skills.rest
china-qijizhifeng
china-qijizhifengOFFICIAL
Explore all skills by this author →
GitHub

agent-debugger-cli

Configure AI models, download Langfuse traces, and run quality checks.

812|90|Updated Apr 23, 2026
Search Tags:#AI#quality check#Langfuse#agent debugging#execution trace
One-click install
npx skills add https://github.com/china-qijizhifeng/agentic-harness-engineering --skill agent-debugger-cli-china-qijizhifeng
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-debugger-cli
Source: https://github.com/china-qijizhifeng/agentic-harness-engineering/tree/main/agents/evolve_agent/skills/agent-debugger-cli
Command: npx skills add https://github.com/china-qijizhifeng/agentic-harness-engineering --skill agent-debugger-cli-china-qijizhifeng

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires nexau, langfuse, pyyaml, python-dotenv, requests, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill unit provides a CLI tool for configuring, downloading, analyzing, and debugging agent execution traces. It simplifies the process of understanding and optimizing agent performance.

Core Features & Use Cases

  • LLM Configuration: Configure the AI model and API key for the agent debugger CLI.
  • Trace Downloading: Download and clean Langfuse traces for analysis.
  • QA Questions: Ask questions about traces and receive detailed responses.
  • Quality Check: Run quality checks on traces to identify issues.
  • Use Case: Suppose you have a complex agent that has generated a large number of execution traces. Use this Skill unit to download the traces, analyze them, and identify any issues that may be affecting the agent's performance.

Quick Start

Install the skill:

python -m pip install ./_source

Configure the LLM:

adb config '{"llm":{"model":"gpt-4.1","base_url":"https://api.openai.com/v1","api_key":"<your-key>"}}'
Download a Langfuse trace:
```bash
adb download --type langfuse --ak <public-key> --sk <secret-key> https://<langfuse-host>/project/<projectId>/traces/<traceId>
Ask a question about a trace:
```bash
adb ask -t ~/.adb/traces/<projectId>/<traceId>.cleaned.json -q "What is the main issue in this trace?"

Frequently Asked Questions about agent-debugger-cli

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

FAQPage Schema
How do I download and analyze Langfuse traces to debug agent performance?▼

To debug agent performance, you can use an agent debugger CLI to download Langfuse traces via public and secret keys, clean the data, and run quality checks to identify execution issues.

What is the best way to run quality checks on AI agent execution traces?▼

Running quality checks on AI agent execution traces involves using a CLI tool to analyze downloaded trace files, allowing you to ask specific questions and identify performance issues.

Do I need Python and local shell access to configure an AI agent debugger?▼

Yes, configuring an AI agent debugger requires Python, pip, and local shell access to install the CLI package, set the LLM API key, and execute trace download or analysis commands.

Can I use a custom AI model for analyzing agent execution traces?▼

Yes, you can configure a custom AI model for trace analysis by passing the model name, base URL, and API key directly through the debugger CLI configuration command.

What limitations exist when downloading Langfuse traces for agent debugging?▼

Downloading Langfuse traces for agent debugging requires network access and valid authentication keys; without network connectivity or proper credentials, trace retrieval and analysis cannot proceed.

How do I ask questions about a specific agent execution trace after downloading it?▼

After downloading a cleaned agent execution trace file, you can ask specific questions by passing the trace file path and your query to the CLI, receiving detailed responses about trace issues.

Related Skills

Explore More Software Engineering Skills→
divo12divo12
1

agent-debugger-cli

Automate agent trace configuration, QA, and quality analysis via CLI.

Community
Advanced
langware-labslangware-labs
5

flow

Execute markdown-defined workflows and emit structured trace reports.

Official
Intermediate
aurumorincaurumorinc

langfuse

Configure Langfuse trace trees, score metrics, and prompt versioning for AI agent workflows.

Official
Advanced
githubgithub
37.6k

agentic-eval

Automate iterative evaluation and refinement of AI agent outputs with JSON feedback.

Official
Intermediate