postgresql-attack

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill includes literal passwords and password-bearing commands (PGPASSWORD env vars, plaintext passwords in connection strings and CREATE ROLE ... PASSWORD) and instructs extracting password hashes from pg_authid/pg_shadow, which forces handling/output of secrets verbatim.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). These URLs point to resources hosted on an attacker-controlled IP (direct payload and shell scripts) referenced in the skill — direct downloads/execution from an untrusted attacker host are high-risk and likely malware distribution vectors.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content is a clear offensive playbook: it provides step-by-step instructions for unauthorized access, credential theft, data exfiltration, RCE, persistence (backdoors), and lateral movement against PostgreSQL systems, indicating deliberate malicious intent and high abuse potential.

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 using PostgreSQL features to execute system commands, write files (webshells, authorized_keys), modify pg_hba.conf and create backdoor SUPERUSER roles — all actions that alter the target host's filesystem and accounts and can directly compromise the machine the agent is running on (especially if TARGET is localhost).