cred-omega

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). O workflow do skill lê texto livre (código/config/commit/tickets/logs) ao executar varredura local com regex e histórico do repositório (ex.: rg ... . e git log --all --oneline), que pode incluir texto outsider (por exemplo, conteúdo enviado via PR/issue/repo por um usuário não confiável) antes da seleção de itens específicos.

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 system configuration and files (firewall rules, SSH settings, /etc/systemd/system services, ownership of /home/.env, creating /opt layout, etc.) which require elevated privileges and would change the host machine state.