[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fs4WysjfNwnAGnQjBaOSVUJ4bZlG-wboJu7U23PcWuq8":3},{"lesson":4},{"id":5,"slug":6,"article_id":7,"title":8,"body":9,"prevention":10,"framework_refs":11,"status":18,"created_at":19,"published_at":20,"article":21,"tags":25,"podcasts":38},"4e71c036-77a6-44af-89e4-872c0c50ce3e","aws-bedrock-sandbox-escape-via-dns-tunneling-and-credential-exposure","7739ed6c-7e76-4a89-8743-d4ea0f6582ce","AWS Bedrock Sandbox Escape Via DNS Tunneling and Credential Exposure","Researchers discovered critical flaws in Amazon Bedrock's AgentCore sandbox that allowed attackers to bypass network isolation through DNS tunneling techniques, enabling data exfiltration from supposedly secure environments. Additionally, a security regression in the microVM Metadata Service removed session token enforcement, creating SSRF vulnerabilities that could expose sensitive AWS credentials. This incident highlights how cloud sandbox technologies can have subtle but critical security gaps that undermine their isolation guarantees. Organizations using cloud AI services must understand that even vendor-managed security controls can have vulnerabilities that require additional defensive layers.","**Immediate actions:**\n- Review and restrict DNS traffic from cloud AI workloads to prevent tunneling attacks\n- Audit current AWS Bedrock configurations and apply AWS-recommended mitigations\n- Implement additional monitoring for unusual network patterns in cloud AI environments\n\n**Long-term improvements:**\n- Establish comprehensive network segmentation policies for cloud AI and ML workloads\n- Develop security requirements for evaluating third-party AI service providers\n- Create incident response procedures specific to cloud service vulnerabilities\n\n**Detection measures:**\n- Deploy DNS monitoring to identify potential tunneling activities\n- Implement anomaly detection for metadata service access patterns\n- Enable comprehensive logging for all cloud AI service interactions",[12,13,14,15,16,17],"CIS Control 12.2","CIS Control 13.1","NIST SC-7","NIST RA-3","NIST SI-4","ISO 27001 A.13.1.3","published","2026-04-13T23:08:39.01946+00:00","2026-04-13T23:08:38.907+00:00",{"id":7,"url":22,"slug":23,"title":24},"https:\u002F\u002Fbit.ly\u002F4t42EpF","cracks-in-the-bedrock-escaping-the-aws-agentcore-sandbox-638f92","Cracks in the Bedrock: Escaping the AWS AgentCore Sandbox",[26,32],{"id":27,"name":28,"slug":29,"description":30,"color":31},"05757c8d-6b93-4194-b35d-7359e7d33b0e","Vulnerability Management","vulnerability-management","Missing scans, no risk prioritization","#fb923c",{"id":33,"name":34,"slug":35,"description":36,"color":37},"f43a7f30-5046-4b10-9dba-1a704139821e","Network Segmentation","network-segmentation","Lateral movement, flat networks, missing firewalls","#06b6d4",[39],{"id":40,"date":41,"edition":42,"title":43,"audio_url":44},"2cf816dd-9e9d-4950-80dd-fdc6088334a3","2026-04-14","morning","ThreatNoir Morning Brief — April 14","https:\u002F\u002Fcdn.threatnoir.com\u002Fpodcasts\u002F2026-04-14\u002Fthreatnoir-morning-brief-2026-04-14.mp3"]