Coding & Development

Configures, writes and upgrades TypeScript 6.0 code and tsconfig.json from the official TypeScript documentation. Use when upgrading from TypeScript 5.x to 6.0 (strict, types, rootDir, baseUrl, moduleResolution node, target es5, outFile, esModuleInterop), choosing module and moduleResolution, writing package.json exports and imports, fixing ESM and CommonJS interop, enabling verbatimModuleSyntax or erasableSyntaxOnly, setting up project references and tsc --build, writing .d.ts files, checking JavaScript with JSDoc, or debugging narrowing, generics, mapped types, decorators and enums. Triggers on TypeScript, tsc, tsconfig, TS5112, ignoreDeprecations, nodenext, bundler, .d.ts, @ts-check, TypeScript 6 upgrade.

Framework de création de projet digital en 5 étapes : Structurer, Outiller, Construire, Livrer, Élargir. Appliquer systématiquement au démarrage de tout nouveau projet web, app, SaaS, ou outil interne. Déclencher dès que l'utilisateur mentionne un nouveau projet, un MVP, un prototype, un produit à construire, une app à développer, ou demande comment organiser le développement d'un produit. Également pertinent quand l'utilisateur a une idée de produit et veut passer à la réalisation, ou quand il se demande par où commencer. Ce skill prend la priorité sur les approches ad-hoc ou les scaffoldings improvisés.

Tauri 2 apps, tauri.conf.json, capabilities and permissions, commands and invoke, events and channels, plugins, bundling, signing and the updater, migration from Tauri 1. Guides Tauri 2 work from the official v2 documentation, including scopes, managed state, plugin authoring, CSP and the isolation pattern, testing and debugging, Linux, Windows and macOS packages, and mobile and store releases. Use when writing or reviewing a Tauri 2 app, adding a plugin or a capability, fixing a "not allowed" or "forbidden path" error, wiring IPC between Rust and the webview, packaging, signing or publishing a desktop or mobile build, or upgrading a Tauri 1 or 2.0 beta project.

Teaches Node.js practice drawn from the official Node.js API documentation, with pitfalls, surprising defaults, deprecations and replacements, each backed by its nodejs.org URL. Use when writing, reviewing, debugging or upgrading Node.js code, or choosing a built-in module (fs, crypto, streams, http, process, CLI flags, node:test, ESM/CommonJS, package.json exports, TypeScript type stripping, child_process, worker_threads, node:sqlite, deprecations).

Set up, update or back up the user's own copy of the acquisition toolkit. Publishes their private Acquisition Scorecard (and, if they want them, the Deal Room, Listing Digest and partner quiz) as claude.ai Artifacts, seeds their config, and records the links in a private my-search/ folder. Use when the user says "set me up", "install the toolkit", "/acq:setup", "update my toolkit pages", or "back up / sync my config".

Teaches current Vite 8 practices drawn from the official documentation. Use when writing or fixing a vite.config, migrating from Vite 7 (Rolldown and Oxc replacing Rollup and esbuild), configuring the dev server, proxy or HMR, handling import.meta.env and modes, tuning the production build, library mode, SSR, writing a Vite plugin, using the HMR or JavaScript API, or the Environment API. Triggers on vite.config, defineConfig, rolldownOptions, import.meta.env, loadEnv, optimizeDeps, ssrLoadModule, ModuleRunner, this.environment, upgrade to Vite 8.

Teaches current ESLint 10 practices drawn from the official documentation. Use when writing, reviewing or fixing an eslint.config.js (flat config, defineConfig, files, ignores, globalIgnores, extends, languageOptions, globals, parsers, processors), configuring ESLint rules and severities, disable comments and inline config, running the CLI in CI (--fix, --cache, --quiet, --max-warnings, bulk suppressions, formatters, feature flags, MCP server), writing custom rules (meta, schema, fixes, suggestions, selectors, scope, code paths) and testing them with RuleTester, building plugins, shareable configs, custom parsers, processors or formatters, using the Node.js API (ESLint, Linter), or migrating to ESLint 9 or 10 from eslintrc. Triggers on eslint, eslint.config.js, flat config, .eslintrc, .eslintignore, @eslint/js, js/recommended, eslint-disable, no-unused-vars, no-restricted-imports, no-restricted-syntax, RuleTester, eslint-plugin, eslint-config, FlatCompat, context.getScope is not a function, ESLint 10.

Screen new business-for-sale listing alerts (BizBuySell, BizQuest, Acquire and similar) against the user's gates, load the survivors into their Acquisition Scorecard facts-only, and write a dated digest to their Listing Digest page. Use when the user says "run my listing digest", "check new listings", "/acq:listing-digest", or from a scheduled task the user set up for it.

Use when asked for an illustration, a thumbnail or a card picture for a card in a grid of content, such as a library of resources (guide, template, deck, article), when a set of cards needs one consistent treatment, or when product screenshots are shared as a reference for card illustrations. Two modes, "noir et blanc" (ink accent) or "accent couleur" (ten-hue palette), chosen by the argument or the wording of the request.

Writes, reviews and upgrades React 19.3 code from the official react.dev documentation. Use when writing components and Hooks, removing unnecessary useEffect calls, fixing Effect dependencies with useEffectEvent, building forms with Actions (useActionState, useOptimistic, useFormStatus, form action), using Server Components, 'use client', 'use server' and Server Functions, Suspense, use(), useTransition, useDeferredValue, ViewTransition, Activity, adopting React Compiler and eslint-plugin-react-hooks, deciding on useMemo, useCallback and memo, server rendering and hydration (createRoot, hydrateRoot, renderToPipeableStream, prerender), or migrating from React 18 (forwardRef, Context.Provider, ReactDOM.render, propTypes, defaultProps). Triggers on React, JSX, hooks, useState, useEffect, useRef, useContext, Server Components, RSC, Actions, React Compiler, Strict Mode, hydration mismatch, React 19 upgrade.

Develop, debug, and validate Primus-Turbo operators and modules on AMD GPUs. Covers the layered architecture (ops / kernels-dispatcher / Triton / HIP-CK csrc / modules), how to add or change a feature end-to-end, accuracy verification (SNR, tolerances, reference implementations), performance benchmarking, the backend dispatch system, and build/test/bench commands. Use for any Primus-Turbo development task (GEMM, Attention, GroupedGEMM, MoE, quantization, normalization, activation) and for accuracy or performance validation.

OS personnel avec Claude Code. À utiliser pour créer ou faire évoluer un second cerveau (un « brain »), un dossier de fichiers Markdown qui sert de mémoire et de tableau de bord à Claude Code. Déclencher quand on veut structurer son travail, traiter une inbox (comptes rendus, transcriptions, notes), mettre en place des règles d'ingestion, des agents, des commandes ou des vues générées, préparer un 1:1, tenir un registre de décisions, ou propager un changement dans tous ses contenus.

Pitfalls, surprising defaults and recommended patterns of Playwright Test for JavaScript and TypeScript, distilled from the official docs. Use when writing, reviewing, debugging or configuring Playwright tests (@playwright/test, playwright.config.ts, npx playwright test), including flaky tests, locators (getByRole, filter, strict mode), web-first assertions (expect, toHaveText, toBeVisible, expect.poll, toPass), fixtures, storageState authentication, project dependencies, timeouts, retries, network mocking (page.route, routeFromHAR, routeWebSocket), API testing, clock, aria snapshots, toHaveScreenshot, axe accessibility scans, traces, UI mode, reporters, CI, Docker, sharding, test locks, component testing with stories and mount, codegen, Playwright MCP, playwright-cli and migration from Puppeteer, Protractor, Testing Library or older Playwright versions.

Teaches current JavaScript (ECMAScript) practice drawn from the MDN JavaScript documentation. Use when writing, reviewing or fixing modern JavaScript, choosing a built-in API, or replacing a deprecated pattern. Covers equality and coercion, let/const and the temporal dead zone, strict mode, ASI, functions, classes and private fields, objects, prototypes, Proxy, Reflect, Symbol, eval, operators (??, ?., spread, BigInt arithmetic), statements, Promise, async/await, Promise.try, iterators and iterator helpers, generators, using/await using, DisposableStack, WeakRef, Array (toSorted, at, groupBy, fromAsync), Map, Set methods, WeakMap, strings and Unicode, encodeURIComponent, RegExp (v flag, RegExp.escape, lastIndex), Number, Math, BigInt, JSON (rawJSON, revivers), typed arrays, ArrayBuffer, base64, SharedArrayBuffer, Atomics, Temporal, Date, Intl (DateTimeFormat, NumberFormat, Collator, Segmenter, PluralRules), errors (cause, Error.isError), ES modules, import attributes, import.meta, dynamic import.

Prepara un plugin de WordPress hospedado en plugins.svn.wordpress.org para internacionalización completa y publica el release al SVN. Cubre: reescribir strings hardcoded en español a inglés (PHP y JS), generar .pot con wp-cli, crear traducción inicial .po/.mo (ej: español neutro), bumpear versión sincronizada, commit al trunk y crear tag. Usar cuando el plugin fue recién aprobado en WP.org, cuando los strings traducibles están en español y bloquean a otros locales, o cuando se quiere añadir/actualizar traducciones y subir un nuevo release.

Builds the information pack Apple asks for when App Review replies "We need additional information to better understand the app". It covers all six required items: (1) a screen recording on a physical device, as a shot list the user records; (2) the app's purpose and target audience; (3) setup and main features, with any login credentials or sample files; (4) the external services the app uses; (5) regional differences; (6) regulated industry or protected third-party material. It also runs Apple's "Prevent common issues" checks. Every answer comes from the app's own code (accounts, user-generated content, in-app purchases, SDKs, locales, screen-capture blocking, data deletion). It reads the connected iPhone's iOS version and checks it is the latest. It ends by telling the user to paste the reply into App Review Information → Notes. Use it when the user pastes that App Review message, says "/new-app-info", "we got rejected, they want more info", "answer App Review", or "make the App Store review pack". It w...

Develop inside an AINAS Coder workspace and ship the result as a NAS app: run commands in the isolated container, manage projects and project instances, trigger builds, and download or install the resulting .zpk. Use when the user asks to build, test, package, or install a project through Coder.

Expert drafting and review of Delaware C-corporation startup legal documents from formation through Series A: certificate of incorporation, bylaws, incorporator and board consents, founder stock purchase agreements, IP assignments, 83(b) elections, equity incentive plans and option/RSU/RSA grants, offer letters, CIIAAs, consulting/advisor/separation agreements, NDAs, SAFEs and side letters, convertible notes, term sheets, NVCA-style priced-round documents, board and stockholder consents, Section 204 ratifications, charter amendments, and compliance (franchise tax, foreign qualification, Form D and blue sky, 409A, Rule 701, ISOs, QSBS). Use when someone forms or runs a Delaware corporation, issues equity, hires, raises money, or asks to draft, fill in, review, or explain any of these documents or deadlines.

Builds, configures and debugs Socket.IO 4.8 servers and clients from the official Socket.IO v4 documentation. Use when writing a Node.js Socket.IO server (Express, Fastify, Koa, Bun, uWebSockets.js) or a socket.io-client app, emitting events with acknowledgements, timeout() and emitWithAck, broadcasting to rooms and namespaces, authenticating in middleware with the auth option, configuring CORS, reverse proxies and sticky sessions, scaling with the Redis, Redis Streams, Postgres, MongoDB or Cluster adapter, enabling connection state recovery and delivery guarantees, tuning reconnection, writing tests, troubleshooting connection issues, or migrating from Socket.IO v2 or v3. Triggers on Socket.IO, socket.io, socket.io-client, io(), new Server, emit, emitWithAck, acknowledgements, rooms, namespaces, broadcast, io.use, middleware, CORS, Redis adapter, sticky sessions, connection state recovery, reconnection, connect_error, Session ID unknown. Not for plain WebSocket servers such as ws without Socket.IO.

Runs desk research and writes a technical guide to .knowledge/{domain}/{topic}/guide.md in the open project. Reads at least five high-relevance primary sources, records the tool that found each source, and draws the macro and micro process. Use when the user says /learn, asks for a study guide or desk research, or gives a subject or link to study.

Run a grilling interview on a local browser page instead of the terminal. Every question is laid out with its recommendation, answerable in any order, with a per-question discussion thread and one "Send to Agent" button. Use when the user says "grill with ui", invokes /grill-with-ui with a topic, or says "/grill-with-ui resume".

Runs a cEDH (competitive Commander) playgroup on top of TopDeck.gg tournament data. Use it to add players and their decks, pull each player's tournament results from TopDeck.gg, report the recent meta (last month by default), log practice games, notes and who can make a session, and build practice pods that face the top meta, stay balanced, avoid repeat pairings and target each player's weak matchups, and post pods, meta reports, results and "who's in?" calls to the group's Discord channel. Use it whenever the user mentions their playgroup, pod or practice night, testing partners, TopDeck stats for people they play with, "what's the meta right now", a practice game result to record, seating people for cEDH, or posting any of that to Discord, even if they don't say "skill" or "TopDeck".

Independent review stage of Full-Stack WP Developer. Reviews a finished WordPress plugin, theme, add-on, or Core patch through a security lens (WordPress-specific vulnerabilities, REST and AJAX authorization with an API risk map, AI-generated-code risks), a quality lens (error handling, data and upgrade safety, performance, i18n, accessibility, comment accuracy, test gaps, class and type design, scope), and an optional lean lens for trimming over-engineering, reporting only high-confidence findings in a fixed format. Invoked by fswp-developer after the verify gate, normally through the fswp-code-reviewer agent; consult directly to review or optimize any WordPress code.

Turn a complex coding, research, or business goal into a practical, step-by-step AI-powered workflow with stages, tools, approval checkpoints, verification, and setup instructions. Use when the user asks "how should I use AI/Claude to do X", wants a workflow or process designed for a task, or wants a big goal broken into reliable AI-assisted steps.