> Computer Hardware Engineers

Enforce strict TypeScript: generics, utility types, discriminated unions, and type guards.

Generate custom React hooks — debounce, fetch, local-storage, media-query, and more.

Scaffold a Model Context Protocol server in TypeScript or Python — tools, resources, prompts wired in.

Thorough PR review — security, performance, correctness, style, with inline-comment-ready output.

Automate browser interactions — web testing, form filling, screenshots, scraping, multi-tab workflows.

Build production-grade Claude API apps with prompt caching, tool use, streaming, and model migration baked in.

Read, extract, merge, split, OCR, and fill PDF forms — all from a single Claude skill.

Drive a real browser to fill forms, scrape data, or test flows. Vision + DOM hybrid.

Design and ship n8n automations — node selection, credential setup, error handling, deploy-ready JSON.

AI commit messages that read like a senior engineer wrote them — Conventional Commits, why-not-what focus.

Scaffold a production-ready SKILL.md from a one-line description. The skill that builds skills.

Generate a README that earns GitHub stars — hero, features, install, quickstart, screenshots, badges.

Generate Claude Artifacts (HTML, SVG, React, Mermaid) the right way — without the common rendering pitfalls.

Bun-native apps — bun.serve, sqlite-native, test runner, package manager, hot reload, single binary.

Generate .cursorrules / .cursor/rules.mdc files that actually teach Cursor your codebase conventions.

TanStack Query (React Query) — queries, mutations, optimistic updates, infinite scroll, prefetch.

Build Figma plugins — manifest, UI, plugin API, network access, and publish-to-Community pipeline.

Run the app, drive the UI, confirm the change actually works — not just that tests pass.

Generate PR descriptions reviewers actually read — summary, test plan, screenshots, breaking changes.

CI/CD workflows that scale — reusable workflows, matrix builds, caching, OIDC, secrets, releases.

Wire Firebase right — Auth, Firestore, Realtime DB, Crashlytics, Analytics, Cloud Messaging — via the BoM.

Pick the right Claude model per task — Fable 5 vs Opus 4.8 vs Sonnet 4.6 vs Haiku 4.5, with cost math.

Shrink images to 25% of their size so AI engines (Claude, GPT, Gemini) read them faster and cheaper.

Wire Hilt dependency injection the right way — modules, scopes, ViewModels, WorkManager, testing.