skills.rest
kingmacth
kingmacthCOMMUNITY
Explore all skills by this author →
GitHub

agent-tracing

Record agent execution snapshots to disk and inspect traces via CLI.

Updated Dec 12, 2024
Search Tags:#LLM#CLI#debug#runtime#trace#snapshots#agent-tracing
One-click install
npx skills add https://github.com/kingmacth/lobe-chat --skill agent-tracing-kingmacth
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-tracing
Source: https://github.com/kingmacth/lobe-chat/tree/main/.agents/skills/agent-tracing
Command: npx skills add https://github.com/kingmacth/lobe-chat --skill agent-tracing-kingmacth

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Agent tracing helps developers debug and analyze agent executions by recording execution snapshots to disk and exposing them for review.

Core Features & Use Cases

  • Capture and store step-by-step execution snapshots as the agent runs.
  • Inspect LLM calls, context data, and tool interactions to diagnose failures and performance issues.
  • Use the CLI to list traces, inspect steps, and view detailed messages for each run.

Quick Start

Use the agent-tracing CLI from the project root to start recording agent execution traces and then inspect them with the trace and inspect commands.

Frequently Asked Questions about agent-tracing

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

FAQPage Schema
How do I trace and inspect LLM calls during agent execution?▼

To trace LLM calls, you use a runtime that records execution snapshots to disk and a CLI to review them. This captures step-by-step interactions and context data for diagnosing failures.

What is the best way to debug agent steps and inspect context data locally?▼

Debugging agent steps locally requires recording partial and final execution snapshots on disk. You can then use CLI commands to list traces, inspect individual steps, and view detailed messages.

How do I view execution snapshots recorded from my agent runs?▼

You view execution snapshots by using the CLI to list recorded traces and inspect specific steps. This exposes detailed messages and context data captured during the agent execution.

Does agent-tracing require any specific runtime components to capture execution traces?▼

Yes, capturing execution traces requires the agent-tracing runtime components. These components record partial and final snapshots, store them on disk, and provide the CLI to view trace details.

Why do I need to record execution snapshots for my LLM workflows?▼

Recording execution snapshots is necessary to audit agent steps and diagnose performance issues across runs. It exposes LLM calls and tool interactions that are otherwise invisible during runtime.

Can I use a CLI to audit agent steps and diagnose failures across multiple runs?▼

Yes, you can use the CLI to audit agent steps and diagnose failures across runs. It lists stored traces from disk and allows you to inspect detailed messages for each recorded execution.

Related Skills

Explore More Software Engineering Skills→
fritzprixfritzprix
8

system-setup

Install and verify MCP runtime dependencies across Windows, Linux, and macOS.

Community
Intermediate
redfox-dataredfox-data
326

xiaohongshu-account-analyzer

Analyze Xiaohongshu accounts to generate structured diagnostic scores and HTML reports.

Community
Advanced
Nick777-PixelNick777-Pixel

agent-tracing

Inspect agent execution snapshots and trace LLM calls via CLI.

Community
Advanced
jeremylongshorejeremylongshore
2.6k

monitoring-database-health

Monitor PostgreSQL and MySQL database health with real-time metrics and alerts.

Community
Advanced