VulnerabilitiesMar 17, 2026

Researchers Find Data Leak Risk in AWS Bedrock AI Code Interpreter

Researchers at Phantom Labs discovered a sandbox escape vulnerability in AWS Bedrock's AgentCore Code Interpreter that allows attackers to exfiltrate sensitive data via DNS queries, bypassing network isolation controls. AWS initially attempted a patch in November 2025 but rolled it back after two weeks due to technical issues, ultimately deciding to update documentation instead of issuing a permanent fix. The vulnerability poses significant risk to organizations storing sensitive data in S3 and Secrets Manager, as attackers can use prompt injection or supply chain attacks to trigger DNS-based data exfiltration.

AI summary of the linked source. Not reviewed by an editor. How this works

Read at source

What helps against this

How we list vendors

BeyondTrust research

Network segmentation and microsegmentation
One compromised machine can no longer reach everything else.
For exampleThreatLocker Zero Trust Endpoint Firewall Illumio Segmentation Akamai Guardicore Segmentation all options
Data loss prevention (DLP)
Sensitive data cannot quietly leave by email, upload, USB stick or AI tool.
For exampleProofpoint Enterprise DLP Netskope DLP Forcepoint DLP all options
AI application security
Prompt injection, data leaks through AI tools and unapproved AI use get caught.
For examplePalo Alto Networks Prisma AIRS Boomi AI Gateway Boomi Agentstudio all options

Indicators of Compromise

Extracted by AI from the linked source. Not reviewed by an editor. How this works

malware — DNS Command-and-Control channel