AWS AI Updates: July 28, 2026

1. AWS Security Hub Adds MCP App For AI-Assisted Exposure Investigation (Preview)

AWS. AWS launched the Security Hub MCP App in preview, a local Model Context Protocol server that connects Security Hub exposure findings to Claude Desktop for natural language investigation. The app exposes read-only tools that let an AI assistant retrieve top findings, analyze attack paths, examine correlated findings, and surface remediation recommendations, running locally under the user’s existing AWS credentials. Responses return both text summaries for the model to reason over and interactive visualizations so users can verify results directly. Source

2. AWS Glue Data Quality Adds ML-Powered Anomaly Detection And Catalog Result Writes

AWS. AWS added anomaly detection to Glue Data Quality that uses ML-powered time-series forecasting to flag unexpected shifts in data statistics without requiring users to define explicit threshold rules. The service can now write evaluation output, including rule results, profiling metrics, and anomaly predictions with confidence bounds, back to Data Catalog tables as queryable records. This produces a persistent audit trail of quality evaluations that can be analyzed with standard SQL, whether checks run inside ETL jobs or directly against Catalog tables. Source