For the record, I opposed the HK agreement as did many other protocol developers. I was not there myself. I had planned on attending but had to cancel last minute.
Nonetheless, several of the developers in attendance did deliver on the hard fork code: https://bitcoinhardforkresearch.github.io/
We had a subsequent meeting in the summer in Palo Alto which I did attend. Bitmain specifically agreed they would signal SegWit on their pools using BIP9, a procedure everyone was well familiar with since the CHECKSEQUENCEVERIFY soft fork had already been activated this way recently.
Instead, Bitmain began insisting upon a consensus-incompatible fork of Bitcoin Core soon after this meeting and broke off communications. In addition, they started pushing the old HK agreement narrative in China.
I’d be happy to provide additional corroborating information upon request.