Execute a training action for anomaly detection models in AWS IoT SiteWise. This specialized function handles training actions on anomaly detection computation models. It supports five optional configurations: target sampling rate, labeled data for supervised learning, model evaluation for point...
High parameter count (21 properties)
Part of the AWS IoT SiteWise MCP Server MCP server. Enforce policies on this tool with Intercept, the open-source MCP proxy.
AI agents invoke execute_training_action to trigger processes or run actions in AWS IoT SiteWise MCP Server. Execute operations can have side effects beyond the immediate call -- triggering builds, sending notifications, or starting workflows. Rate limits and argument validation are essential to prevent runaway execution.
execute_training_action can trigger processes with real-world consequences. An uncontrolled agent might start dozens of builds, send mass notifications, or kick off expensive compute jobs. Intercept enforces rate limits and validates arguments to keep execution within safe bounds.
Execute tools trigger processes. Rate-limit and validate arguments to prevent unintended side effects.
tools:
execute_training_action:
rules:
- action: allow
rate_limit:
max: 10
window: 60
validate:
required_args: trueSee the full AWS IoT SiteWise MCP Server policy for all 72 tools.
Agents calling execute-class tools like execute_training_action have been implicated in these attack patterns. Read the full case and prevention policy for each:
Other tools in the Execute risk category across the catalogue. The same policy patterns (rate-limit, validate) apply to each.
execute_training_action is one of the high-risk operations in AWS IoT SiteWise MCP Server. For the full severity-focused view — only the high-risk tools with their recommended policies — see the breakdown for this server, or browse all high-risk tools across every MCP server.
Execute a training action for anomaly detection models in AWS IoT SiteWise. This specialized function handles training actions on anomaly detection computation models. It supports five optional configurations: target sampling rate, labeled data for supervised learning, model evaluation for pointwise diagnostics, model metrics for training insights, and retraining scheduler configuration for automated retraining. Note: Get the training_action_definition_id from describe_computation_model response's actionDefinitions array (actionType: "AWS/ANOMALY_DETECTION_TRAINING"). Args: training_action_definition_id: ID of the training action definition (required) training_mode: Training mode - TRAIN_MODEL, START_RETRAINING_SCHEDULER, or STOP_RETRAINING_SCHEDULER (required) target_resource: Resource containing computationModelId (required, computationModelId must be in UUID format) region: AWS region (default: us-east-1) # TRAIN_MODEL mode parameters: export_data_start_time: Unix epoch timestamp for training data start (REQUIRED for TRAIN_MODEL) export_data_end_time: Unix epoch timestamp for training data end (REQUIRED for TRAIN_MODEL) target_sampling_rate: Sampling rate (PT1S to PT1H) - higher rates offer detail but increase cost (optional for TRAIN_MODEL) label_bucket_name: S3 bucket for labeled training data CSV (optional for TRAIN_MODEL, requires label_s3_prefix) label_s3_prefix: S3 prefix for labeled training data CSV (optional for TRAIN_MODEL, requires label_bucket_name) evaluation_start_time: Unix epoch timestamp for evaluation data start (optional for TRAIN_MODEL, requires all evaluation params) evaluation_end_time: Unix epoch timestamp for evaluation data end (optional for TRAIN_MODEL, requires all evaluation params) evaluation_bucket_name: S3 bucket for evaluation results (optional for TRAIN_MODEL, requires all evaluation params) evaluation_s3_prefix: S3 prefix for evaluation results (optional for TRAIN_MODEL, requires all evaluation params) metrics_bucket_name: S3 bucket for comprehensive training metrics (optional for TRAIN_MODEL, requires metrics_s3_prefix) metrics_s3_prefix: S3 prefix for training metrics JSON (optional for TRAIN_MODEL, requires metrics_bucket_name) # START_RETRAINING_SCHEDULER mode parameters: lookback_window: Historical data window for retraining (P180D, P360D, P540D, P720D) (REQUIRED for START_RETRAINING_SCHEDULER) retraining_frequency: How often to retrain (P30D to P1Y) (REQUIRED for START_RETRAINING_SCHEDULER) promotion: Model promotion mode (SERVICE_MANAGED, CUSTOMER_MANAGED) (optional for START_RETRAINING_SCHEDULER, defaults to SERVICE_MANAGED) retraining_start_date: Unix epoch timestamp for when retraining should start (optional for START_RETRAINING_SCHEDULER) # STOP_RETRAINING_SCHEDULER mode parameters: # No additional parameters required or accepted for STOP_RETRAINING_SCHEDULER mode # Common optional parameters: client_token: Optional unique identifier for idempotent requests resolve_to: Optional resource containing assetId Returns: Dictionary containing action execution response with trainingPayload for reference. Example: # Basic training result = execute_training_action( training_action_definition_id='12345678-1234-1234-1234-123456789012', training_mode='TRAIN_MODEL', target_resource={'computationModelId': '87654321-4321-4321-4321-210987654321'}, export_data_start_time=1717225200, export_data_end_time=1722789360 ) # Start retraining scheduler result = execute_training_action( training_action_definition_id='12345678-1234-1234-1234-123456789012', training_mode='START_RETRAINING_SCHEDULER', target_resource={'computationModelId': '87654321-4321-4321-4321-210987654321'}, lookback_window='P360D', retraining_frequency='P30D', promotion='SERVICE_MANAGED', retraining_start_date=1730332800 ) # Stop retraining scheduler (no additional parameters needed) result = execute_training_action( training_action_definition_id='12345678-1234-1234-1234-123456789012', training_mode='STOP_RETRAINING_SCHEDULER', target_resource={'computationModelId': '87654321-4321-4321-4321-210987654321'} ) # Complete training with all configurations result = execute_training_action( training_action_definition_id='12345678-1234-1234-1234-123456789012', training_mode='TRAIN_MODEL', target_resource={'computationModelId': '87654321-4321-4321-4321-210987654321'}, export_data_start_time=1717225200, export_data_end_time=1722789360, target_sampling_rate='PT5M', label_bucket_name='anomaly-detection-data-bucket', label_s3_prefix='Labels/model-id/Labels.csv', evaluation_start_time=1719817200, evaluation_end_time=1720422000, evaluation_bucket_name='anomaly-detection-eval-bucket', evaluation_s3_prefix='Evaluations/model-id/', metrics_bucket_name='anomaly-detection-metrics-bucket', metrics_s3_prefix='ModelMetrics/model-id/', resolve_to={'assetId': '11111111-1111-1111-1111-111111111111'} ). It is categorised as a Execute tool in the AWS IoT SiteWise MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Add a rule in your Intercept YAML policy under the tools section for execute_training_action. You can allow, deny, rate-limit, or validate arguments. Then run Intercept as a proxy in front of the AWS IoT SiteWise MCP Server MCP server.
execute_training_action is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the execute_training_action rule in your Intercept policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the Intercept policy for execute_training_action. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
execute_training_action is provided by the AWS IoT SiteWise MCP Server MCP server (awslabs.aws-iot-sitewise-mcp-server). Intercept sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic policy on every MCP tool call. Per-identity grants. Full audit log.