Intelligent Log Analysis

Intelligent Log Analysis Skill

When to Use This Skill

Automatically invoke this Skill when:

• User mentions "check logs", "console errors", "debug logs"
• User reports JavaScript errors or runtime issues
• User wants to analyze a running web application
• User needs to monitor browser console output
• Keywords: "logs", "console", "errors", "debugging", "Chrome DevTools"

What This Skill Does

Intelligent Log Analyzer provides:

1. 90% Noise Filtering - Removes repetitive framework logs
2. Smart Deduplication - Collapses identical logs into summaries
3. Priority Detection - Highlights [CLAUDE-*] markers
4. Context Optimization - Saves 10K-25K tokens vs raw logs

Instructions

When this Skill is invoked:

Step 1: Determine the Mode

Ask yourself what level of analysis is needed:

• Quick Mode (10s) - Fast error detection
• Deep Mode (60s) - Comprehensive analysis
• Pure Mode - Maximum context optimization
• Claude-Priority - Focus on [CLAUDE-*] markers
• Errors-Only - Critical issues only

Step 2: Execute the Analysis

IMPORTANT: You MUST execute this bash command:

Quick Mode (Default):

~/claude-log-system/claude_console_v2.sh [target] --quick

Deep Mode:

~/claude-log-system/claude_console_v2.sh [target] --deep

Pure Mode:

~/claude-log-system/claude_console_v2.sh [target] --pure

Claude-Priority Mode:

~/claude-log-system/claude_console_v2.sh [target] --claude-priority

Errors-Only Mode:

~/claude-log-system/claude_console_v2.sh [target] --errors-only

Where [target] is usually localhost:PORT (e.g., localhost:3001)

Step 3: Analyze and Present

Organize the log output into:

## 📊 Log Analysis Results

**Target**: [Application URL]
**Analysis Mode**: [Quick/Deep/Pure/Errors-Only]
**Duration**: [Time analyzed]

### 🔴 Critical Issues
[Errors and exceptions that need immediate attention]

### 🟡 Warnings & Important Patterns
[Repeated warnings, performance issues, degraded functionality]

### 🟢 System Health Indicators
[Successful operations, normal metrics]

### 💡 Recommendations
1. [Specific action to fix issue #1]
2. [Specific action to fix issue #2]
...

### 📈 Log Statistics
- Total logs captured: [number]
- After filtering: [number] (XX% reduction)
- Critical errors: [number]
- Warnings: [number]

Step 4: Provide Context

Explain:

• What the errors mean
• Likely root causes
• Specific code locations if identifiable
• Next debugging steps

Examples

Example 1: Quick Error Check

User: "Check if there are any console errors on localhost:3001"

You execute:

~/claude-log-system/claude_console_v2.sh localhost:3001 --errors-only

You present: Filtered list of errors with explanations and fixes.

Example 2: Deep Analysis

User: "Analyze all logs from my app, it's running slow"

You execute:

~/claude-log-system/claude_console_v2.sh localhost:3001 --deep

You present: Comprehensive analysis including performance warnings, network issues, and optimization suggestions.

Intelligent Filtering Features

The system automatically:

• ✅ Collapses [repeated 10x] identical logs
• ✅ Aggregates progress bars and loading streams
• ✅ Prioritizes developer-marked [CLAUDE-*] logs
• ✅ Filters React/Vue/Angular framework noise
• ✅ Groups related error sequences

Claude-Specific Log Markers

Developers can add markers for your attention:

• [CLAUDE-CRITICAL] - Immediate attention required
• [CLAUDE-FOCUS] - Important for analysis
• [CLAUDE-INFO] - Useful context
• [CLAUDE-DEBUG] - Debug information

Use --claude-priority mode to focus on these.

Performance Characteristics

• Log Reduction: 80-95% fewer logs
• Token Savings: 10K-25K tokens saved
• Analysis Time: 10-60 seconds
• Memory Usage: Optimized with auto-cleanup

Prerequisites

Before executing:

1. Chrome browser must be running
2. Target web page must be open
3. Script automatically sets debugging port (9222)

Fallback Strategy

If the command fails:

1. Verify Chrome is running with the target page
2. Try extending timeout: --timeout 60
3. Check if page URL is correct
4. Manual Chrome DevTools as last resort

Important Notes

• Always execute the bash command, don't simulate results
• Analyze patterns, not just list errors
• Provide actionable fixes, not just descriptions
• Group related issues for clarity
• Suggest next steps for debugging

Common Use Cases

• Frontend debugging (React, Vue, Angular)
• API error investigation
• Performance bottleneck identification
• Memory leak detection
• Network request analysis
• Runtime error tracking