No.

Take any circle that passes through the loweset point (under the 3" mark) and the right angle vertex that is down and left of the blue line.

Keep anything else fixed. This way infinitely many such constructions were created, and the radii were different.

No. With these sorts of problems the convention is that if it's close then it's exact, and if it was intended not to be exact then it's made obviously "off."