Skip to content
Kitbay
SkillOfficialTechnical Preview

elasticsearch-anomaly-detection-explainer

Apache-2.0Updated on Nov 10, 2026605GitHub stars

Elastic's skill for explaining the results of machine learning anomaly detection jobs in Elasticsearch.

What the skill teaches

Elastic's skill for explaining the results of machine learning anomaly detection jobs in Elasticsearch. When a user wants to know what drives a particular score, how the model learns, or why results look odd, the agent reads the job configuration through the ML REST API and the result documents through ordinary searches on the .ml-anomalies indices, then explains scoring, renormalisation and model behaviour and works through a troubleshooting checklist. It does not use ES|QL, so it also works on Serverless, and it needs the monitor_ml privilege. Creating or starting jobs is left to a sibling skill.

Capabilities

  • Explain why an anomaly score is high or low
  • Describe how anomaly detection models learn
  • Troubleshoot unexpected anomaly scores
  • Read job configuration and anomaly results