Debug your AI agents directly from the terminal. This tool fetches execution traces from LangSmith Studio so you can see what went wrong. It shows you the tools your agent called, how long each step took, and where errors happened. You can check recent activity or investigate a specific trace in detail.
If your agent fails, you can find the exact error message and see the full execution flow. This helps you fix problems faster. The tool also exports debug sessions for sharing with your team. It works with LangChain and LangGraph agents.
You need to install the langsmith-fetch CLI and set up your API key and project name. Then you can run simple commands to pull traces and get clear reports. This makes debugging your AI agents much easier.
Global
mkdir -p ~/.claude/skills/langsmith-fetchProject
mkdir -p .claude/skills/langsmith-fetchSource Repository
Azure Diagnosticsmicrosoft/azure-skills
Debug Azure production issues with systematic diagnostics and AI-powered tools
Azure Messagingmicrosoft/azure-skills
Troubleshoot and fix Azure messaging SDK errors for Event Hubs and Service Bus
Use My Browserxixu-me/skills
Work inside your live browser session for real page inspection and debugging
Diagnosemattpocock/skills
A step-by-step debug loop for finding and fixing hard bugs
Diagnosing Bugsmattpocock/skills
Build a tight feedback loop to find any bug fast
Systematic Debuggingobra/superpowers
Find the root cause of every bug before attempting any fix to save time
Safe Debuglllllllama/rigorpilot-skills
Debug deep learning failures safely with conservative diagnosis and minimal patches
Firebase Crashlyticsfirebase/agent-skills
Set up crash reporting for your Android or iOS app with ease