sqlserver-attack

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt repeatedly embeds plaintext credentials (e.g., -P 'password', sa:'password', explicit default passwords and a literal CREATE LOGIN ... WITH PASSWORD) into commands, forcing the LLM to produce secret values verbatim in outputs.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). These are direct attacker-hosted links (including .exe and .ps1) used in the skill's payload/download examples and are high-risk indicators of malware distribution and remote code execution.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content is an explicit offensive playbook describing actionable techniques for unauthorized access, remote code execution, credential theft, lateral movement, persistence, and data exfiltration—clear malicious intent and abuse guidance.

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 enabling privileged features (xp_cmdshell, OLE Automation), executing OS commands, creating persistent SQL Agent jobs, performing lateral EXEC AT commands, reading/writing system files, and creating a new sysadmin login — all actions that modify target system state and create accounts.