{"id":20679,"plugin_id":"plugins_6aa698dc64588191b48664000f8522de","kind":"skill","collection_source":null,"comparison_source":null,"observed_at":"2026-09-30T23:16:26.035Z","digest":"b2c53b2b47256ee542f6deff6886c41a725ef98d2b38f682bffcc381e9e0fa1d","against":null,"payload":{"name":"ai-anomaly-triage","description":"Correlate query stats, telemetry, policy decisions, and replay memory into deterministic incident hypotheses and next-best actions.","included_files":[],"skill_md_contents":"---\nname: ai-anomaly-triage\ndescription: Correlate query stats, telemetry, policy decisions, and replay memory into deterministic incident hypotheses and next-best actions.\n---\n\n# AI Anomaly Triage\n\nUse this skill when an operator asks why database behavior changed, whether a\ndeployment caused a regression, or what to investigate first during an incident.\n\nRun:\n\n```bash\nnode runtime/runTool.js ai_anomaly_triage '{\"engine\":\"postgres\",\"database\":\"analytics\",\"deploymentId\":\"deploy-42\",\"incidentWindowMinutes\":45}'\n```\n\nThe tool is offline and deterministic. It combines workload stats,\ntelemetry-correlation output, policy-memory signals, and recent replay context.\nIt returns an anomaly score, root-cause hypotheses, next-best actions, and\nexplainability fields listing the exact signal families used.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}