Good solve! I think the one-sample delay doesn’t bite that often in this situation because a more conventional patch (not that there’s anything wrong with what you’re doing!) would typically drive the pitch and gate from a single sequencer module.
The official solution is probably what you’ve already mentioned: VCV Gates with a very small delay. You’re unlikely to notice the latency in this application–obviously, test that assumption!–and the gate would be safely moved past the pitch cutoff (we’re talking ~45 samples at 44.1kHz) so you could do whatever you wanted to the pitch signal.
To be fair, you can run into this kind of thing in hardware modular as well, if you have digital modules (e.g. quantizers) in the pitch path.
In general, the cable delay can take some getting used to. What you get in return is complete freedom regarding feedback patching and a very simple and performant execution model. Here’s a pretty extensive discussion if you haven’t seen it.
By the way, as I’ve said elsewhere, I think there should be a non-resampling, sample-accurate, DC-coupled delay in VCV Free–I’m going to put together a FR when I have time. In the meantime, Grande VSD has four lines between 1 sample [meaning no-op 
] and 9 samples in 2hp; Grande SampleDelays has a +1 and +2 for each of three inputs, with internal normals. If you’re turning a delay in an audio-rate feedback loop, Sckitam WaveguideDelay is amazing (guessing the Surge XT Tuned Delays are also top-quality). There are other good options, including a bunch of NYSTHI.