Overview

Live Log Insight – An intelligent agent system for automated log analysis, anomaly detection, and actionable insights generation.

:::info Looking for the CLI version? This version uses the Python pipeline directly (non-CLI run). If you prefer using the interactive CLI with animations and prompts, switch to the CLI version. :::

Live Log Insight

An intelligent agent skills system that analyzes logs, detects anomalies, and produces actionable insights — using composable skills, not monolithic prompts.

Powered By

What is Live Log Insight?

Live Log Insight is a modular Agent Skills system designed for automated log analysis. It orchestrates specialized skills to transform raw logs into meaningful insights for operators and developers. Unlike traditional monitoring tools, it uses an AI-powered pipeline that can understand context, detect anomalies, and suggest actions.

Key Features

• 🔍 Automated Discovery: Automatically identifies log sources across environments
• 📊 Intelligent Parsing: Normalizes diverse log formats into structured events
• 📈 Pattern Recognition: Aggregates logs to identify trends and metrics
• 🚨 Anomaly Detection: Detects spikes, new errors, and unusual behavior
• 💡 Root Cause Analysis: Generates plausible hypotheses for detected issues
• 📝 Actionable Summaries: Produces clear, human-readable reports
• ✅ Smart Recommendations: Suggests concrete next steps for operators

Architecture

The system follows a sequential pipeline workflow:

Each stage is implemented as a modular skill, making the system flexible and extensible.

When to Use

Activate Live Log Insight when:

• User requests log analysis or system health insights
• Daily/weekly operational summaries are needed
• Investigating incidents, spikes, or unusual behavior
• Monitoring production or staging systems
• Troubleshooting application errors

Quick Example

# Run a complete log analysis pipeline
uv run main.py --time-range 24h --environment production

This will:

1. Discover available log sources
2. Fetch logs from the past 24 hours
3. Parse and normalize log entries
4. Aggregate metrics and patterns
5. Detect anomalies
6. Generate hypotheses (optional)
7. Create a summary report
8. Recommend next steps

Supported Log Sources

• Filesystem: /var/log and custom paths
• Elasticsearch: Index-based log retrieval
• Custom sources: Extensible via configuration

Output Formats

Output	Description	Location
summary.md	Human-readable analysis report	output/summary.md
anomalies.json	Detected issues with evidence	output/anomalies.json
recommendations.json	Prioritized action items	output/recommendations.json
metrics.json	Aggregated statistics	output/metrics.json
hypotheses.json	Root cause hypotheses	output/hypotheses.json

Get Started

Quickstart

Get up and running in 5 minutes.

CLI Version

Switch to CLI version for interactive mode.

Core Concepts

Learn about the system architecture.

Skills Reference

Explore all available skills.

Examples

See real-world usage examples.

Configuration

Configure log sources and thresholds.

API Reference

Integrate with your own tools.

Project Structure

live-log-insight/
├── .agents/
│   ├── AGENTS.md              # Agent orchestration instructions
│   └── skills/
│       ├── livelogs_insights/     # Orchestration skill
│       ├── logsource_discovery/   # Log source discovery
│       ├── fetch_logs/            # Log retrieval
│       ├── parse_logs/            # Log normalization
│       ├── aggregate_logs/        # Metrics computation
│       ├── detect_anomalies/      # Anomaly detection
│       ├── high_hypothesis/       # Root cause analysis
│       ├── generate_summary/      # Report generation
│       └── recommend_actions/     # Action recommendations
├── src/agentX/                  # Python source code
│   ├── config/                  # Configuration modules
│   ├── pipeline/                # Pipeline orchestration
│   └── shared/                 # Shared utilities
├── config/                      # Configuration files
├── metadata/                    # Sample data for testing
├── output/                      # Generated outputs
└── docs/                        # Documentation

License

Apache-2.0