Skill UI

Guide to instrument Azure-hosted ASP.NET Core or Node.js web apps with telemetry reporting to Azure App Insights for improved observability and health tracking.

A comprehensive skill for creating, managing, and querying datasets within the Arize platform. Use it for dataset CRUD operations, appending examples, exporting data, and generating test or evaluation examples for machine learning models.

This skill facilitates comprehensive LLM-as-judge evaluation within the Arize platform. Users can create, update, and manage evaluators (defining prompt templates, classification criteria, and models). It allows running evaluations against live spans, historical experiments, or continuous data streams. It supports advanced features like column mapping and fine-grained data granularity (span, trace, session) for rigorous model performance monitoring.

This skill guides the structured, two-phase process of instrumenting applications with Arize AX for comprehensive observability. It covers analyzing various codebases (Python, JS, Java), detecting dependencies, and implementing robust tracing. Special emphasis is placed on LLM applications, ensuring manual spans are added for tool and function calling to provide complete system visibility.

This skill guides the process of optimizing, improving, or debugging Large Language Model (LLM) prompts. It teaches users how to extract structured prompts and performance signals from production trace data, leveraging attributes like `llm.input_messages` and evaluation annotations. Users can perform data-driven tuning and quality improvement using the `ax` CLI.

This skill guides users on how to download, export, and inspect Arize traces and spans using the `ax` CLI. It covers exporting data by specific IDs (Trace ID, Span ID, Session ID), performing root-cause investigations, and handling complex export scenarios, including best practices for filtering and security considerations.

Aspire is a code-first polyglot toolchain that orchestrates containers, executables, and cloud resources through an AppHost project, guiding creation, debugging, configuration, deployment, and troubleshooting for distributed applications across languages.

Analyzes IaC files and live Azure resources to validate SKUs, usage, and costs, then generates GitHub issues (one per recommendation plus an EPIC) so teams can track implementation of prioritized savings.

Design, analyze, and refine Azure infrastructure through natural language prompts, generate interactive architecture diagrams locally with bundled scripts, and produce deployable Bicep templates to automate infrastructure-as-code workflows.

Conducts a multistep preflight validation for Azure Bicep deployments, covering syntax checks, what-if previews, permission verification, and report generation so teams can confidently run azd or az deployments with known outcomes.

Fetches real-time Azure retail prices from the public Azure Retail Prices API and guides credit estimations for Copilot Studio agents, covering compute, storage, networking, databases, AI, and Copilot billing scenarios.

This advanced workflow automates the process of analyzing the operational health of specific Azure resources. It ingests and analyzes raw logs, telemetry data, and performance metrics from diverse Azure services (e.g., VMs, Web Apps, Databases). The system performs comprehensive root cause analysis, classifies issues (Critical, High, Medium), and generates actionable, step-by-step remediation plans, significantly reducing manual troubleshooting time.