Re: a), the best thing is to document the protocol, then let someone else (who has not seen the insides of the original application) implement it based on the documentation.
It doesn't sound like the OP has seen the insides of the original application either, sounds like it was just the protocol that was reverse engineered, from the packets.
In the US at least, clean room reverse engineering is not required (see the "case law" section of the article, which documents two cases where clean-room separation was not maintained and the reverse engineers still prevailed) nor does it necessarily protect you (all the examples in the article predate the DMCA, which puts significant restrictions on reverse-engineering itself, and as the article points out, the clean-room approach is it relevant for patents).
