps you request only what each job needs, which matters for app review and least-privilege security. Use it during backend or integration work on headless storefronts, internal ops tools, or agent-driven merchandising—not as a substitute for Shopify’s full API docs, but as a fast checklist your agent can follow when scaffolding credentials.875installs36React NativeReact Native is an agent skill for solo and indie builders shipping iOS and Android apps with React Native and Expo. It organizes mobile engineering guidance by impact so you fix CRITICAL list and scroll problems before polishing medium-priority details. The skill covers virtualized lists, stable keys, Reanimated animations, navigation with Expo Router, state management, and platform-specific code paths. Use it when starting a new Expo project, debugging janky scrolling, or reviewing a codebase before release. It is framed as a living pattern library rather than a one-off snippet dump, which helps agents give consistent reviews across sessions. Pair it with your agent’s normal lint and test flow; it does not replace device profiling or native toolchain debugging but steers implementation toward patterns that keep scroll and animation performance acceptable on real hardware.867installs37Social Media PostsSocial-media-posts is an agent skill that writes platform-native social copy for solo founders who need distribution without hiring a social manager. You can start from a topic and bullet points, repurpose an existing blog or newsletter, or run a coordinated campaign across LinkedIn, Facebook, Instagram, and Reddit with suggested posting order. The workflow gathers required inputs—topic or source material, target platforms, and optional tone—then shapes hooks, hashtag strategy, and length to each network’s conventions so posts are ready to paste into schedulers or native composers. It fits launch week announcements, feature drops, and grow-phase content recycling when you already have long-form assets but lack short-form variants. The skill uses read and write tooling to pull source files when you point at content paths. It does not replace analytics, paid ads setup, or community moderation; it focuses on draft quality and format compliance so you ship consistent voice everywhere you show up.861installs38Shopify ContentShopify Content is a compact API reference for solo and indie builders who automate Shopify storefront content from Claude Code, Cursor, or similar agents. It maps everyday content operations—static pages, blog articles, URL redirects, and related meta—to the correct Shopify Admin surface, mostly GraphQL for pages and articles and REST for redirects. Use it when you are implementing theme-adjacent content, migration scripts, or agent-driven store updates and need exact mutation names and field shapes without opening Shopify docs in another tab. The skill does not replace OAuth setup, theme Liquid, or checkout customization; it accelerates the integration layer where your code or agent talks to Shopify. It fits ecommerce SaaS, headless-adjacent workflows, and small teams shipping on Shopify who want repeatable, citable procedure in the repo instead of ad-hoc chat guesses about API versions and field names.853installs39Design ReviewDesign Review is a Claude Code skill that inspects whether a web app or page looks professional, consistent, and polished—not whether users can complete tasks. Solo builders use it when something feels “off” after a build, before client or team demos, or as a periodic quality pass on a live product. The agent walks structured visual checks (spacing grids, type scale, colour use, hierarchy, component consistency, motion/interaction cues, breakpoints) and outputs a findings report backed by screenshots. It deliberately excludes UX workflow auditing; pair it with ux-audit when you need both usability and visual craft. Browser automation is required, with the same URL and MCP detection patterns as ux-audit. Ideal for indie SaaS dashboards, marketing sites, and extension UIs where developer-built aesthetics need a design-conscious second opinion without hiring a full review cycle.825installs40React PatternsReact Patterns is an agent skill that packages React 19 performance and composition architecture guidance for Vite and Cloudflare Workers–style projects. Solo builders install it when they are writing pages and components, reviewing PRs, or chasing “why is this slow?” and unnecessary re-renders. The playbook orders fixes by impact so you address eliminating async waterfalls first, then bundle and render work, then composition and boundary patterns. It doubles as a refactoring guide for prop-heavy components and as a review rubric before merge. Triggers include react patterns, react review, react performance, composition patterns, and reduce re-renders. It is procedural knowledge meant to be invoked repeatedly across frontend work—not a one-off generator—so agents read existing code with Read, Glob, and Grep and apply the ranked rules consistently.822installs41Git WorkflowGit Workflow is an agent skill that turns messy git requests into repeatable rituals for solo builders using Claude Code. It focuses on pull request preparation—gathering commits and diffs, drafting a concise title and evidence-based body, pushing, and opening a PR with GitHub CLI—as well as branch cleanup, conflict resolution, and monorepo tagging called out in its description. Use it when you are ready to ship a slice of work, not when you are still brainstorming features. The skill is multi-phase in practice: you create commits during Build, but the canonical Prism shelf is Ship review because the documented steps center on PRs and merge safety. Following the flow reduces forgotten test plans and branch rot on one-person repos. Deliverables are pushed branches, PR text grounded in git log, and optionally cleaned local branches or release tags after verification.809installs42Project DocsProject-docs is a Claude Code agent skill that generates structured project documentation by analysing the real codebase. Solo and indie builders install it when a repo has no docs, outdated architecture write-ups, or when someone new needs to understand how routes, APIs, and the database actually work today. The workflow starts by detecting project type from familiar root files—wrangler for Workers, vite and App.tsx for React SPAs, drizzle config for ORM layers, pyproject for Python, and similar signals—then reads source, schema files, route definitions, and configuration to produce accurate ARCHITECTURE.md, API_ENDPOINTS.md, and DATABASE_SCHEMA.md. That matters because agent-assisted refactors and fast shipping often outpace human-written docs; this skill closes the gap so your coding agent and collaborators operate on facts from the tree, not memory. Triggers include phrases like generate docs, document architecture, create API docs, and document schema.783installs43D1 MigrationD1 Migration is an agent skill that walks solo builders through Cloudflare D1 database migrations when using Drizzle ORM. It is for developers who already have a Worker or Pages project with D1 wired in and need repeatable generate–review–apply steps instead of guessing why a migration failed mid-apply. The skill stresses a critical gate: always read generated `.sql` before applying, because Drizzle can emit full table recreation when defaults change, producing INSERT statements that reference columns that do not exist on the old table. It documents how to simplify migrations to safe ALTER statements when you are only adding columns, then apply to local Wrangler D1 and remote with consistent commands. It also covers fixing stuck migration metadata and handling partial failures so you are not left with a broken remote schema. Use it whenever you are setting up initial D1 schemas, shipping schema iterations, or debugging migration errors after generate or deploy.776installs44Deep ResearchDeep Research is an agent skill for solo and indie builders who want evidence-backed decisions before they write code or lock a spec. Instead of improvising from training data, the workflow goes wide and deep: scan local projects for reusable patterns, study competitors and reviews, read forums, inspect plugin ecosystems, and compare technical architectures. You pick a depth—focused for a single choice like editor libraries, wide for a new feature or product landscape, or deep for a major build where nothing important is left unexplored. Each depth changes scope of ambition, not just timer length. The deliverable is a research brief that downstream planning and implementation skills can consume, reducing rework and wrong forks early in the journey.755installs45React Native Exporeact-native-expo is a production-oriented agent skill for solo builders shipping iOS and Android apps with Expo SDK 52 and above. It encodes React Native’s New Architecture requirements, React 19 deltas, and the sharp edges introduced in SDK 54—deprecated packages, Reanimated v4 shifts, and Swift-based iOS templates—so agents do not repeat common migration failures. Claude Code can auto-discover it from keywords like expo sdk, fabric, turbomodules, and version-specific RN releases. Use it when bootstrapping a new Expo app, enabling Fabric, or unblocking crashes after an SDK bump. It complements generic React skills by focusing on Metro, dev clients, simulators, and native bridge errors rather than web-only patterns.745installs46Resume Cover LetterResume Cover Letter is an agent skill that turns your background and a real job posting into application-ready documents—not a generic template dump. It is built for solo builders and indie operators who are applying for roles, consulting gigs, or founder-adjacent positions while shipping products on the side. Before writing, the skill expects a target title and company, the listing text or URL, mode (resume, cover letter, or both), region (AU/NZ, US, or UK), and enough candidate detail to substantiate bullets. It mirrors vocabulary from the posting, emphasizes measurable outcomes, and respects regional CV versus resume conventions. Special circumstances such as employment gaps, pivots, or relocation get explicit framing guidance rather than one-size-fits-all advice. Use it when you have a concrete opening to pursue; skip it when you only want a career brainstorm without a role to target.714installs47Proposal WriterProposal Writer is an agent skill for solo and indie service providers who need to win B2B work without sounding generic. It walks you through gathering client context, the ask, your recommended solution, numbers, deadlines, and relationship tone, then assembles a proposal that shows you understood the problem and have a credible plan. Sections cover project understanding (often lifted from discovery notes), scope, timeline, pricing presentation, and terms—each optional if it does not fit the engagement. The skill is triggered whenever you need a proposal, quote, project proposal, client proposal, SOW, statement of work, or engagement letter. It optimizes for clarity and trust on the page rather than flashy marketing copy, which helps independent builders compete with larger shops on professionalism alone.712installs48RoadmapRoadmap is a Claude Code–oriented agent skill that plans and executes entire application builds. In plan mode it generates a comprehensive technical roadmap detailed enough for hours of autonomous work per phase—not a vague strategy doc but an ordered delivery blueprint with actionable tasks. Start, resume, and status modes let you kick off execution, pick up after interruptions, and see which phase you are on. It is meant after deep research or when you have a product brief and before you say build this, especially for multi-week products spanning many sessions. The skill reads artifacts from your workspace, writes the roadmap, then drives phase-by-phase implementation including milestone commits, deployment, testing, and continuation until finished or stuck. Solo and indie builders use it to avoid ad-hoc chat planning and to keep a single source of truth for what to build next.707installs49Team Updateteam-update is an agent skill for drafting and delivering team status updates using whatever collaboration stack is actually loaded in the session. Solo builders wearing PM and eng hats use it when they want the agent to discover MCP chat and issue tools instead of assuming Slack or GitHub exist. The readme defines a capability-first detection strategy: list tools, match servers for messaging and issues, note threading support, and detect git locally when MCP does not expose version control. When chat MCP is missing, the skill still produces channel-ready text for manual paste. For indie teams on Claude Code with a mix of Google Chat, Linear, or GitHub MCP plugins, this procedural map reduces failed sends and vague "post an update" prompts. It is not a full incident comms playbook—it optimizes structured discovery and outbound updates tied to repo and tracker context during ongoing delivery.694installs50Gws Setupgws-setup is an agent skill that walks solo builders through greenfield configuration of Google's Workspace CLI so Claude Code can call Gmail, Drive, Calendar, Sheets, Docs, Chat, Tasks, and related APIs with a proper OAuth desktop flow. It is intentionally incremental: pre-flight commands detect whether the CLI, client_secret.json, and gws auth status already succeed, then the workflow only covers missing pieces from npm install through Cloud Console project and credential creation to authentication and skill installation. The skill targets first-time setup and auth troubleshooting rather than day-to-day message search or spreadsheet edits—those live in the downstream gws skills it installs. Prerequisites are Node.js 18+, a Google account, and console access. Because compatibility is Claude Code–oriented, other agents may need adapted paths for skill installation even if gws itself is generic. Successful runs yield a authenticated gws binary and a broad skill library for workspace automation across later Build, Grow, and Operate tasks.693installs51Gws Installgws-install is a Claude Code workflow skill for jezweb’s Google Workspace CLI pack. Solo builders who already completed gws-setup on one machine use it to bring the same OAuth client to a laptop reinstall, CI runner, or second desk without redoing Google Cloud project steps. The skill runs pre-flight commands, installs @googleworkspace/cli globally via npm, writes client_secret.json under ~/.config/gws, and finishes authentication so bundled gws agent skills work immediately. It explicitly defers newcomers to gws-setup when they lack credentials. Prism tags it under Build integrations because it connects your coding agent to Workspace APIs rather than shipping product features. Expect shell, network, and secret-handling steps; review scopes on the auth panel before trusting automation on production accounts.692installs52Stripe Paymentsstripe-payments is an agent skill for solo builders who need to accept money on a SaaS or ecommerce web app without getting lost in Stripe’s API surface. It walks through which Stripe product to pick—hosted Checkout for speed, Payment Element when you own the UI, subscription mode for recurring revenue, Setup Intents for saved cards, Connect for marketplaces, and Customer Portal for self-serve billing—and produces working integration code with webhook verification. Triggers include add payments, stripe checkout, subscription, pricing page, billing, and stripe webhook. It is most often used while building backend routes and frontend checkout flows, but the same skill supports validate-phase pricing page prototypes when you want real Stripe prices behind a landing experiment. No MCP server is required; the agent uses the Stripe npm SDK and your repo files. Start here before layering tax, invoicing edge cases, or Connect payouts.673installs53Product ShowcaseProduct-showcase is an agent skill that turns a running web app into a marketing website people can actually learn from. Solo builders with hard-to-explain tools—especially agentic or multi-step products—use it when a landing template would undersell the workflow. You choose quick, standard, or thorough depth; standard is the default multi-page home, features, and how-it-works path with captured screens. The skill emphasizes browsing the real product, recording sequences, and structuring pages from high-level positioning down to step-by-step demonstrations. Quick mode stays a fifteen-to-twenty-minute single page; thorough mode can spend hours on per-feature pages, comparisons, and doc-like demos. It is intermediate because you need a runnable app, coherent positioning, and time to review generated copy and assets. Pair it after Validate when you have a prototype worth showing, and before paid distribution pushes. Claude Code is the stated compatibility surface for browse-and-build automation.664installs54Db SeedDatabase Seed Generator is a Claude Code–oriented agent skill that reads your real schema and writes seed scripts you can run in dev, demo, or CI—not hand-rolled fake rows that violate foreign keys. It scans typical locations for Drizzle schema modules, Prisma schema, Drizzle/SQL migrations, and standalone SQL, then models tables, constraints, defaults, and relationships including JSON stored as TEXT on D1/SQLite. The workflow asks for purpose (dev, demo, testing), volume, and domain-appropriate sample data so generated fixtures feel plausible for your product niche. Output emphasizes idempotency and safe reruns, with ordering that respects referential integrity and awareness of D1 batch limits and uniqueness collisions. Solo builders use it when the app shell exists but empty databases block UI work, stakeholder demos, or automated tests. It accelerates backend setup; it does not replace migration authoring or production data governance.659installs55Strategy DocumentStrategy Document Writer is an agent skill for solo and indie builders who need business strategy in writing, not slide fluff. It walks through choosing one of four modes—SWOT, lean or full business plan, OKRs, or competitive analysis—then gathers context (name, industry, size, goals) before drafting. The bar is specificity: claims should be falsifiable and recommendations implementable within a stated horizon. Use it when you are validating what to build, pitching a partner, entering a market, or resetting quarterly focus. It fits Claude Code, Cursor, and similar agents as procedural knowledge packaged as SKILL.md, and pairs naturally with later build and launch work once scope is locked.649installs56Drizzle Orm D1Drizzle ORM for Cloudflare D1 is an agent skill package for solo builders shipping serverless APIs and SaaS on Workers who want TypeScript-first SQL without SQLite foot-guns. It walks through schema definition, Drizzle Kit migrations, relations, prepared statements, and the D1 batch API patterns that matter at the edge. The skill is explicitly aimed at preventing repeated production failures—transaction semantics that break on D1, foreign key surprises, cascade deletes, and SQLite variable limits during bulk inserts. Use it when you are standing up a new D1 database, evolving tables safely, or debugging ORM-generated SQL that fails at runtime. It pairs naturally with a Workers backend stack and assumes you are comfortable with TypeScript migrations rather than raw SQL-only workflows. Prism lists it for builders in the Build phase who need a disciplined ORM layer instead of ad-hoc D1 queries copied from forums.648installs57Cloudflare ApiCloudflare API is a procedural skill that documents REST patterns for Cloudflare’s developer platform—D1, R2, KV, Workers, Vectorize, Queues, and Durable Objects—using the v4 account API. Solo builders who already use wrangler for day-to-day deploys reach for this skill when they need scripted bulk operations, direct SQL against D1, debugging queries with parameters, or automation that spans multiple services under one account ID. Each section pairs curl examples with Authorization Bearer headers and jq parsing, including parameterized D1 queries to reduce injection risk. The skill does not replace wrangler for local dev and publish flows; it extends what you can do from CI, one-off terminals, or agent-driven runbooks. Expect intermediate familiarity with API tokens, account IDs, and Cloudflare service boundaries. It is especially useful while integrating edge data stores into a SaaS or API product and when operating cron-style maintenance without opening the dashboard for every change.647installs58App DocsApp-docs is an agent skill for solo builders who need publishable user documentation without manually clicking every path and pasting screenshots into a doc. It detects available browser automation tooling, resolves the best URL to exercise the app, then walks screens while generating structured prose: quick-start paths for shallow passes and full guides with CRUD flows, mobile views, and troubleshooting when you choose thorough or exhaustive depth. The skill explicitly positions itself against screenshot dumps—it wants teaching-oriented sections with annotated visuals and reference tables. Standard depth targets on the order of thirty captures and roughly thirty to sixty minutes of agent time, while exhaustive mode aims at a documentation suite suitable for external users. It is marked claude-code-only for compatibility, so plan your agent environment accordingly. Use it late in Build when features are stable enough to document, or refresh docs before a Launch push when onboarding copy must match the live UI.633installs59Onboarding UxOnboarding UX is a Claude Code–oriented agent skill that closes the gap between working features and a welcoming first login. It instructs the agent to resolve a live or staging URL, run a browser-based walkthrough using the same MCP detection patterns as ux-audit, and identify moments where new users face empty data, unexplained navigation, or hidden capabilities. Rather than stopping at critique, the skill produces concrete guidance content and implementation-ready UI patterns—onboarding checklists, empty-state messaging, progressive tooltips, feature tours, sensible defaults, and contextual help tied to the actual screens observed. Solo builders shipping SaaS or browser extensions use it after core features exist but before launch traffic arrives. It complements ux-audit in a two-step chain: structural UX issues first, onboarding-specific solutions second. Triggers include phrases like new user experience, empty states, feature tour, or when the app feels confusing on day one.629installs60Hono RoutingHono Routing is an agent skill for solo builders who want a single, modern HTTP layer that runs everywhere from Cloudflare Workers to Node without relearning framework quirks each deploy. The skill consolidates routing, middleware composition, validation with Zod or Valibot, RPC typing for end-to-end clients, and streaming interfaces including SSE and WebSockets—common needs for indie SaaS APIs and agent backends. Auto-trigger keywords help agents surface the skill when errors mention middleware response typing, validator failures, or JWT verification requirements. It assumes you are actively coding handlers rather than drafting product specs, and it pairs well with edge-first stacks where cold start and bundle size matter. Documented production status and a 2025-10-22 refresh signal maintained troubleshooting notes for real-world Hono releases. Use it when scaffolded Express knowledge fights your target runtime or when you need disciplined validation hooks before Ship-phase security review.624installs61Design SystemDesign System is a Claude Code skill that extracts a complete, semantic design system from an existing website, HTML file, or screenshot and writes it to DESIGN.md. Solo builders use it when they want a site's visual language documented before generating new pages—whether bootstrapping a clone, formalising undocumented UI, or prepping .design/DESIGN.md for the design-loop skill. The workflow starts by identifying the source (live URL via Playwright or scraper, local HTML read, or visual capture), then inspects markup and visuals for tokens and component patterns. It emphasises atmosphere and consistency across multi-page projects and can pull cues from Google Stitch-style sources. Tools include Read, Write, Edit, Bash, Glob, and Grep with claude-code-only compatibility, so browser automation assumptions fit Claude Code agents rather than every IDE.623installs62Award ApplicationAward Application Writer is a procedural skill for solo founders and indie operators who need Telstra-style business awards, chamber prizes, industry grants, or competition entries that actually match how judges score. It walks the agent through collecting the exact criteria and limits, optionally pulling organiser pages for rubrics and past winners, then mapping achievements to each criterion in a table before drafting. That order prevents the common failure mode of beautiful prose that never addresses weighted score lines. The skill is for Grow-phase content when you are demonstrating merit against defined rules—not for inventing product specs. Resubmissions and first-time entries are both in scope when the user supplies real metrics and stories. Treat it as a structured writing workflow rather than a one-shot prompt so every section earns points on the evaluator’s sheet.603installs63Nz Business EnglishNZ Business English is a writing-style skill for anyone communicating with a New Zealand business audience. It encodes EN-NZ spelling (colour, organise, centre), noun–verb pairs (licence vs license), and a voice that is professional yet approachable—inclusive by default without sounding like a caricature of Kiwi culture or a stiff corporate template. Solo builders use it when drafting or revising emails, Slack messages, proposals, support replies, blog posts, and marketing pages where local credibility matters. Because the trigger is audience-based rather than feature-based, it travels across Validate landing copy, Build docs, Launch distribution, Grow lifecycle messaging, and Operate support. Invoke it whenever you are producing or polishing text for NZ clients, partners, or regulators and want consistency without hiring a local copy editor for every send.591installs64Project Planningproject-planning is an agent skill that turns fuzzy web project goals into structured, agent-safe planning packages centered on IMPLEMENTATION_PHASES.md. Solo builders invoke it when starting a new codebase, landing a major feature, or splitting oversized work into phases that each have verification steps and clear exit conditions—so Claude or Cursor does not drown in context or skip checkpoints. The skill emphasizes practical PM detail for indie stacks: per-phase file maps, Mermaid data-flow sketches, internal and external dependency callouts, and a gotchas section that captures binding and config surprises before implementation begins. Version 1.1 specifically strengthens file-level templates across phases and ties planning output to SESSION.md for continuity between sessions. It is not a substitute for product discovery or automated deployment; it is the scaffolding that makes downstream coding passes predictable. Use it when you want phased truth on disk that agents can re-read instead of re-negotiating scope in chat.584installs65Wordpress Plugin CoreWordPress Plugin Core is an agent skill for solo and indie builders who extend WordPress instead of only configuring themes. It walks through plugin bootstrap, hook registration, database reads and writes with prepared statements, the Settings API for admin options, custom post types and taxonomies, meta boxes, and REST routes with sane permission checks. You get explicit architecture choices—minimal procedural plugins, class-based OOP, or PSR-4 autoloaded structure—so your repo stays maintainable as features grow. Security guidance aligns with real-world patch stacks: input sanitization, output escaping, capability checks, and CSRF nonces, with pointers when wpdb::prepare or REST callbacks fail. Use it when you are greenfielding a plugin or debugging vulnerabilities and compatibility after WordPress core updates through the 6.7–6.9 line.569installs66Nemoclaw Setupnemoclaw-setup is a Claude Code–oriented workflow skill for installing and configuring NVIDIA NemoClaw, a sandboxed agent platform built on OpenClaw inside Docker and k3s. Solo builders reach for it when they want a self-hosted agent with a web UI and terminal CLI, network policies for outbound access, and Nemotron models without wiring everything by hand. The skill sequences pre-flight validation, dependency installation, sandbox creation, OpenShell integration, and optional Cloudflare Tunnel remote access, including explicit workarounds for cgroup and Docker edge cases on real Linux servers. It assumes Ubuntu 22.04 or newer, Node 20+ (22 recommended), and an NVIDIA API key from build.nvidia.com when using hosted models. Use it during Build when standing up agent tooling on a Spark or DGX-class machine—not as a daily coding assistant for application features.558installs67Fork DisciplineFork Discipline is an agent skill for solo and indie builders running one codebase across multiple client deployments. It audits the core/client boundary so shared platform logic stays in core while tenant-specific config, content, schema, and custom features live under per-client trees. The agent reads the repo with Glob and Grep, flags anti-patterns like inline client branching, config files that override instead of merge, and migrations that collide with shared numbering conventions. You get a structured boundary map, violation report, and refactoring plan, with optional FORK.md documentation and scripts to realign the tree. Use it before onboarding a new white-label customer, after a rushed fork, or when you cannot tell whether a change belongs in platform or client code. It is Claude Code oriented via allowed-tools metadata and fits teams shipping multi-tenant or multi-brand SaaS from a single repository.551installs68Seo Metaseo-meta is an agent skill that produces end-to-end page metadata for solo builders launching sites on modern stacks. It generates title patterns, meta descriptions, Open Graph and Twitter Card fields, canonical links, and JSON-LD structured data types including LocalBusiness, Service, FAQ, and BreadcrumbList—matching how Jeremy Dawes packaged the jezweb claude-skills workflow for search and social parity. Invoke it when you are building new routes, adding share cards, or implementing schema.org markup without copying boilerplate from five blog posts. The skill aligns with Launch-phase SEO and GEO readiness while still fitting Build when you embed tags in templates. It is a generator-style package focused on HTML head output rather than running audits against live Search Console. Expect concrete tag snippets you can drop into frameworks that support server or static head injection.546installs69Tanstack Routertanstack-router is an agent skill for solo builders creating TypeScript React SPAs that need file-based, type-safe navigation instead of ad-hoc React Router setups. It walks through TanStack Router patterns—route trees, loaders, search-param validation, authenticated beforeLoad guards, error boundaries, and optional virtual file routes—while calling out integration with TanStack Query and Cloudflare Workers-style deployment contexts from the upstream skill README. The skill is explicitly oriented toward fixing and preventing failure modes: roughly twenty documented issues such as lost validation structure, param parsing mistakes, devtools dependency breaks, Vite bundling surprises, and SSR streaming crashes. Use it during Build when you are scaffolding routes, tightening types, or debugging routing in production-like Docker or worker environments. It is not a generic CSS or component library skill; it assumes you already chose TanStack Router and need procedural guardrails so navigation stays correct as the app grows.546installs70Design LoopDesign Loop is a Claude Code–oriented workflow skill for solo builders who want a whole small site built page by page without re-explaining context each time. You seed .design/next-prompt.md with the first task; each iteration reads that baton, produces HTML/Tailwind for one page, wires it into navigation and internal links, optionally checks layout with browser automation, refreshes site documentation, then writes the next prompt so the loop can continue when you say “keep going” or run unattended. It fits early product sites, landing clusters, and docs-style multi-page builds where consistency matters more than a one-shot single-page mockup. The skill is not a design system generator for React SPAs by itself—it targets baton-driven static or Tailwind site assembly. Pair it with your brand constraints in the initial baton so later pages stay on-theme.545installs71Brains TrustBrains Trust is an agent skill that sends your question—code, architecture, strategy, prompting, or writing—to other leading models for a structured second opinion. Indie builders running Claude Code use it when one assistant’s take feels thin, risky, or untested, especially before merging, choosing infra, or locking a product bet. It wires OpenRouter, Gemini, and OpenAI with patterns for a single consult, multi-model consensus, or deliberate devil’s advocate, and it defaults to two-provider consensus when you invoke it without extra instructions. Session context steers the mode toward code review, architecture, or debugging. You need API keys and network access; the skill is marked Claude Code–oriented. Treat outputs as advisory input to your own judgment, not a substitute for tests, security review, or production monitoring.539installs72Firecrawl Scraperfirecrawl-scraper is an agent skill package that teaches coding agents how to use the Firecrawl API to turn live websites into markdown or structured data suitable for LLMs. Solo and indie builders install it when they need reliable scraping, multi-page crawls, URL maps, combined search-and-scrape flows, schema-guided extraction, batch URL processing, or change monitoring—without maintaining headless browsers and parser glue themselves. The skill aligns with Firecrawl v2, points to official documentation, and covers practical concerns such as dynamic JavaScript pages, common anti-bot friction, and document formats beyond HTML. It is aimed at Claude Code, Cursor, Codex, and similar agents building research pipelines, competitor snapshots, content ingestion, or brand-reference extraction. Use it during Build when integrating external data sources; it also supports Idea and Validate research passes and Operate-style content drift checks via change tracking.527installs73Ai Sdk UiAi-sdk-ui is an agent skill for solo builders shipping AI-native web apps on Next.js and React with the Vercel AI SDK v6 frontend layer. It concentrates on the hooks product teams touch daily—useChat for streaming conversations, useCompletion for single-shot text, and useObject for structured streaming JSON—plus the message-parts model and tool-approval flows that separate demo chat from production UX. The readme explicitly targets migration from v5 to v6 and a long tail of failure modes: failed to parse stream, undefined state reads, maximum update depth loops, and race conditions when multiple requests overlap. That makes it valuable when you already have a /api/chat route but the UI flakes in Strict Mode or after deploy. Treat it as the UI companion to server-side AI SDK skills: wire endpoints first, then invoke this skill to harden the client. Production-ready framing assumes Next.js 14+/15+ and current stable SDK pins.526installs74AccessibilityAccessibility is an agent skill that teaches solo builders and small teams how to implement WCAG 2.1 AA–compliant web interfaces using semantic HTML, correct ARIA, focus management, and screen reader–friendly patterns. It is framework-agnostic but includes React-oriented examples for forms, modals, dropdowns, tabs, toasts, and data tables. Use it when you are building or refactoring UI, fixing “focus outline missing,” “insufficient contrast,” or “aria-label required” errors, or preparing a site for real keyboard and assistive-tech use—not when you only need a one-line lint rule. The skill bundles copy-paste patterns (skip links, live regions, focus traps), troubleshooting flows, and checklists so agents produce shippable markup instead of generic “make it accessible” advice. It pairs naturally with pre-ship testing and review passes without replacing formal third-party audits.511installs75Playwright Localplaywright-local is a production-oriented Claude skill for driving real browsers on your machine with Playwright instead of brittle shell curls or rented cloud grids. Indie builders use it when they need trustworthy end-to-end checks before ship, daily scraping with sane session and cookie handling, or visual artifacts such as screenshots and PDFs from dynamic pages. The skill emphasizes pitfalls that waste solo time: CI jobs that hang on browser boot, extension testing gaps, and Linux-specific browser dependencies. It also covers stealth-oriented scraping when sites throttle datacenter IPs, plus authenticated flows and infinite-scroll content. Documented as processing high page volume in real scraping systems, it bridges Ship testing and Build integration work. Pair it with your app stack and CI so agents generate stable selectors, local launch flags, and debugging steps rather than one-off Puppeteer snippets.508installs76Skill CreatorSkill Creator is an agent skill for solo and indie builders who package reusable procedural knowledge into Claude Code skills. It explains what skills provide—multi-step workflows, API and format integrations, company-specific schemas, and bundled assets—and why brevity matters when every token competes with conversation history and other skills. The skill should run whenever you want to create a new capability or refresh an existing SKILL.md, before you dump long tutorials into the context window. It aligns with Prism’s mission to surface trustworthy agent extensions by helping you ship focused, self-contained skills that other builders can discover and install. Intermediate complexity: you need enough product sense to scope a domain and enough discipline to cut redundant explanations. Use it across Idea through Operate whenever you formalize how your agent should behave for a recurring job.504installs77Tanstack Tabletanstack-table is an agent skill by jezweb that guides production headless data tables using TanStack Table v8, with emphasis on server-driven pagination, filtering, and sorting for indie SaaS dashboards and admin tools. Solo builders on Cloudflare Workers and D1 often hit duplicated fetch logic, broken controlled state, and React 19 friction; the skill packages patterns that coordinate TanStack Query keys with table state and optional virtualization for large datasets. Auto-trigger keywords cover datagrids, API-backed sort/filter, D1-backed tables, and performance work on long lists. It is framed as error prevention—twelve documented failure modes—rather than a generic table tutorial, which speeds Build-phase delivery when you already have an API shape in mind. Use it when you are wiring UI to your backend, not when you only need design mockups or SEO landing copy.504installs