specialized-attack-playbooks

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs constructing requests and extracting/including secrets (AK/SK, X-Cloud-Access-Token, base64 PEM keyData, MinIO accessKey/secretKey, visitor tokens) in request bodies/headers and outputs, which requires the LLM to handle and emit 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 playbook: it contains step-by-step instructions and tools for stealing private keys, exfiltrating credentials, performing ARP MITM, abusing S3/STS to achieve remote code injection, abusing MinIO/OCS uploads, and related persistence/automation — clearly intentional malicious behavior.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill includes explicit instructions to modify kernel/network settings (echo to /proc/sys/net/ipv4/ip_forward, ip neigh replace), run packet-capture/ARP spoofing tools, and create persistent crontab/rc.local entries — all actions that require root and change the host system state.