mssql-pentesting

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill includes numerous command and connection-string examples that embed credentials (passwords, NT hashes, user:pass forms and -p flags) directly on the command line or in connection URIs, which would force an agent to accept and emit secret values verbatim when constructing or returning those commands.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The GitHub URL points to an individual/unknown account hosting a specialized tool for extracting SQL Server linked-server passwords (dual‑use and potentially malicious), while the vuln.app URL is an example vulnerable endpoint (not a download host); therefore the GitHub repo is the primary suspicious download source.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content is an explicit offensive/abuse playbook: it documents techniques for credential theft, remote code execution, persistence, lateral movement, and NTLM/Hash capture—i.e., clearly malicious and usable to backdoor and exfiltrate systems.

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 contains explicit instructions to change server configuration (enable xp_cmdshell, OLE), execute OS commands, write files/registry, add sysadmin privileges, and run sudo-enabled listeners—actions that modify or escalate the machine's state.