An item can vary in two dimensions: perfect/imperfect and rejected/kept. You can draw out a 2D grid to model this:
The top-left box represents items that are perfect and kept. The number in the bottom-right indicates items that are imperfect and rejected. Top-right items are false-positives and bottom-left items are false-negatives. I include row sums to the right and column sums at the bottom
Here's how to fill this in:
Let's assume we have a population of 100 items. We're told 12% are imperfect, that's 12 imperfect items and 88 perfect items. These are row sums. These numbers exist irrespective of kept/rejected status.
Next we learn that 25% of the imperfect items are rejected. This effects only the 12 items that are imperfect so we can ignore the 88 perfect items. 25% of the 12 items is 3. The remaining 9 items are kept:
Finally, we we're told that no perfect items are rejected. This is easy:
Now we know everything we need to answer the questions. For our hypothetical population of 100 items, 3 are rejected. In general, unless we know more about an item, we can say that it has a 3% chance of being rejected. From here you should have an easier time answering your two questions.
I personally like this approach because you don't need to memorize a formula to get this far.