AWS News - 2026-11-11

2026-11-11
最終更新: 2026-11-11 19:13:51 JST

AI による概要

概要 14 記事 / 一覧 23 記事

この日はセキュリティ速報 3 件と、AWS Nitro Sandboxes・AgentCore まわりの記事が中心でした。セキュリティでは CVE-2026-125271 が公表され、AWS API MCP Server が操作名を誤って正規化してから比較していたため、セキュリティポリシーを回避できる問題があったとされています。Architecture Blog では、規制業界の企業で中央管理された ingress プロキシの背後にあっても、Amazon Application Recovery Controller (ARC) Region Switch を使ってアプリケーションチームがセルフサービスでリージョンフェイルオーバーを実行する構成が紹介されました。Machine Learning Blog では、複数アカウントの Bedrock 呼び出しログを集約してトークン消費を分析する観測用データレイク (tokenomics data lake) の構築方法と、Tata Consultancy Services (TCS) が AgentCore 上で要件から統制されたテストを生成・実行するエージェント型テストオーケストレーターを構築した BFSI 向け事例が取り上げられています。あわせて CVE-2026-124848 と CVE-2026-126087 も公表されました。CVE-2026-124848 は macOS と Linux の Kiro CLI 2.27.0 より前のバージョンで、ローカルのセッション記録とプロンプト履歴のファイルが全ローカルユーザーから読める権限で作られていた問題で、同じホストの他のユーザーが記録に含まれる認証情報やソースコードを読めるおそれがあります (影響範囲は 1.16.0 以上 2.27.0 未満)。CVE-2026-126087 は OpenJD Sessions for Python 0.11.0 未満 (AWS Deadline Cloud ワーカーエージェント 0.33.0 未満が利用) のクロスユーザープロセスランチャーにおける信頼できない検索パスの問題で、キューにジョブを投入できる認証済みリモートユーザーが PATH 環境変数を書き換えるジョブを送ることで、ワーカーエージェントの OS ユーザーとして任意コードを実行し、そのエージェントが保持する AWS 認証情報を取得できるおそれがあります。Compute Blog では、コンテナ・EKS ポッド・ECS タスクのそれぞれにハードウェア分離された仮想マシンの境界を与える新しい Amazon EC2 のビルディングブロック AWS Nitro Sandboxes が解説され、Nitro System の強い分離・オペレーターアクセスなし・暗号的な構成証明をほぼネイティブの性能でコンテナワークロードに持ち込むとしています。Machine Learning Blog では、AgentCore Evaluations の組み込み評価器・DeepEval・AutoEval を使い分けて、エージェントがタスクを完了したか、話題から逸れていないか、データを漏らしていないか、手順を飛ばしていないかを評価する方法と、Trovata が Bedrock と AgentCore 上に RBAC で保護した銀行ブランド向けの自律型トレジャリーエージェントを構築した事例が公開されました。このほか Amazon RDS for PostgreSQL が Extended Support のマイナーバージョン 13.23-rds.20260819・12.22-rds.20260820・11.22-rds.20260822 を提供し、Amazon RDS for MySQL 26.10 が RDS Database Preview Environment で評価できるようになり、AWS Partner Central の事業計画機能が刷新されて AI による計画の準備状況スコアやリアルタイムの進捗追跡、計画のエクスポートが加わりました。セキュリティブログでは AWS ProcessUnity Global Risk Exchange 評価の更新と、re:Invent 2026 (11 月 30 日〜12 月 4 日、2,200 以上のセッション) のデジタルソブリンティ関連セッションの案内が出ています。

主要トピック
  • セキュリティ: CVE-2026-125271 (AWS API MCP Server の操作名正規化の誤りによるセキュリティポリシー回避)

  • レジリエンス: ARC Region Switch によるセルフサービスのリージョンフェイルオーバー

  • 生成 AI: Bedrock 呼び出しログを集約する tokenomics データレイク

  • 事例: TCS が AgentCore でエージェント型テストオーケストレーターを構築

  • セキュリティ速報: CVE-2026-124848 (Kiro CLI 2.27.0 未満の macOS / Linux で、セッション記録とプロンプト履歴が全ローカルユーザーから読める権限で保存される。影響範囲は 1.16.0 以上 2.27.0 未満)

  • セキュリティ速報: CVE-2026-126087 (OpenJD Sessions for Python 0.11.0 未満と AWS Deadline Cloud ワーカーエージェント 0.33.0 未満で、PATH を書き換えるジョブからワーカーエージェントの権限で任意コード実行と AWS 認証情報の取得)

  • コンピューティング: AWS Nitro Sandboxes の解説、コンテナ・EKS ポッド・ECS タスクごとにハードウェア分離された VM の境界を与え、オペレーターアクセスなしと暗号的な構成証明をほぼネイティブ性能で提供

  • 生成 AI: AgentCore Evaluations の組み込み評価器・DeepEval・AutoEval の使い分け、Trovata が RBAC で保護した銀行ブランド向けの自律型トレジャリーエージェントを構築

  • データベース: RDS for PostgreSQL が Extended Support のマイナーバージョン 13.23・12.22・11.22 を提供、RDS for MySQL 26.10 が Database Preview Environment で評価可能に

  • パートナー・イベント: AWS Partner Central の事業計画機能を刷新 (AI による準備状況スコアと進捗追跡)、ProcessUnity Global Risk Exchange 評価の更新、re:Invent 2026 のデジタルソブリンティ関連セッション案内

AI (Claude Opus 5) が生成 · 2026-11-11 08:09:23 JST

AWS What's New

AWS Partner Central now provides an improved business planning experience

AWS Partner Central now offers a new business planning experience in the AWS Console. AWS Partners can initiate plans for growing their AWS business or collaborate on plans initiated by AWS teams.

Partners can collaborate with AWS partner development, technology, marketing, and specialist teams to define plan narratives, goals, and initiatives. AI-assisted plan readiness scores help partners identify areas to improve, while real-time tracking shows progress against goals. Partners can also add attachments and export plans to prepare for business reviews.

Partners can use application programming interfaces (APIs) and software development kits (SDKs) to integrate business planning into their workflows. Programmatic notifications are also available.

Business planning is available to AWS Partners who are actively engaged with AWS Partner management teams. AWS Partner Central is available in the US East (N. Virginia) Region. To get started, sign in to AWS Partner Central and choose Business Plan.

Amazon RDS for PostgreSQL announces Extended Support minor versions 13.23-rds.20260819, 12.22-rds.20260820 and 11.22-rds.20260822

Amazon Relational Database Service (RDS) for PostgreSQL announces Amazon RDS Extended Support minor versions 13.23-rds.20260819, 12.22-rds.20260820 and 11.22-rds.20260822.

Amazon RDS Extended Support provides up to three additional years of fixes for critical Common Vulnerabilities and Exposures (CVE) and bugs beyond a major version's end of standard support date, giving you more time to upgrade to a new major version.

You can upgrade your database using Amazon RDS Blue/Green Deployments, in-place upgrade, or restore from a snapshot. Learn more about performing version upgrades in the Amazon RDS for PostgreSQL User Guide.

Amazon RDS for PostgreSQL makes it simple to set up, operate, and scale PostgreSQL deployments in the cloud. Create or update a fully managed Amazon RDS database in the Amazon RDS Management Console or by using the AWS Command Line Interface (AWS CLI).

Amazon RDS for MySQL 26.10 is now available in Amazon RDS Database Preview Environment

Starting today, Amazon RDS for MySQL 26.10 is available in the Amazon RDS Database Preview Environment, allowing you to evaluate the pre-release of MySQL 26.10 on Amazon RDS for MySQL.

MySQL 26.10 includes bug fixes, security patches, and some improvements to READ COMMITTED workload performance and purge lag warnings in the error log. On busy READ COMMITTED workloads, NUMA systems performance could be impacted as each read view that closed required a global mutex. Read-view objects can now be reused which eliminates the global mutex on closure. The error log now reports warnings when the history-list length (HLL) crosses innodb_max_purge_lag and when purge lag begins delaying DML, so you can detect purge backlogs before they affect performance. To learn more, please refer to the MySQL 26.10 release notes.

Amazon RDS Database Preview Environment database instances are retained for a maximum period of 60 days and are automatically deleted after the retention period. Amazon RDS database snapshots that are created in the Preview Environment can only be used to create or restore database instances within the Preview Environment. Amazon RDS Database Preview Environment database instances are priced as per the pricing in the US East (Ohio) Region.

Amazon OpenSearch Service announces managed cross-Region disaster recovery

詳細を表示

Amazon OpenSearch Service now offers managed cross-Region disaster recovery, so you can meet strict availability and business continuity requirements for mission-critical search, AI, and analytics workloads without building or maintaining replication yourself. After you create domains in two Regions, you can pair them with a single API call or a few steps in the AWS Management Console.

Managed disaster recovery delivers a recovery point objective (RPO) of under one minute and a recovery time objective (RTO) of under 20 minutes. The secondary domain stays continuously in sync with the primary. For planned events such as maintenance or Region rotation, a switchover swaps the two domains' roles with zero data loss. If your primary domain becomes unavailable, a failover promotes the secondary, with data loss limited to the replication lag. After a failover, you restore protection by pairing the new primary with a secondary domain. You can monitor replication health and lag with Amazon CloudWatch metrics.

Managed disaster recovery requires OpenSearch 3.7 or later and supports domains in the same AWS account, including domains encrypted with customer managed AWS Key Management Service (AWS KMS) keys. Domains that use UltraWarm storage or OpenSearch Optimized instances (OR1 or OR2) aren't supported. If you're on an earlier version, clone-based upgrades let you test OpenSearch 3.7 against your workload, cut over when you're ready, and roll back if needed.

There's no additional charge for managed disaster recovery. You pay standard rates for the secondary domain and cross-Region data transfer. For details, see Amazon OpenSearch Service pricing.

Managed disaster recovery is available today in all commercial AWS Regions where Amazon OpenSearch Service is available (Region list). To get started, see the disaster recovery documentation.

Amazon EC2 G7 instances now available in additional regions

Amazon Elastic Compute Cloud (Amazon EC2) G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs are now available in Europe (London), Asia Pacific (Mumbai, Sydney) and South America (Sao Paulo) Regions. G7 instances deliver up to 4.6x AI inference performance and up to 2.1 graphics performance compared to G6 instances. G7 instances also deliver faster performance for GPU-accelerated data analytics workloads.

Customers can use G7 instances for deploying AI models for language translation, video and image analysis, speech recognition and recommender systems. They also accelerate graphics workloads such as creating and rendering real-time, cinematic-quality graphics and game streaming. Additionally, G7 instances support video transcoding, spatial computing, and data analytics workloads. G7 instances feature up to 8 NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs with 32 GB of memory per GPU and custom Intel Xeon 6 processors. They support up to 192 virtual CPUs (vCPUs) and up to 700 Gbps of Elastic Fabric Adapter (EFA) networking bandwidth. They also support up to 768 GiB of system memory, and up to 7.6 TB of local NVMe SSD storage.

You can use G7 instances for Amazon EC2 in the following AWS Regions: US West (Oregon), US East (N. Virginia, Ohio) , Europe (London, Spain, Frankfurt, Stockholm) , Asia Pacific (Mumbai, Sydney) and South America (Sao Paulo). You can purchase G7 instances as On-Demand Instances, Spot Instances, or as part of Savings Plans.

To get started, visit the AWS Management Console, AWS Command Line Interface (CLI), and AWS SDKs. To learn more, visit the G7 instance page.

AWS Billing and Cost Management now supports AWS Billing Conductor and AWS Billing Transfer views as source views for custom billing views

Custom billing views now support AWS Billing Conductor and AWS Billing Transfer views as source views. These are the views AWS maintains per billing group and per billing transfer relationship. You can combine them into a single custom billing view, apply filters, and use that view in Cost Explorer, AWS Budgets, and the Billing and Cost Management console. There are three source view types. Billing group views contain the cost of each Billing Conductor billing group, which you can combine for chargeback and showback across business units. Billing transfer views contain the cost transferred to you for each end customer, and transfer showback views contain the cost that end customer sees. Aggregating each type gives you your total transferred cost and the total bill your end customers see.

Your access does not change: you can use an AWS-managed view as a source only where you can already access it. Once created, you can share a custom billing view using AWS Resource Access Manager with accounts inside or outside your organization, so your finance team or a downstream seller works with the data in their own account. Whether you are an AWS Partner reselling AWS services or an organization using billing groups for pro forma billing, you get one consolidated view instead of one per customer or group.

This feature is available in all commercial AWS Regions at no additional cost. To get started, navigate to Billing Views in the AWS Billing and Cost Management console or view the user guide.

AWS Japan Blog

AIが進め、人が決める ― AI駆動開発の本当のボトルネックはどこにあったか

こんにちは。AWS ソリューションアーキテクトの崔 祐碩(Woosuk Choi)です。AI 開発を進めている多くのお客様から「開発を加速したい」というご相談をいただいています。お客様と議論を重ねるたびに、本当のボトルネックは開発そのものではなく、意思決定をめぐるプロセスと組織ではないか、と感じることが増えてきました。本ブログでは、2026年7月29日に MBK Wellness 株式会社と実施した AI-DLC ワークショップを題材に、私たちがこの仮説をどう確かめ、どんな答えが返ってきたのかをご紹介します。

製造業 7 社が率直に語った AI エージェント全社基盤の実践と課題:生成 AI 基盤・活用推進 ラウンドテーブル 2026年10月 開催報告

製造業のお客様を支援しているソリューションアーキテクトの西田です。 2026 年 10 月 26 日に AWS […]

Kiro の「Approve for me」のご紹介

2026 年 11 月 10 日、Kiro IDE、Kiro CLI、Kiro Web で「Approve for me」の提供を始めました。Approve for me を有効にすると、ルールで決まっていない操作について、Kiro が LLM によるレビューを行います。

AWS Security Blog

Digital sovereignty at re:Invent 2026: Sessions you won’t want to miss

AWS re:Invent returns to Las Vegas for cloud and AI’s biggest week from November 30–December 4, 2026. This year we have over 2,200 sessions, 70 percent of which are interactive, led by experts from Amazon Web Services (AWS), our customers, and our partners. With that much to choose from, we’ve pulled together the sessions built […]

AWS refreshes ProcessUnity Global Risk Exchange assessment to enhance customer’s third-party due diligence

Amazon Web Services (AWS) is pleased to announce renewal of the AWS ProcessUnity Global Risk Exchange assessment. This validated assessment report helps customers perform effective cloud supplier due diligence on AWS and enhances their third-party risk management process. What is the AWS ProcessUnity Global Risk Exchange assessment Many customers use third-party cyber risk management (TPCRM) […]

Reduce alert fatigue by customizing threat detection features in Amazon GuardDuty

The teams responsible for security today don’t have a problem with the number of alerts they receive; they have a problem turning those alerts into something truly actionable. When an analyst concludes an investigation with “this was caused by our vulnerability scanner” or “this account is configured to share Amazon Machine Images (AMIs) publicly on […]

AWS Security Bulletins

CVE-2026-125271 - AWS API MCP Server Security Policy Bypass via Incorrect Operation Name Normalization

Bulletin ID: 2026-176-AWS
Publication Date: 11/10/2026 08:30 AM PST

Description:

We identified CVE-2026-125271 of AWS API MCP Server. The server normalized operation names incorrectly before comparing them to the operator ' s security policy, so operations whose names contain consecutive capital letters never matched a correctly written denyList or elicitList entry. This might allow an actor to execute AWS API operations that the operator's security policy was configured to deny or require elicitation for.

Impacted versions: >= 0.2.13 AND < 1.5.1

Refer to the article for the most up-to-date and complete information.

CVE-2026-126087 - Untrusted search path in OpenJD Sessions for Python and the AWS Deadline Cloud worker agent

Bulletin ID: 2026-177-AWS
Publication Date: 11/10/2026 10:45 AM PST

Description:

Untrusted search path in the cross-user process launcher in Open Job Description (OpenJD) Sessions for Python before 0.11.0, as used in the AWS Deadline Cloud worker agent before 0.33.0, on Linux might allow an authenticated remote user with permission to submit jobs to a queue to execute arbitrary code as the worker agent OS user and obtain the AWS credentials held by the worker agent via a crafted job that modifies the PATH environment variable.

Impacted versions:
- openjd-sessions < 0.11.0
- deadline-cloud-worker-agent < 0.33.0

Refer to the article for the most up-to-date and complete information.

CVE-2026-124848 - Incorrect default permissions on session transcripts and prompt history in Kiro CLI

Bulletin ID: 2026-178-AWS
Publication Date: 11/10/2026 12:00 PM PST

Description:

Incorrect default file permissions in Kiro CLI before version 2.27.0 on macOS and Linux might allow other local users on the same host to read a user's session transcripts and prompt history, including any credentials or source code they contain. The CLI created its local session storage files with permissions that made them readable by all local users.

Impacted versions: Kiro CLI >= 1.16.0 AND < 2.27.0

Refer to the article for the most up-to-date and complete information.

AWS Architecture Blog

Self-service regional failover with Amazon Application Recovery Controller Region Switch and centralized ingress controls

Application teams in regulated enterprises can execute self-service regional failover with Amazon Application Recovery Controller (ARC) Region Switch, even behind centrally managed ingress proxy fleets. Learn how Transit Gateway peering, Route 53 Private Hosted Zones, and VPC Lattice preserve security controls while removing cross-team coordination during DR.

AWS Machine Learning Blog

Amazon Bedrock tokenomics data lake

Running Amazon Bedrock across many accounts shows you the bill but not what happened inside each call. This post builds a centralized observability data lake that aggregates Bedrock invocation logs from every linked account into one queryable table, so you can see prompts, token counts, model choice, and per-invocation cost across your organization.

How TCS built an agentic test orchestrator with Amazon Bedrock AgentCore

Tata Consultancy Services (TCS) built an agentic test orchestrator on Amazon Bedrock AgentCore that turns requirements into executable, governed tests for its Banking, Financial Services, and Insurance (BFSI) clients. Learn how specialized agents, per-tenant isolation, and workload-aware execution cut testing effort by 88% in an internal benchmark.

Autonomous treasury agents: How Trovata built an AI-native solution on Amazon Bedrock and AgentCore

Learn how Trovata built autonomous treasury agents on Amazon Bedrock and Amazon Bedrock AgentCore, using an RBAC-secured architecture for bank-branded AI deployment and agentic workflows that move corporate treasury from manual operations to real-time decision-making at enterprise scale.

The agent evaluation playbook: Built-in, DeepEval, and AutoEval on Amazon Bedrock AgentCore Evaluations

When you deploy an agent, whether it is working is really several questions: did it finish the task, stay on topic, leak data, or skip steps? This post shows how to evaluate agents on Amazon Bedrock AgentCore with built-in, DeepEval, and AutoEval evaluators, and how to match each behavior that matters to the right evaluator at the right level.

Build a daily dependency watcher with LangGraph and Amazon Bedrock AgentCore web search

The Web Search API in Amazon Bedrock AgentCore gives agents current, cited web results signed with the AWS credentials they already use. In this post, build a daily dependency watcher as a LangGraph graph that researches package ecosystems in parallel, keeps its rules in code, and checks every finding against its cited sources.

Introducing Kimi K3.1 Pro on Amazon Bedrock

Kimi K3.1 Pro from Moonshot AI is now available on Amazon Bedrock. Learn how to get started through the OpenAI-compatible Chat Completions, Responses, and Converse APIs, optimize inference with explicit prompt caching, and use the model as a coding assistant in OpenCode.

AWS Compute Blog

Nitro Sandboxes: Hardware isolation for every EKS pod, ECS task, and EC2 workload

AWS Nitro Sandboxes is a new Amazon EC2 building block that gives each container, EKS pod, or ECS task the boundary of a hardware-isolated virtual machine, bringing the Nitro System's strong isolation, zero operator access, and cryptographic attestation to container workloads at near-native performance.