Menu
Runs integrated TSFM anomaly detection on a provided time-series dataset for explicit target columns.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Gathers live sensors, failure modes, and maps their relevance for diagnostics and root-cause insights.
Verifies IoT observability and maintenance history, then applies a basic safety clearance check.
Retrieves known failure modes for one explicit asset type or asset name.
Retrieves sensors and failure modes for one explicit asset and maps which sensors can monitor or detect those failures.
Lists available IoT sites and assets for a specified site.
Retrieves historical IoT observations for one explicit asset over an explicit time point or time window.
| id | tsfm_anomaly_detect_sensor |
| name | TSFM Anomaly Detect Sensor |
| version | 1.0.0 |
| description | Runs integrated TSFM anomaly detection on a provided time-series dataset for explicit target columns. |
| required_servers | ["tsfm"] |
| asset_types | ["chiller","equipment","time_series"] |
| keywords | ["tsfm","anomaly detection","tsad","anomalies","time series"] |
| default_enabled | true |
| inputs | {"dataset_path":{"type":"string","required":true},"timestamp_column":{"type":"string","required":true},"target_columns":{"type":"array","required":true},"model_checkpoint":{"type":"string","required":false},"false_alarm":{"type":"number","required":false}} |
| execution | {"type":"declarative"} |
Use this skill when the user provides or clearly references an existing time-series dataset and asks to detect anomalies for one or more target sensor columns with TSFM/TSAD.
Do not use this skill when the user asks for forecasting only, when IoT data must first be retrieved, or when the dataset path or target/timestamp columns are missing.
The final answer should state:
results_file path containing anomaly labels.{
"steps": [
{
"name": "anomaly_detection",
"server": "tsfm",
"tool": "run_integrated_tsad",
"arguments": {
"dataset_path": "$dataset_path",
"timestamp_column": "$timestamp_column",
"target_columns": "$target_columns",
"model_checkpoint": "$model_checkpoint",
"false_alarm": "$false_alarm"
}
}
]
}