Live Logs Insights
The main orchestration skill that coordinates the complete log analysis pipeline.Overview
This skill is the entry point for comprehensive log analysis. It activates and coordinates all other skills in sequence to transform raw logs into actionable insights.When to Use
Activate this skill when:- User requests log analysis or system health insights
- Daily/weekly operational summaries are needed
- Investigating incidents or unusual behavior
- Monitoring production or staging systems
Directory Structure
Instructions
-
Read configuration: Load
config/log_sources.yamlandconfig/anomaly_thresholds.yaml - Execute pipeline: Run the orchestration script with input parameters
-
Pipeline stages: The script coordinates:
logsource_discovery- Identifies log sourcesfetch_logs- Retrieves raw logsparse_logs- Normalizes log formatsaggregate_logs- Computes metricsdetect_anomalies- Identifies issueshigh_hypothesis- Generates explanations (optional)generate_summary- Creates reportrecommend_actions- Suggests next steps
-
Review outputs: Results are written to
output/:summary.md- Human-readable reportanomalies.json- Detected issuesrecommendations.json- Action itemsmetrics.json- Aggregated statisticshypotheses.json- Root cause hypotheses
-
Present findings: Include:
- Executive summary of system health
- Key anomalies with confidence levels
- Root cause hypotheses (when available)
- Prioritized recommendations
- Supporting metrics and evidence
-
Handle errors: If pipeline fails:
- Log the error with context
- Return partial results if available
- Suggest manual investigation steps
- Never hallucinate missing data
