Skill UI

This skill acts as an expert architect for scaffolding robust, production-ready applications using TypeScript. It generates complete, structured project templates for various types, including Next.js full-stack apps, React/Vite SPAs, and Node.js APIs. It incorporates modern tooling like pnpm, Vite, and Next.js features, ensuring type safety, proper dependency management, comprehensive testing setups, and adherence to current industry best practices. Use it to accelerate project setup and maintain scalable architecture.

This guide provides comprehensive metrics, instrumentation code (TypeScript), health check logic, and alerting rules for monitoring a SaaS platform like Juicebox. It covers critical metrics such as search latency (p95), data ingestion rates, API error rates, and quota usage, ensuring the maintenance of an optimal user experience and data integrity. Use this to establish robust operational monitoring dashboards and alerts.

This guide provides comprehensive strategies and TypeScript implementations for managing rate limits when interacting with the Juicebox AI API. It includes a token bucket rate limiter class, a robust retry mechanism for handling 429 errors and transient failures, and utilities for efficient batch processing of large datasets. Use this when building high-throughput applications that rely on Juicebox data analysis or enrichment endpoints.

Auto-activating helper that provides step-by-step guidance, best practices, and validation for generating production-ready k6 load test scripts, configs, and benchmarking workflows.

This comprehensive guide details how to safely upgrade LangChain versions (e.g., from 0.1.x to 0.3.x+). It covers critical breaking changes across both TypeScript and Python implementations, including updating import paths, converting legacy LLMChain to the modern LCEL pipe syntax, updating agent creation, and migrating memory management. Essential for maintaining and modernizing existing LLM applications.

A minimal example demonstrating how to implement basic tracing using the Langfuse SDK. It illustrates the hierarchy of traces, spans, and generations (LLM calls), providing code examples in TypeScript and Python for easy setup and integration testing.

This comprehensive guide details the installation and configuration of the Langfuse SDK for LLM observability. It covers setup instructions for both JavaScript/TypeScript and Python environments. Users will learn how to retrieve necessary API keys, configure environment variables, and initialize tracing using OpenTelemetry, enabling full monitoring and debugging of AI application calls.

Provides a deep dive and step-by-step scripts for complex Langfuse migrations. Covers comprehensive data export (traces, scores, prompts, datasets) and import processes. This guide is essential for platform transitions, moving between different Langfuse instances, or migrating from observability tools like LangSmith, supporting zero-downtime dual-write patterns.

A comprehensive checklist and guide for deploying Langfuse observability to a production environment. It covers best practices for verifying production configuration, implementing robust error handling, ensuring graceful shutdown, and executing pre-deployment validation scripts to guarantee system stability and reliability.

A comprehensive guide providing step-by-step instructions for upgrading Langfuse SDK versions (v3 to v5) and migrating between major API changes. It covers best practices for both TypeScript/JavaScript and Python implementations, focusing on adopting OpenTelemetry standards to ensure seamless LLM observability integration.

A comprehensive guide for deploying the application layer that interacts with Lindy AI agents. This includes setting up webhook receivers, managing callback handlers, and ensuring production readiness across multiple platforms like Docker, Vercel, and AWS. It covers prerequisites, multi-step deployment instructions, post-deployment verification scripts, and rollback plans for robust MLOps practices.

This skill provides comprehensive patterns for integrating external applications with the Lindy AI platform. It covers three primary channels: Webhook triggers for receiving inbound data, HTTP actions for making outbound API calls, and executing custom Python/JavaScript code (Run Code). Use this when building complex, multi-step automation workflows or connecting internal systems to Lindy agents.