troubleshoot

Analyze debug logs to identify root causes of unexpected agent behavior.

Updated Apr 1, 2026
One-click install
npx skills add https://github.com/quickxz35-ux/bookmap-crypto-addon --skill troubleshoot-quickxz35-ux
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: troubleshoot
Source: https://github.com/quickxz35-ux/bookmap-crypto-addon/tree/main/skills/troubleshoot
Command: npx skills add https://github.com/quickxz35-ux/bookmap-crypto-addon --skill troubleshoot-quickxz35-ux

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Investigate and explain unexpected chat agent behavior by analyzing debug logs and session traces, enabling quick root-cause understanding.

Core Features & Use Cases

  • Root-cause analysis: Correlate errors, latency, and tool usage to identify the source of misbehavior.
  • Trace-driven debugging: Follow main and child sessions via runSubagent, system prompts, and tools logs to reconstruct events.
  • Remediation guidance: Provide concrete steps to fix configuration, routing, or model behavior issues.

Quick Start

Run an initial diagnostic by selecting a session and requesting a concise root-cause report with recommended fixes.

Frequently Asked Questions about troubleshoot

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

FAQPage Schema
How do I find the root cause of unexpected chat agent behavior from debug logs?▼

To find the root cause of unexpected agent behavior from debug logs, you analyze the main.jsonl session file to trace tool invocations, correlate latency or load failures, and output a concise root-cause analysis with actionable remediation steps.

What is the best way to trace latency and tool invocation issues in Copilot session logs?▼

The best way to trace latency and tool invocation issues in Copilot session logs is by following child_session_ref pointers from the main session to related logs, reconstructing events to pinpoint load failures or tool errors.

How does session tracing work for diagnosing agent misbehavior in JSONL files?▼

Session tracing for diagnosing agent misbehavior works by reading the main.jsonl log, following child_session_ref pointers to child sessions, and correlating system prompts and tools logs to reconstruct the full event timeline.

Can I diagnose tool load failures using debug logs from a specific session ID?▼

Yes, you can diagnose tool load failures using debug logs from a specific session ID by analyzing the corresponding debug-logs/<sessionId>/main.jsonl and tools_*.json files to trace invocation errors and output remediation steps.

Do I need specific file formats to perform root-cause analysis on chat agent sessions?▼

Yes, you need specific file formats to perform root-cause analysis on chat agent sessions, specifically Copilot-style debug logs structured as debug-logs/<sessionId>/main.jsonl alongside related tools_*.json files.