cloud-attack-methods

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill includes explicit examples of credentials and instructs sending/reading tokens (e.g., POST bodies with accessKey/secretKey and metadata endpoints that return tokens), which requires the LLM to handle or reproduce secret values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document is an explicit offensive/cloud-attack playbook describing token/credential theft, metadata abuse, remote code execution, service enumeration, and techniques to exfiltrate data and deploy backdoors—highly malicious intent.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The prompt explicitly describes techniques to access local cloud metadata endpoints, steal tokens, exploit kubelet/etcd and privileged pod host-escape paths and other privilege-escalation methods that would enable an agent to obtain credentials and compromise the machine or its cloud identity.