OK, the way you describe the deck being vented eliminates the "airbubble under the deck" idea.
Then, the auto-dive phenomenon still sounds like a change in vessel weight and C/G, characteristic of free-flooding hulls with a WTC inside.
I re-submit the possibility that, while the boat is underway at full speed, water is entering the hull through the forward deck limbering holes and perhaps the forward ballast vents, too.
Under ram pressure, this additional water ballast is rising upwards inside the hull, increasing weight, and causing the boat to settle to a deeper waterline.
When the additional water ballast starts filling the deck area, we have a forward shift in C/G because the naviform foredeck and wheelhouse have greater volume than the aft deck does.
This forward shift in C/G engenders a slight nose-down attitude (wouldn't take much at full speed, and might not be readily perceptible from shore) which imparts a negative planing attitude to the straight portions of the side fins and upper surfaces of the hull, initiating a dynamic dive. (Note: this is not a function of "negative lift" produced by Bernouli's reduced pressure effect; it is resultant of increased impact pressure above the planing surfaces.)
As this slight diving trend continues, the top surface of the deck goes awash, and it, too, becomes a negative planing surface, driving the boat down still further.
Without stern planes to counter the effect, the trend continues to produce what, for lack of a better word at the moment, I'll call a "crash dive".
The stock quad planes are reminiscent of what was state-of-the-art design theory for the 1860's; but there are many drawbacks: potential for "slop" in the linkages to produce unequal control forces; and the short / unequal moment-arm distance of the fore / aft plane pairs from the C/G minimalizing / unbalancing control surface efficiency, to name two of the best reasons why designers eventually switched to stern planes.
As you've mentioned, the auto-dive phenomenon is controllable with the use of stern planes; and I'll add that, while your arrangement with an additional plane aft of the prop is most efficient, hinging the trailing section of the side fins themselves to turn them into a pair of stern planes, will also work pretty well.
As an alternative for those who don't want stern planes: a mechanized, R/C actuated, "sliding weight" device that enables the C/G to be adjusted slightly fore / aft while the boat is underway, could compensate for the dynamic variation in C/G, too; but it would add weight, complexity, and be more difficult to control. I've never liked such devices, but some people do use them. Could work in this case, so I just thought I'd throw it out for others to consider.
VULCANIA SUBMARINE: Home of the Nautilus Minisub and 20,000 Leagues Diving Suits