Claude Skills by FluxonLab

Build production UI: component APIs, React state/batching, TS/JS patterns, Tailwind v4 tokens, shadcn/ui, motion.

Test-first work: unit, integration and API tests (Jest, Vitest, pytest, Bats, Temporal), contracts, quality playbook.

Advanced Git: rebase, cherry-pick, bisect, worktrees, reflog recovery.

Design agent systems: roles, handoffs, tool boundaries, routing; LangGraph, Copilot SDK, MCP server scaffolds.

Prompts: write and review, safety and bias, hallucination risk, final execution prompts, AI studio prototypes.

RAG and retrieval: chunking, embeddings, vector and hybrid search, ranking, grounding.

LLM evals: rubrics, datasets, automated metrics, human review, regression checks.

Design APIs and interfaces: REST/GraphQL principles, OpenAPI 3.1 specs, contracts, spec-to-code generation.

Review server-side code: correctness, validation, maintainability, third-party, webhook and queue boundaries.

Auth: sessions, cookies, tokens, origins; authorization, roles, row-level access, tenancy, denied paths.

Errors and observability: error patterns, logs, metrics, traces, OpenTelemetry, dashboards, safe diagnostics.

Python: uv envs, packaging/PyPI, style, anti-patterns, asyncio, jobs, config, errors, resilience, FastAPI.

Cloud and infra: service choices, Terraform/OpenTofu, Kubernetes/Helm, hardened multi-stage Dockerfiles.

Inspect a repo before edits: package manager, scripts, entrypoints, env, DB markers, repro, module and data-flow map.

Review architecture: coupling, duplication, integration fit, keep/adapt/replace; record decisions as ADRs.

Specs and plans: spec-first briefs, file-level plans, feature breakdown, incremental slices, plan updates, estimates.

Refactor safely: scope, preserve behavior, cut complexity, sequence multi-file refactors, verify the changed surface.

Release gates: demo/handoff/deploy readiness, env and DB preflight, canary and rollback, hotfix, game release.

Review a diff before commit or merge: assumptions, simplicity, surgical scope, regression surface, impacted tests.

Set up project agent context: minimal AGENTS/rules files, session context, recovery when output quality degrades.

Data and ML pipelines: ETL/ELT, data quality, lineage, training reproducibility, model serving, notebooks.

Database changes: Prisma schema, safe reversible and zero-downtime migrations, backups, seeds, fixtures.

Postgres/Supabase: table design, indexes, RLS, Edge Functions, query plans, N+1, pagination, SQL tuning.

CI/CD: workflow triggers, caches, matrices, artifacts, release gates, GitHub Actions hardening.

Review built UI: hierarchy, consistency, polish, responsive layout, dashboards, design-system drift, prototypes.

Audit accessibility to WCAG 2.2: semantic HTML, keyboard, focus, contrast, labels, screen readers.

Conversion: landing pages, offer clarity, trust, CRO, forms, signup and onboarding activation, page copy.

Localization: locale parity, missing keys, i18n setup, hreflang and multilingual SEO, cultural adaptation.

Next.js App Router: server/client component boundaries, caching, streaming, parallel routes, data fetching.

3D in the browser with Three.js or React Three Fiber: scenes, GLTF, lighting, raycasting, instancing.

Game engineering: loop/ECS, code review, tech debt, frame budget, profiling; Godot, Unity, Unreal, Phaser, PixiJS.

Game design: concept, GDD, systems, balance, levels, narrative, audio, live ops, HUD and menu UX, prototypes.

Review mobile apps: architecture, navigation, state, release surface, iOS HIG and Android Material conventions.

Ecommerce: catalog, checkout, payments, inventory, orders, taxes; WordPress/WooCommerce plugins and themes.

NFC card, linktree, menu builder, QR and tap-to-open flows, plus print-ready artifacts and export.

Local LLM runtimes: model routing, resource limits, privacy boundaries, offline behavior.

Review Chrome extensions: MV3 manifest, content scripts, service workers, permissions, store readiness.

Maps: routing, geocoding, markers, clustering, location privacy, MapLibre styling and tile sources.

Analytics: tracking plans, event naming, GA4/GTM setup, conversions, attribution, privacy and consent.

SEO: technical and on-page audits, metadata, schema, sitemaps, E-E-A-T, content ops, clusters, AI search/GEO.

Cost control: LLM token and API spend, model choice, subagent fanout, cloud and hosting waste.

Monorepos: pnpm workspaces, Turborepo caching, Nx, Bazel, build graph, package boundaries.

Web performance: Core Web Vitals (LCP, INP, CLS), budgets, bundle size, Lighthouse CI, measure then fix.

Backend performance: p95/p99 latency, flame graphs, memory and FD leaks, Python profiling, caching strategy.

Requirements: user stories, acceptance criteria, scope, prioritization, outcome roadmaps, pre-mortem risks.

Market and customer research: competitors, audience, positioning, interviews, surveys, cited synthesis.

Business model: monetization, unit economics, pricing and packaging, lean canvas, North Star, network effects.

Write or review docs: README, docs site, tutorials, how-tos, API reference, runbooks, changelog, release notes.

Diagnose runtime failures: startup, ports, env, health and readiness gates, log bundles, systematic debugging.

Local app launchers: shortcuts, browser-first local UI, desktop wrappers, logs, ports, clean shutdown.