Regex for unicode invisible control characters does not match all cases

While using the regex `\p{Cf}`, which is expected to match invisible formatting indicators in Unicode, we found some inconsistent behavior. Working through this list, for example: https://www.compart.com/en/unicode/category/Cf

| Character | Type | Result |
|--|--|--|
| `\00AD` | Soft Hyphen | Match |
| `\0602` | Arabic Footnote Marker | No match |
| `\202A` | Left to Right Embedding | No match |
| `\200B` | Zero-width Space | No match |
| `\2064` | Invisible Plus | No match |

I can see that `\p{Cf}` is supported, but are there any rules governing which matches are valid? Is this expected behavior?