AWS Security Blog
ICYMI: July 2026 @AWS Security
If you found time for a bit of vacation this summer, you might be in catch-up mode. Here’s a list to help: all the expert blog posts, new service capabilities, code samples, and workshops, in case you missed it, from July 2026.
AWS Security Blog post
This month’s AWS Security Blog posts covered AI agent security, supply chain protection, network firewall automation, DDoS mitigation, and compliance readiness. Read on for guidance on securing AI coding agents, implementing dependency cooldowns, choosing the right key management solution, and preparing for HIPAA Technical Safeguard requirements.
AI Security
Enforce least-privilege authorization in multi-agent AI chains using Cedar
Authors: Dhananjay Karanjkar | Published: July 6, 2026
Learn to implement a three-layer Cedar policy model with OAuth 2.0 authentication to prevent authorization scope expansion across multi-agent delegation chains using Amazon Verified Permissions.
Enforce zero data retention on Amazon Bedrock with Bedrock Projects and service control policies
Author: Rob Higareda | Published: July 7, 2026
Learn to use Amazon Bedrock Projects and SCPs to centrally enforce zero data retention policies, preventing accounts from enabling data sharing with third-party model providers across your organization.
Designing for the inevitable: System prompt leakage and mitigations in generative AI applications
Author: Manideep Konakandla | Published: July 8, 2026
Learn to implement defense-in-depth mitigations for system prompt leakage using Amazon Bedrock Guardrails prompt attack filters, canary tokens, semantic similarity detection, and sandwich instruction patterns.
Balancing speed and safety: A control framework for AI coding agents
Authors: Daniel Begimher, Danny Cortegaca | Published: July 30, 2026
Learn to implement an application security control framework for AI coding agents, with author-time controls that shape what agents produce and build-time controls that verify what reaches production.
Data Protection
How to use the AWS Workload Credentials Provider for cross-account secret retrieval and prefetching secrets
Authors: Derik Wang, Paras Dhawan | Published: July 1, 2026
Learn to configure the AWS Workload Credentials Provider for cross-account secret retrieval using IAM role chaining and prefetching secrets at startup to reduce cold-start latency.
The CISO’s guide to post-quantum mandates and migrations
Author: Rushir Patel | Published: July 8, 2026
A strategic playbook for CISOs navigating post-quantum cryptography migration, covering regulatory timelines, dependency classification, cryptographic telemetry, and building crypto-agile organizations.
AWS KMS or AWS CloudHSM: Choose the right key management solution
Author: Derek Tumulak | Published: July 28, 2026
Learn how to choose between AWS KMS and AWS CloudHSM based on integration needs, cost, and whether you require traditional HSM interfaces or legacy algorithms.
Secure your npm and pip package updates in Amazon Linux
Author: Norbert Manthey | Published: July 29, 2026
Learn to implement a one-line dependency cooldown for npm and pip that skips packages published in the last 24 hours, protecting against supply chain events while still allowing urgent security patches.
Infrastructure security
Secure Amazon container workloads using container attribute-based rules in AWS Network Firewall
Authors: Amit Gaur, Amish Shah, Preetkumar Shah, Akash Kumar Sinha | Published: July 1, 2026
Learn to define AWS Network Firewall rules for Amazon EKS and Amazon ECS workloads using native container attributes like namespaces, pod names, and labels instead of ephemeral IP addresses.
Authenticate legitimate AI agent traffic with AWS WAF Bot Control
Authors: Harith Gaddamanugu, Kaustubh Phatak | Published: July 14, 2026
Learn to use Web Bot Authentication (WBA) in AWS WAF Bot Control to cryptographically verify legitimate AI agent traffic using HTTP message signatures and ed25519 keys.
Accelerating AWS Network Firewall troubleshooting with AWS DevOps Agent
Author: Salman Ahmed | Published: July 24, 2026
Learn to use AWS DevOps Agent to automate root cause analysis for AWS Network Firewall connectivity issues, including domain deny lists, stateless rule priority misconfigurations, and asymmetric cross-AZ routing drops.
AWS Shield Advanced is embracing the AWS WAF Anti-DDoS managed rule group: What changes and how to prepare
Authors: Eitav Arditti, Andrew Chen, Justin Kurpius | Published: July 27, 2026
AWS Shield Advanced is adopting the AWS WAF Anti-DDoS managed rule group as its default application-layer DDoS protection, with a phased migration from July 2026 through January 2027.
Threat detection and incident response
Introducing the Amazon GuardDuty investigation agent: on-demand AI-powered threat assessment
Author: Allan Holmes | Published: July 20, 2026
Learn to use the new Amazon GuardDuty investigation agent (public preview) to automate threat correlation and receive structured assessments with risk levels, confidence scores, MITRE ATT&CK mappings, and actionable recommendations.
Amazon identifies North Korean hacker group behind open-source supply chain attacks
Author: CJ Moses | Published: July 29, 2026
Learn how Amazon Threat Intelligence linked the compromises of axios, debug, chalk, and typo-crypto NPM packages to a single DPRK-linked threat actor, and how attacker tradecraft is evolving with generative AI.
Extend Amazon Inspector SBOM Generator with plugins
Authors: Michael Long, Anthony Verleysen, Charlie Bacon | Published: July 30, 2026
Learn to write custom Lua plugins for the Amazon Inspector SBOM Generator to inventory package ecosystems that aren’t supported out of the box, without modifying source code or waiting for an official release.
Security Hub adds AI workload protection and multicloud support for Microsoft Azure
Author: Michael Fuller | Published: July 14, 2026
AWS Security Hub now monitors Microsoft Azure resources for misconfigurations and vulnerabilities, adds GuardDuty AI Protection for Amazon Bedrock and Amazon SageMaker AI, and introduces an AI inventory for organization-wide visibility.
Governance and compliance
AWS designated as a critical third party to the UK financial sector
Author: Michael Jefferson | Published: July 10, 2026
AWS has been designated as a critical third party to the UK financial sector by HM Treasury, establishing direct regulatory oversight by the Bank of England, PRA, and FCA.
New compliance guidance available: HITRUST i1 on AWS
Authors: Abdul Javid, Shreya Singh | Published: July 13, 2026
AWS published new implementation guidance for HITRUST i1 certification, covering 11 technical control domains with AWS-specific controls for healthcare organizations seeking i1 assessment readiness.
HIPAA Security Rule on AWS – Technical Safeguards Implementation and Readiness Guidance
Authors: Abdul Javid, Hector Rodriguez, Kapil Temghare, Shreya Singh | Published: July 31, 2026
New guidance helping covered entities and business associates implement and evidence compliance with HIPAA Security Rule Technical Safeguards (§164.312) on AWS, including 2025 NPRM proposed changes.
Identity
Introducing OAuth support for AWS MCP Server
Authors: Vaibhav Chowla, Jaimin Bhatt, Ankur Joshi | Published: July 9, 2026
AWS MCP Server now supports OAuth 2.1 authorization through AWS Sign-In, enabling agents like Claude Code,Kiro, and Gemini CLI to connect using existing IAM credentials with browser-based authentication.
July Security Bulletins
In July 2026, AWS published 21 security bulletins (2026-049 through 2026-069) addressing vulnerabilities across open-source SDKs, MCP servers, and developer tools. Key themes include credential disclosure and SSRF, affecting HealthLake, HealthOmics, and API MCP servers, plus Strands Agents tools that could inadvertently expose secrets to unauthorized endpoints. Command and code injection impacted aws-cdk-lib, jsii-diff, Bedrock AgentCore SDK, and Amplify Codegen UI. The smithy-rs framework received three patches for denial-of-service via uncontrolled recursion and Slowloris issues.
Other notable issues include insecure file permissions in the AWS CLI, deserialization remote code execution in the Advanced JDBC Wrapper, SQL injection in mcp-gateway-registry, TLS 1.3 flaws in s2n-tls, and stored XSS in AWS Ops Wheel. A common thread: insufficient input validation in tools interacting with AI agents, reflecting the expanded surface area of LLM-integrated workflows. All patches are available, upgrade promptly. For more information, see AWS Security Bulletins.
AWS Samples
This month brings 14 new AWS samples spanning AI security, identity, data protection, governance, threat detection, and security posture management. From deploying governed AI agent platforms on Amazon Bedrock AgentCore to building data-residency-compliant chatbots and DevSecOps baselines for Kiro, these repositories help you implement security and governance best practices across your AWS environment.
AI Security
Lark MCP on AgentCore
Learn to deploy a hosted remote MCP service on Amazon Bedrock AgentCore that lets AI agents operate Feishu/Lark through 450+ tools, with per-user identity isolation and smart multi-step orchestration via 20+ domain Skills.
Lark CLI MCP Wrapper on AgentCore Runtime and Identity
Learn to securely wrap a CLI tool as an MCP server on AgentCore Runtime using a sidecar credential-isolation pattern, where the CLI process never holds real tokens and all secrets are resolved through AgentCore Identity’s Token Vault.
LiteLLM Bedrock Gateway on EKS
Learn to deploy a production-grade LiteLLM proxy on Amazon EKS as a unified OpenAI/Anthropic-compatible gateway to Amazon Bedrock, with four progressive layers covering network isolation, cross-region inference profiles, and cross-account delegation.
Enterprise Agentic AI Platform Accelerator on AgentCore
Learn to deploy a secure, governed foundation for production AI agents on Amazon Bedrock AgentCore with CDK stacks covering identity, gateway, memory, runtime, and observability; supporting multiple agent frameworks (Strands, LangGraph, Claude SDK) and opt-in security controls including VPC isolation, KMS encryption, Cedar policies, and Bedrock Guardrails.
FlowAMP: AI Agent Governance on AWS
Learn to deploy a single-pane-of-glass agent management platform on Amazon Bedrock AgentCore that discovers, monitors, scores, controls, and cost-accounts AI agents across an AWS Organization with agentic discovery, compliance scanning (NIST AI RMF, ISO 27001, SOC 2), Responsible-AI scoring, FinOps via Cost Explorer, and Cedar-based policy enforcement.
Kiro SecOps Baseline
Learn to deploy a DevSecOps security baseline for Kiro as a single Go CLI that installs global guardrails (permissions.yaml, steering, skills, a security-review agent) and per-project workspace hooks (fail-closed guard, PR/pipeline review gates, scanner configs for gitleaks, trivy, and checkov) with enterprise fleet distribution via MDM and Administration scope.
Identity
OAuth 2.0 Token Exchange with Amazon Cognito
Learn to implement RFC 8693 OAuth 2.0 Token Exchange using Amazon Cognitowith a true delegation pattern, enabling services to act on behalf of users while maintaining distinct service identities and least-privilege access in microservices architectures.
Lark Identity on AgentCore — Gateway Interceptor
Learn to implement enterprise identity pass-through on Amazon Bedrock AgentCore using a Gateway Request Interceptor that forwards the user’s identity and injects per-user credentials to downstream MCP tools, so the agent never holds a token and tools act only as the authenticated user against Lark.
Data Protection
Automated PII Detection Pipeline with Amazon Macie
Learn to build an event-driven pipeline that automatically detects PII in Amazon S3 objects using Amazon Macie, AWS Step Functions, and custom data identifiers, with CSV/JSON reporting and SNSalerting for high-severity findings.
Data-Residency Chatbot with Amazon Bedrock AgentCore
Learn to build a data-residency-compliant natural-language chatbot on Amazon Bedrock AgentCore that keeps all data and AI inference within a single AWS Region, using governed text-to-SQL with whitelist-validated queries, Aurora PostgreSQL in private subnets, and AgentCore Gateway for secure tool access.
Governance and compliance
Video Compliance Agent
Learn to build an end-to-end automated video compliance verification pipeline using Amazon Bedrock, ECS Fargate, and AWS Step Functions that processes videos shot-by-shot, extracting frames, audio transcripts, and OCR text, then flags potential broadcast guideline violations with structured per-shot reports.
Contract Compliance Search with Amazon OpenSearch
Learn to build a contract compliance search system that combines semantic search with semantic highlighting using Amazon OpenSearchService, Amazon Titan V2 embeddings, and a SageMaker-hosted highlighting model to surface relevant clauses across contract documents.
Threat detection and incident response
Multicloud Security Posture Assessment
Learn to deploy a centralized security assessment solution that scans AWS, Azure, Google Cloud Platform, and Oracle Cloud Infrastructure environments from a single AWS deployment using Prowler, with AWS CloudFormation templates for each provider and unified reporting in HTML, CSV, and JSON-OCSF formats.
Centralize AWS Security Agent Findings
Learn to deploy a AWS CloudFormation stack that automatically exports AWS Security Agentpenetration test findings to Amazon S3 and queries them centrally with Amazon Athena, using Amazon EventBridge, Step Functions, and a AWS Glue catalog for tracking findings over time.
Sentinel Harness — Production SecOps Agents as Configuration
Learn to build production security-operations agents as pure configuration on Amazon Bedrock AgentCore Harness, declaring model, prompt, tools, skills, memory, and limits in YAML while AWS runs the agent loop with human-in-the-loop gates, detection-engineering tools, adversary emulation, and a self-improvement closed loop.
AWS Labs
This month brings 1 new AWS Labs repository focused on data protection, helping organizations build automated PII detection and redaction pipelines with AI-powered processing across documents and audio files.
Data Protection
PII Anonymizer
Learn to build an automated PII detection and redaction pipeline using AWS Step Functions, Amazon Bedrock, Amazon Textract, and Amazon Transcribe; supporting PDFs, Word, Excel, images, and audio files with synthetic replacement or blackout modes, concurrency control, and customer-managed KMS encryption.
Conclusion
July 2026 provides guidance and examples for securing AI agent architectures at scale, from governed text-to-SQL with data residency controls and agent management platforms to DevSecOps baselines for AI coding tools. The posts and samples provide patterns for least-privilege authorization in multi-agent chains using Cedar, post-quantum migration planning, container-aware network firewall rules, and multicloud security posture management. Each resource includes deployment steps or runnable code so you can validate in your own environment before adopting. 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.
If you have feedback about this post, submit comments in the Comments section below.