Advanced AI Prompt for Automated AWS Cost Anomaly Remediation

Act as an AWS FinOps Automation Engineer. Design a ruthless, fully autonomous cost-control architecture leveraging AWS Cost Anomaly Detection. When an anomaly breaches the absolute [SPEND_THRESHOLD] (e.g., $100 spike in 1 hour) specifically targeting [TARGET_SERVICES] (e.g., SageMaker or EC2), it must trigger an SNS topic. Write a highly defensive Python Lambda function subscribed to this topic that parses the JSON payload. If the anomaly is traced to an un-tagged, rogue developer instance, the Lambda must execute an automated 'StopInstances' or 'DeleteEndpoint' API call to instantly halt the billing bleed, followed by pushing a forensic summary payload directly to the [NOTIFICATION_WEBHOOK].