The Shai Hulud supply chain worm, attributed to TeamPCP, compromises CI/CD pipelines by injecting malicious npm/PyPI packages that harvest build credentials and pivot into production AWS cloud infrastructure. In a confirmed breach, attackers stole Jenkins EC2 instance role credentials via the Instance Metadata Service (IMDS), used them from external IPs, escalated privileges by creating an IAM user with AdministratorAccess, modified Redshift and Aurora security groups to open network paths, enumerated Secrets Manager for warehouse credentials, and exfiltrated data via the Redshift Data API. The attack demonstrates that pipeline identity equals production identity, with explicit attacker naming conventions (exfil-s3-* policies, exfil STS session names) providing high-fidelity detection opportunities.
IAM
4 posts
From CI/CD to Cloud Data: How Shai Hulud Persistence Leads to Redshift Breach 19 Cloud Security Challenges and How to Mitigate Risk The article outlines 19 critical cloud security challenges facing organizations, emphasizing that misconfigurations, weak identity and access management (IAM), and human error are the primary drivers of cloud compromise. It highlights emerging threats such as AI-powered deepfake social engineering, MFA fatigue, and cloud-targeted extortion, underscoring the need for unified visibility and robust configuration management.
Emerging Enterprise Security Risks of AI The rapid adoption of agentic AI in enterprise environments introduces significant security risks by amplifying existing software supply chain and identity management vulnerabilities. Threat actors can leverage prompt engineering, input manipulation, and malicious packages to weaponize AI agents, necessitating zero-trust principles, robust IAM for non-human identities, and human-in-the-loop safeguards.
Cracks in the Bedrock: Agent God Mode The AWS Bedrock AgentCore starter toolkit automatically provisions overly permissive IAM roles that grant wildcard access across the AWS account. This "Agent God Mode" misconfiguration allows a compromised AI agent to exfiltrate ECR images, access other agents' memories, and escalate privileges by invoking other code interpreters or agents.
