arize-trace

Export and inspect Arize trace and span data using the ax CLI.

Updated May 5, 2026
One-click install
npx skills add https://github.com/ethan-rio/skills --skill arize-trace-ethan-rio
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: arize-trace
Source: https://github.com/ethan-rio/skills/tree/main/skills/arize/arize-trace
Command: npx skills add https://github.com/ethan-rio/skills --skill arize-trace-ethan-rio

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

When working with LLM applications, it is difficult to get visibility into runtime execution, especially when debugging errors, unexpected outputs, or behavior regressions. This skill solves that by enabling easy download, export, and inspection of Arize trace and span data to understand exactly how your LLM app is performing.

Core Features & Use Cases

  • Targeted Export: Pull exact trace or span data by ID, session ID, or custom SQL-like filters to get only the data you need.
  • Error Debugging: Investigate root causes of LLM app failures by inspecting span attributes, LLM inputs/outputs, tool calls, and error messages.
  • Bulk Analysis: Export large datasets of traces and spans for offline analysis of behavior regressions or performance issues. Use Case: If your LLM chatbot is returning incorrect answers for specific user queries, use this skill to export the relevant trace, inspect the LLM input messages and tool call parameters to identify where the logic is breaking down.

Quick Start

Use the arize-trace skill to export all error spans from your Arize project for the last 24 hours to debug failing LLM requests.

Frequently Asked Questions about arize-trace

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I export Arize traces to debug LLM application errors?

Investigate LLM execution path failures by exporting Arize trace data to inspect span attributes, LLM inputs and outputs, tool call parameters, and error messages. This allows you to review the exact execution path that caused the failure.

Can I bulk export Arize spans for offline analysis of behavior regressions?

Filter Arize trace exports by applying custom SQL-like filters, or by specifying exact trace IDs, span IDs, and session IDs. This targeted export capability ensures you only download the specific LLM runtime data needed.

Are there guardrails for handling untrusted span content during trace inspection?

Guardrails are applied during trace inspection to handle untrusted span content safely. This ensures that downloading and analyzing potentially malicious or malformed LLM execution data does not compromise your debugging environment.

What is the best way to inspect LLM tool calls and inputs causing incorrect chatbot responses?

The best way to inspect incorrect LLM chatbot responses is to export the relevant Arize trace and inspect the LLM input messages and tool call parameters. This helps identify exactly where the application logic is breaking down.