{"skill":{"slug":"aloudata-anomaly-detection","displayName":"Aloudata CAN SKILLS - anomaly-detection","summary":"对指标进行异常检测，判断当前数据是否偏离正常范围，输出结构化的异常检测报告。当用户希望检查指标是否异常、做健康巡检、或对一组指标做批量异常扫描时，必须使用此 Skill。 触发场景包括但不限于：用户提到\"异常检测\"\"有没有异常\"\"是否正常\"\"健康检查\"\"巡检\"\"波动是否正常\"\"数据是不是有问题\"\"帮我看看有没有问...","tags":{"latest":"1.0.0"},"stats":{"comments":0,"downloads":298,"installsAllTime":0,"installsCurrent":0,"stars":1,"versions":1},"createdAt":1774966925788,"updatedAt":1774967515349},"latestVersion":{"version":"1.0.0","createdAt":1774966925788,"changelog":"aloudata-anomaly-detection 1.0.0\n\n- Initial release of the anomaly-detection skill.\n- Provides structured reports to determine if metrics deviate from normal ranges.\n- Designed to support health checks and bulk anomaly scans for various business metrics.\n- All data queries are delegated to the metric-query skill; performs anomaly judgment locally after obtaining data.\n- Distinct workflow from related skills: focuses exclusively on abnormality detection (not raw querying or attribution).\n- Offers guidance on baseline building (statistical, trends, thresholds) and clear differentiation from adjacent skills.","license":"MIT-0"},"metadata":null,"owner":{"handle":"jackyujun","userId":"s172qm3yezbbjev7q8rsxxbd1s83he2f","displayName":"jackyujun","image":"https://avatars.githubusercontent.com/u/4506268?v=4"},"moderation":null}