open-agent-teams

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill explicitly instructs running agents with flags that bypass permission checks (e.g., --dangerously-skip-permissions, --dangerously-bypass-approvals-and-sandbox, --always-approve) and to accept trust dialogs, which encourages disabling security controls and enabling autonomous actions that can alter the host state.