How do I filter for specific errors in my trace logs?

You can use the command 'cat [filename].jsonl | jq 'select(.status == "error")'' to isolate and examine failed execution events.

What is the difference between quick-validator and trace-analyzer?

quick-validator uses the Haiku model for fast, cost-effective validation tasks like schema checks, while trace-analyzer uses reasoning-heavy models like Sonnet or Opus for root cause analysis.

Where are ctx-monitor trace files stored?

Trace files are stored as structured JSONL records in the .claude/ctx-monitor/traces/ directory within your project.

How can I check if monitoring is currently active?

Verify that "enabled": true is set in .claude/ctx-monitor/config.json and confirm your hooks are loaded using the /hooks command in the CLI.

What should I do if a hook is not firing as expected?

Check your hooks.json configuration for matcher errors, verify the patterns match the intended tools, and review trace logs to see if the session started correctly.

Trace Interpretation for ctx-monitor

Name: Trace Interpretation for ctx-monitor
Author: murillodutt

bymurillodutt

0•

分析与监控

Analyzes and deciphers Claude Code execution traces to debug agent behavior and identify performance bottlenecks.

This skill empowers developers using the ctx-monitor plugin to navigate and understand complex execution logs within the Claude Code CLI environment. It provides structured guidance for reading JSONL trace files, identifying various event types like tool usage and session boundaries, and diagnosing common failure patterns such as intermittent errors or hook misconfigurations. By leveraging specialized analysis agents like quick-validator and trace-analyzer, it streamlines the debugging process, offering actionable insights into why an agent stopped or how to optimize tool performance through historical data analysis.

主要功能

01Includes a comprehensive troubleshooting checklist for monitoring and hook verification.

02Deciphers structured JSONL trace formats and core event metadata.

03Provides optimized jq commands for advanced log filtering and data extraction.

04Identifies common failure patterns including cascade failures and performance degradation.

050 GitHub stars

06Explains the roles of specialized analysis agents for validation and deep reasoning.

使用场景

01Troubleshooting missing hooks or unexpected session terminations in the Claude Code CLI.

02Monitoring and auditing AI agent performance and duration across multiple sessions.

03Debugging failed tool executions and identifying root causes in agentic workflows.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add murillodutt/ctx-monitor trace-interpretation

For use in Claude.ai and ChatGPT

Download Skill