SkillRegistry

WCAG 2.2 accessibility audit and remediation guidance for frontend code, UI reviews, keyboard/focus testing, ARIA usage, contrast, forms, and regression checks.

Local-first personal AI agent gateway architecture for Windows with messaging integrations, tool execution, and real-time streaming.

Anthropic Claude API expert - master Messages API requests, responses, streaming, tool use, prompt caching, and current Claude 4-family model practices.

REST and HTTP API design with OpenAPI 3.1, resource modeling, versioning, pagination, idempotency, errors, validation, auth boundaries, and production compatibility.

Expert in converting between AI provider APIs - Anthropic, Gemini, OpenRouter, Groq, Mistral, DeepSeek, and NVIDIA NIM.

Comprehensive authentication system for AI agent gateway with JWT tokens, session management, and multi-factor authentication.

Role-based access control (RBAC) and permission system for AI agent gateway with fine-grained resource authorization and policy enforcement.

Secure shell command execution for AI agent gateway with sandbox isolation, input validation, and resource limits using node-pty.

Server-side canvas rendering service for AI agent gateway using @napi-rs/canvas for image generation, drawing commands, and visual output.

CI/CD pipeline patterns for reliable builds, tests, artifacts, deployments, environment promotion, supply-chain security, and rollback-safe delivery.

Zod-validated configuration system for AI agent gateway with security modes, tool permissions, and platform-specific settings.

React-based control panel and dashboard for AI agent gateway with widget system, real-time monitoring, and configuration UI.

DeepSeek API expert - specialized AI models for coding, reasoning, and chat with competitive pricing.

Docker and containerization best practices for Node/TypeScript services, multi-stage builds, BuildKit cache/secrets, non-root runtime images, health checks, and image hardening.

File read/write operations for AI agent gateway with security controls, path validation, and resource limits.

Production guide for proxying Google Gemini API requests with the current Google GenAI SDK, multimodal content, streaming, safety settings, and gateway integration.

GitHub Actions workflow design, secure permissions, pnpm/Node CI, caching, matrix builds, OIDC deploys, artifacts, reusable workflows, and action hardening.

GraphQL schema, resolver, authorization, DataLoader, pagination, query cost/depth limiting, error masking, and production server guidance.

Groq API expert - ultra-low latency LLM inference with optimized models for speed and cost efficiency.

Secure input validation with Zod 4, allowlists, canonicalization, semantic validation, file upload checks, API boundary parsing, and safe error reporting.

JWT hardening for authentication systems, including RFC 8725 best practices, jose verification, algorithm allowlists, claims validation, token confusion prevention, revocation, and secure storage.

Build Model Context Protocol servers and clients with the TypeScript SDK, tools, resources, prompts, stdio and Streamable HTTP transports, Zod schemas, auth, and security review.

Message routing engine for AI agent gateway with channel+sender lookup, binding resolution, mention detection, and authorization checks.

Mistral AI API expert - access cutting-edge open-source models with European data residency and enterprise features.

Next.js App Router development with React Server Components, Server Actions, route handlers, caching, streaming, metadata, security, testing, and deployment guidance.

NVIDIA NIM expert - master NVIDIA NIM deployment, GPU-accelerated inference, custom model serving, and MCP integration for enterprise AI workloads.

OpenRouter API expert - access 400+ models with a unified API, dynamic model discovery, provider routing, fallbacks, and app attribution headers.

OWASP-aligned application security review and remediation for access control, cryptography, injection, insecure design, misconfiguration, vulnerable dependencies, auth, integrity, logging, SSRF, and secure coding controls.

Advanced pnpm workspace management, dependency resolution strategies, and performance optimization for monorepos.

PostgreSQL schema design, query optimization, indexing, migrations, transactions, EXPLAIN analysis, JSONB, full-text search, and operational safety.

Prompt engineering for reliable LLM outputs, including task framing, context selection, structured outputs, tool-use prompts, eval-driven iteration, prompt caching, and safe rationale handling.

Retrieval-augmented generation architecture with ingestion, chunking, embeddings, hybrid retrieval, reranking, citations, freshness, evaluation, and hallucination controls.

React application development with components, hooks, state design, effects, accessibility, performance, React Compiler awareness, testing, and production debugging.

React/Vite control UI for AI agent gateway with TypeScript, Tailwind CSS, and responsive design.

Redis caching, TTLs, key design, distributed locks, streams, pub/sub, rate limiting, invalidation, observability, and production reliability patterns.

Safe behavior-preserving refactoring guidance for code cleanup, decomposition, dependency boundaries, test-backed migrations, and code smell remediation.

Tailwind CSS v4 utility and design-system guidance, theme variables, responsive layouts, class composition, accessibility, component variants, and migration-aware styling.

Test-driven and test-first development with Vitest, Testing Library, behavior-focused tests, fixtures, regression tests, mocks at boundaries, and red-green-refactor workflow.

Approval system for dangerous tools with configurable modes (always-require-approval, owner-only, yolo) and sandbox execution.

Vue 3 application development with Composition API, script setup, TypeScript props/emits, Pinia, reactivity, component design, testing, and performance patterns.