attack-chain

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). This skill explicitly instructs hunting for and using credentials (grep for api_key/secret/password, cloud metadata access) and shows many commands that embed plaintext passwords/NTLM hashes/API keys in CLI invocations and scripts, so an LLM coordinating or emitting those commands would likely need to handle or output secret values verbatim, creating a high exfiltration risk.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). High risk: the list contains direct links to Windows executables/installer/HTA/SCT hosted on an "evil" host (classic malware delivery), plus the AWS metadata endpoint (169.254.169.254) which is often targeted by SSRF to steal cloud credentials.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content contains explicit, actionable offensive playbooks (initial access, webshell/RCE, credential theft with Mimikatz, persistence, C2/evasion, log wiping and data access) clearly intended to enable compromise and unauthorized data exfiltration.

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 explicitly instructs creating user accounts, planting SSH keys, writing/modifying root-owned system files (e.g. /etc/ld.so.preload, /etc/systemd/system), clearing logs, and using sudo-based escalation commands — all actions that alter the host state and require elevated privileges.