Skill UI

SGLang is a high-performance serving framework for LLMs and VLMs that uses RadixAttention prefix caching to deliver structured outputs (JSON/regex/grammar) and agentic workflows with function calls, providing up to 5× faster inference than vLLM across multi-GPU production deployments.

This guide details how to configure the Speak language learning platform across multiple environments (dev, staging, production). It is essential for implementing robust SaaS deployments, ensuring that separate API keys, mock modes, and environment variables are correctly managed for seamless development and operational management.

A comprehensive guide to securely implementing Supabase across diverse modern application stacks. This skill covers critical patterns for Next.js Server Components (SSR), Single Page Applications (SPA), React Native mobile apps, Edge Functions (Serverless), and complex multi-tenant deployments using RLS or schema isolation. It guides developers on correctly managing client types (anon vs. service_role) and handling authentication flows in various environments.

This comprehensive skill guides the full lifecycle management of Supabase projects in a production environment. It covers essential DevOps practices, including running database migrations, deploying Edge Functions, managing secrets, implementing zero-downtime blue/green rollouts, and executing robust rollback procedures and post-deployment health checks. Ideal for ensuring stable and reliable production deployments.

Optimizes large language model inference on NVIDIA GPUs with TensorRT, delivering 10-100× faster throughput, quantized precision (FP8/INT4), multi-GPU scaling, and serving-ready tooling for production deployments.

A comprehensive, multi-phase workflow designed for Infrastructure as Code (IaC) using Terraform. It guides users through provisioning cloud resources, creating reusable modules, managing complex state, handling multi-environment deployments, and integrating robust CI/CD pipelines and security practices. Ideal for establishing scalable, repeatable, and auditable cloud infrastructure.

Tracks and runs regression suites, automating execution of critical tests and analyzing outcomes so Claude highlights failures, flaky checks, and stability impacts before deployments.

Automatically checks application metrics such as load times, bundle sizes, and API responses against predefined performance budgets, surfacing regressions and preventing slippage before deployments—ideal for CI/CD workflows.

This expert skill validates the deployment readiness of Vertex AI Agent Engines. It executes a weighted audit across five critical dimensions: security, monitoring, performance, compliance, and best practices. The output provides a comprehensive 0-100% composite score, a detailed pass/fail report for each check, and a prioritized, actionable remediation plan, ensuring high-quality, production-grade deployments.

This guide details how to implement environment isolation (dev, staging, prod) for GPU cloud deployments on Vast.ai. It uses environment variables and configuration classes to enforce separate API keys, budget caps, instance limits, and GPU whitelists for each deployment tier, ensuring efficient resource management and cost control.

A comprehensive checklist for deploying large-scale, production-ready GPU workloads on Vast.ai. It covers critical lifecycle stages including account authentication, instance validation, data encryption, persistent checkpointing strategies, fault tolerance for spot instances, robust monitoring (utilization & cost), and strict cost controls, ensuring reliable and stable ML/AI deployments.

This comprehensive guide helps developers select and implement optimal Vercel architecture blueprints based on project scale and complexity. It covers five validated patterns, ranging from simple static JAMstack sites and robust full-stack SaaS applications, to advanced multi-zone micro-frontend enterprise deployments. Use this resource when planning new deployments, optimizing scaling strategies, or structuring complex monorepos.