Comment by renrutal

How does one tell a "well-meaning" request from an attack?