Read all about the latest AWS security features, compliance updates, and hands-on resources in our new, monthly digest posts. You’ll find expert blog posts, new service capabilities, code samples, and workshops.

This month’s AWS Security Blog posts covered AI security, identity and access management, threat intelligence, data protection, and multicloud operations. Whether you’re securing agentic AI systems, upgrading to post-quantum cryptography, or streamlining forensic collection, these posts offer practical guidance across the security landscape.

Access control with IAM Identity Center session tags
Author: Rashmi Iyer | Published: April 28, 2026
Learn to combine AWS IAM Identity Center permission sets with session tags from Microsoft Entra ID to implement fine-grained attribute-based access control (ABAC) across multiple AWS accounts.

Can I do that with policy? Understanding the AWS Service Authorization Reference
Authors: Anshu Bathla, Prafful Gupta | Published: April 27, 2026
Learn to use the AWS Service Authorization Reference to determine what’s achievable with IAM policies, recognize scenarios needing alternative solutions, and build more effective security controls.

Secure AI agent access patterns to AWS resources using Model Context Protocol
Author: Riggs Goodman III | Published: April 14, 2026
Learn to secure AI agent access to AWS resources via MCP using three principles: least privilege, organizational role governance, and differentiating AI-driven from human-initiated actions.

Four security principles for agentic AI systems
Authors: Mark Ryland, Riggs Goodman III, Todd MacDermid | Published: April 2, 2026
Learn four security principles from AWS’s NIST response for securing agentic AI: secure development lifecycle, traditional controls, deterministic external enforcement, and earned autonomy through evaluation.

Designing trust and safety into Amazon Bedrock powered applications
Author: Victor Lungu | Published: April 29, 2026
Learn to integrate responsible AI concepts into Amazon Bedrock applications, including abuse detection, Amazon CloudWatch monitoring, Bedrock Guardrails configuration, and the abuse response process.

Building AI defenses at scale: before the threats emerge
Author: Amy Herzog | Published: April 7, 2026
AWS CISO announces Project Glasswing with Anthropic, introducing Claude Mythos Preview for vulnerability research, plus the general availability of AWS Security Agent for autonomous penetration testing.

Shift-Left Tag Compliance using AWS Organizations and Terraform
Authors: Welly Siauw, Sourav Kundu, Manu Chandrasekhar | Published: April 27, 2026
Learn to validate tag compliance during development using AWS Organizations tag policies, a reusable Terraform tagging module, and a test-driven approach that dynamically validates against live organizational policies.

What the March 2026 Threat Technique Catalog update means for your AWS environment
Authors: Shannon Brazil, Cydney Stude | Published: April 28, 2026
The AWS CIRT’s latest Threat Technique Catalog update covers Amazon Cognito refresh token abuse, AMI image deletion targeting recovery, and trust policy modifications for persistence and privilege escalation.

A framework for securely collecting forensic artifacts into S3 buckets
Authors: Jason Garman, Vaishnav Murthy | Published: April 8, 2026
Learn to securely collect forensic artifacts into Amazon S3 using time-limited, least-privilege credentials with AWS STS session policies and automated AWS Step Functions workflows.

Transform security logs into OCSF format using a configuration-driven ETL solution
Authors: Vivek Gautam, Arpit Gupta, Ryan Gomes | Published: April 17, 2026
Learn to transform custom security logs into OCSF format using an AWS ProServe configuration-driven ETL solution with AWS Step Functions, AWS Glue or Amazon EMR Serverless, and Amazon Security Lake integration.

A technical walkthrough of multicloud full-stack security using AWS Security Hub Extended
Authors: Matt Meck, Michael Fuller | Published: April 22, 2026
Learn how AWS Security Hub Extended simplifies multicloud security procurement and operations through curated partner solutions, unified billing, and OCSF-based findings consolidation.

Protecting your secrets from tomorrow’s quantum risks
Authors: Stéphanie Mbappe, Tobias Nickl | Published: April 24, 2026
Learn to upgrade AWS Secrets Manager clients to use hybrid post-quantum TLS with ML-KEM, protecting secrets against harvest-now-decrypt-later attacks, and verify connections via AWS CloudTrail.

How AWS KMS and AWS Encryption SDK overcome symmetric encryption bounds
Authors: Panos Kampanakis, Matthew Campagna, Patrick Palmer | Published: April 3, 2026
Learn how AWS Key Management Service and the AWS Encryption SDK use derived key methods to automatically handle AES-GCM encryption limits, eliminating the need to manually track bounds or rotate keys.

How to clone an AWS CloudHSM cluster across Regions
Authors: Desiree Brunner, Rickard Löfström | Published: April 20, 2026
Learn to clone an AWS CloudHSM cluster to another Region using CopyBackupToRegion, then synchronize keys—including non-exportable keys—across cloned clusters for disaster recovery.

Investigations of reported security vulnerabilities affecting Amazon and AWS services, software, and products.

• CVE-2026-7791 – Local Privilege Escalation via TOCTOU Race Condition in Amazon WorkSpaces Skylight Agent
• CVE-2026-7461 – OS Command Injection in Amazon ECS Agent via FSx Windows File Server Volume Credentials
• Issue with FreeRTOS-Plus-TCP – IPv6 Router Advertisement Memory Safety Issues
• CVE-2026-7424 – Integer Underflow in DHCPv6 Sub-Option Parser in FreeRTOS-Plus-TCP
• Issue with FreeRTOS-Plus-TCP – MAC Address Validation Bypass and ICMP Echo Reply Integer Underflow
• CVE-2026-7191- Arbitrary Code Execution via Sandbox Bypass in QnABot on AWS
• Issues in tough library and tuftool CLI utility
• Issue with AWS Ops Wheel (CVE-2026-6911 and CVE-2026-6912)
• CVE-2026-6550 – Key commitment policy bypass via shared key cache in AWS Encryption SDK for Python
• CVE-2026-6437 – Mount Option Injection in Amazon EFS CSI Driver
• CVE-2026-5747 – Out-of-bounds Write in Firecracker virtio-pci Transport
• Issues with AWS Research and Engineering Studio (RES)
• Issues with Amazon Athena ODBC Driver
• CVE-2026-5429 – Kiro IDE Webview Cross-Site Scripting via Workspace Color Theme

This month brings 16 new AWS samples spanning identity, governance, compliance, detection and incident response, AI Security, data protection, and infrastructure security. From beginner-friendly AI agent development on Amazon Bedrock to automated Control Tower re-registration at scale, these ready-to-deploy repositories help you implement security best practices across your AWS environment.

Amazon Cognito OAuth2 Token Proxy with Caching
Learn to deploy an Amazon API Gateway proxy for Cognito’s OAuth2 token endpoint with intelligent caching and AWS WAF protection, reducing M2M authentication costs by over 90%.

Cognito API Gateway Authorization Demo
Learn to implement user-specific data protection using Amazon Cognito, API Gateway, and an AWS Lambda authorizer that enforces JWT sub claim matching to prevent cross-user data access.

Securely Connecting On-Premises Data Systems to Amazon Redshift with IAM Roles Anywhere
Learn to deploy a fully private environment connecting on-premises workloads to Amazon Redshift using X.509 certificate authentication via IAM Roles Anywhere for short-lived credentials.

AWS IAM Access Key Lifecycle Management with Human Approval
Learn to automate organization-wide detection, disabling, and deletion of unused IAM access keys using Step Functions, IAM Access Analyzer, and a secure human-in-the-loop approval workflow.

Secrets Manager Audit
Learn to resolve and report who can access your AWS Secrets Manager secrets—across accounts, through Identity Center, and down to the human behind the IAM role—in a single command.

Control Tower Organization Re-Registration Automation
Learn to automate AWS Control Tower OU re-registration and account updates at scale using lifecycle events, Amazon EventBridge, and AWS Lambda to resolve mixed governance after landing zone changes.

Sample Agent Skills for Builders
A curated collection of installable agent skills that extend AI coding agents (Claude Code, Cursor, Copilot) with production-ready AWS, CDK, security scanning, and engineering workflows.

How to Stop AI Agent Hallucinations: 5 Techniques + Production on Amazon Bedrock AgentCore
Learn to detect, prevent, and self-correct AI agent hallucinations using Graph-RAG, semantic tool selection, multi-agent validation, neurosymbolic guardrails, and agent steering with Strands Agents.

Compliance Lens
Learn to deploy a serverless solution that analyzes AWS Config snapshots across an AWS Organization, compares them against conformance pack rule sets, and visualizes compliance posture via Amazon QuickSight dashboards.

AWS Security Agent Terraform Configuration
Learn to provision AWS Security Agent resources using the AWSCC Terraform provider, automating agent space creation, IAM roles, target domain registration, and penetration test setup.

AWS Security Agent Demo Suite
Learn to use AWS Security Agent across three scenarios: automated design reviews, AI-generated infrastructure code review via GitHub, and penetration testing against intentionally vulnerable applications.

Agentic SOC Workshop — CDK Infrastructure
Learn to build an AI-powered Security Operations Center agent that investigates Amazon GuardDuty findings, queries CloudTrail logs, and takes automated containment actions using Amazon Bedrock AgentCore.

Implementing Kerberos Authentication for Apache Spark Jobs on Amazon EMR on EKS to Access a Kerberos-Enabled Hive Metastore
Learn to configure Kerberos authentication for Spark jobs on Amazon EMR on Amazon Elastic Kubernetes Service, connecting to a Kerberos-enabled Hive Metastore using Microsoft Active Directory as the KDC.

AWS Nitro Enclaves with Kubernetes – Hello World Example
Learn to deploy a Hello World application inside an AWS Nitro Enclave on Amazon EKS, covering cluster creation, device plugin setup, and enclave image building.

Multi-Tenant OpenClaw on Firecracker
Learn to deploy isolated, multi-tenant OpenClaw AI agents on AWS using Firecracker microVMs with per-tenant kernel/network isolation, auto-scaling, backup/restore, and a web management console.

Amazon Bedrock for Beginners – From First Prompt to AI Agent
Learn to build AI applications on Amazon Bedrock, from basic API calls to a full agent with RAG, guardrails, tool use, and the Strands Agents SDK.

April 2026 reinforces that securing AI workloads now requires the same rigor applied to traditional infrastructure. The posts and samples in this edition provide concrete patterns for enforcing least privilege on agentic systems, automating governance at organizational scale, and preparing cryptographic implementations for post-quantum requirements. The security bulletins address vulnerabilities across compute, networking, and developer tooling, reinforcing the need to apply patches consistently. Each resource includes deployment steps or runnable code so you can validate the approach in your own environment before adopting it. Subscribe to the AWS Security Blog RSS feed to receive updates as they publish, and revisit this digest monthly for a consolidated view of what changed and what to act on.

As AI agents, bots, and programmatic access become an increasingly significant portion of web traffic, organizations need better tools to understand, analyze, and manage this activity. Today, we’re excited to announce AI Traffic Analysis dashboards for AWS WAF protection packs—also known as web access control lists (web ACLs)—providing comprehensive visibility into AI bot and agent behavior across your applications.

The rapid proliferation of AI bots—from search engine crawlers to research agents—has fundamentally changed the nature of web traffic. Organizations across industries are discovering that AI agents now represent 30–60% of their total traffic, driving significant infrastructure costs without always generating business value.

Traditional bot management tools weren’t designed for the nuances of AI traffic. Teams need to answer critical questions such as: Which AI organizations are accessing our content? What are they trying to accomplish? Which endpoints are most frequently targeted? How has this activity changed over time? Most importantly, how can we turn this visibility into actionable business decisions?

The new AI Traffic Analysis dashboard provides specialized visibility into AI bot and agent activity, available directly within your AWS WAF protection pack (web ACL) console. With this launch, AWS WAF Bot Control expands its detection coverage to track more than 650 unique bots and agents, offering one of the most comprehensive AI bot detection catalogs available. A detection catalog that will keep growing and be updated to align with the pace of the industry’s changes.

This dashboard goes beyond standard security metrics to deliver AI-specific insights that help you understand and manage this critical traffic segment.

• Bot identification and verification: See which AI bots are accessing your applications, including bot names, owning organizations, and verification status. Quickly distinguish between legitimate AI agents from known organizations and potentially suspicious activity.
• Intent classification: Understand the purpose behind AI bot requests. The dashboard categorizes bot behavior patterns—whether crawling for search indexing, conducting research, gathering training data, or other activities—helping you align access policies with business objectives.
• Access pattern analysis: Identify your most frequently accessed URLs and endpoints by AI agents. This visibility helps you understand which content is most valuable to AI organizations and optimize your infrastructure accordingly.
• Temporal trends and historical analysis: Track AI bot activity patterns by time of day and analyze historical trends over the past 14 days. Detect anomalies, understand peak usage periods, and identify emerging patterns in AI traffic.
• Organization breakdown: View traffic volume segmented by bot owner organization, giving you clear visibility into which AI companies are accessing your content and at what scale.

AI Traffic Analysis dashboards integrate seamlessly with AWS WAF Bot Control for common bots using the same traffic evaluation engine while providing specialized analytics for AI-specific patterns. The dashboards display near real-time summaries based on Amazon CloudWatch metrics collected as AWS WAF evaluates your web traffic.

To access the AI Traffic Analysis dashboard:

1. Navigate to your protection pack (web ACL) in the AWS Management Console for AWS WAF.
2. Select the AI Traffic Analysis tab.
3. Apply filters for bot organization, intent type, or verification status as needed.
4. Analyze the comprehensive visualizations across bot identity, intent classification, access patterns, and temporal trends.

The dashboard populates automatically once your protection pack begins receiving AI bot traffic, so you have visibility exactly when you need it.

This new capability addresses a critical need as organizations navigate the evolving landscape of AI-driven web traffic. With detailed insights into AI bot behavior, you can:

• Make informed access decisions: Understand bot intent before implementing allow or block rules.
• Optimize infrastructure investment: Identify high-traffic endpoints and plan capacity accordingly. Know whether your infrastructure costs are supporting business value or used without programmatic compensation mechanism.
• Implement tiered access strategies: Serve different content or pricing based on AI agent verification and intent.
• Detect anomalies and emerging patterns: Spot unusual patterns that might indicate emerging threats or opportunities. Real-time visibility helps you respond quickly to changes in AI bot behavior.
• Support cross-organizational strategy: Provide data to stakeholders across security, product, and business teams for informed decisions about AI bot access policies and monetization opportunities.
• Customize as needed: AI Traffic analyses are emitted as CloudWatch metrics that an organization can use to customize CloudWatch or another supported observability product as needed. Moreover, by using CloudWatch metrics, an organization can build proactive measures such as alerts or business actions such as rate or limit changes.
• Monetize AI traffic at the edge: For a reference architecture that combines WAF Bot Control AI visibility, traffic control, and content monetization using the x402 payment protocol, see the sample-x402-content-monetization-with-cloudfront-and-waf project on GitHub. It demonstrates how to classify AI bot traffic, enforce per-path pricing policies, and settle payments at the edge using Amazon CloudFront and Lambda@Edge – with zero changes to your existing origins.
  

  Note: This AWS Samples solution is not a supported product in their own right, but educational examples to help our customers use our products for their applications. As our customer, any applications you integrate this example into should be thoroughly tested, secured, and optimized according to your business’s security standards & policies before deploying to production or handling production workloads. Deploying it will provision resources that incur additional AWS charges, so review costs before deploying and delete the stack when no longer needed.

In addition to the console dashboard, you can programmatically query AI bot traffic data using the GetTopPathStatisticsByTraffic action, available through the AWS WAF API, AWS SDKs, and AWS CLI. This action returns the top URI paths by bot traffic volume for a given web ACL and time window. Each path in the response includes request counts, traffic percentages, and the top bots accessing it. You can filter results by bot category (for example, ai), organization, or specific bot name, and use a URI path prefix (for example, /api/) to drill down into specific areas of your application. The following AWS CLI example shows how to query the top paths accessed by AI bots for a specific web ACL.

The following AWS CLI example shows how to query the top paths accessed by AI bots for a specific web ACL:

aws wafv2 get-top-path-statistics-by-traffic \
  --web-acl-arn "arn:aws:wafv2:us-east-1:123456789012:global/webacl/ExampleWebACL/a1b2c3d4-5678-90ab-cdef-EXAMPLE11111" \
  --scope "CLOUDFRONT" \
  --time-window StartTime=2026-02-25T00:00:00Z,EndTime=2026-02-26T00:00:00Z \
  --bot-category "ai" \
  --uri-path-prefix "/api/" \
  --limit 5 \
  --number-of-top-traffic-bots-per-path 3

A sample response:

{
  "TopPathStatistics": [
    {
      "Path": "/api/v1/products",
      "RequestCount": 145320,
      "TrafficPercentage": 32.4,
      "TopBots": [
        { "BotName": "ExampleBotA", "Organization": "ExampleOrgA", "RequestCount": 98210 },
        { "BotName": "ExampleBotB", "Organization": "ExampleOrgB", "RequestCount": 47110 },
        { "BotName": "ExampleBotC", "Organization": "ExampleOrgC", "RequestCount": 0 }
      ]
    },
    {
      "Path": "/api/v2/search",
      "RequestCount": 87650,
      "TrafficPercentage": 19.5,
      "TopBots": [
        { "BotName": "ExampleBotA", "Organization": "ExampleOrgA", "RequestCount": 52300 },
        { "BotName": "ExampleBotC", "Organization": "ExampleOrgC", "RequestCount": 35350 },
        { "BotName": "ExampleBotB", "Organization": "ExampleOrgB", "RequestCount": 0 }
      ]
    }
  ],
  "TimeWindow": {
    "StartTime": "2026-02-25T00:00:00Z",
    "EndTime": "2026-02-26T00:00:00Z"
  }
}

Programmatic access enables you to:

• Build custom dashboards or integrate AI traffic data into existing observability platforms.
• Automate alerting when specific paths see unusual bot traffic spikes.
• Feed traffic data into business intelligence pipelines for content monetization decisions.
• Investigate and debug AI bot activity within a specific timeframe to identify the root cause of traffic anomalies or incidents.

For detailed usage information, see the GetTopPathStatisticsByTraffic API reference and the AWS CLI command reference. This API pairs naturally with the CloudWatch metrics approach described above, giving you both real-time metric streams and on-demand path-level analytics for comprehensive AI traffic management.

For customers on flat-rate pricing plans, the AI Traffic Analysis dashboard is included with all paid plans. Read more about CloudFront flat-rate pricing in the launch blog post. For AWS WAF customers not subscribed to flat-rate plans, the AI traffic analysis dashboard is available at no additional cost. See AWS WAF pricing for details.

The AI Traffic Analysis dashboard represents a significant step forward in managing the intersection of AI and web security. As AI agents continue to grow as a percentage of overall web traffic, having the right visibility tools becomes essential for both security and business success.

To learn more about AWS WAF Bot Control and AI Traffic Analysis dashboards, visit the AWS WAF Developer Guide or explore the feature directly in your AWS WAF console.

If you have feedback about this post, submit comments in the Comments section below.

AWS Security Assurance Services is announcing the release of our latest compliance guide, ISO 31000:2018 Risk Management on AWS, which provides practical guidance for organizations establishing and operating a risk management program in AWS environments using ISO 31000:2018 principles.

The guide explains how organizations can integrate AWS services into their risk management processes to support the core components of ISO 31000:2018, including establishing context and criteria, conducting risk assessments, implementing risk treatments, and enabling continuous monitoring and review. It also highlights how AWS security, automation, and monitoring capabilities can help customers identify areas for improvement and help enforce controls at large. The guide includes:

• An overview of the ISO 31000:2018 risk management framework, including context and criteria, risk assessment, risk treatment, and monitoring and review. You will learn how to apply ISO 31000’s core principles within AWS environments and use AWS services for risk identification, detection, treatment, and monitoring.
• Governance and risk treatment considerations aligned with the AWS Shared Responsibility Model. This includes strategies for risk avoidance, mitigation, transfer, and acceptance.

By combining ISO 31000 risk management principles with AWS security services, organizations can build scalable, automated environments that help support continuous risk identification, proactive treatment, operational visibility, and ongoing compliance readiness.

Download Available: ISO 31000:2018 Risk Management on AWS Compliance Guide

For further assistance, contact AWS Security Assurance Services