You should be fine stacking on a whole set of extension tubes. You will increase diffraction, however you'll also be magnifying your subject by a greater factor, possible several times more...so fine details will still stand out more than they would at a lower magnification level because the effects of diffraction remain smaller than the magnified details (up to a certain point...diffraction will grow faster than detail magnification, however long before it gets to the point where the airy disc is larger than your original details, other things will limit your ability to keep extending.) The facets of an insects eye become gigantic, and the fine details of EACH FACET could be visible with enough magnification, to the point where they span large clusters of pixels...where the airy disc of diffraction may only span a couple pixels. Extension tubes do not add any optical elements to the light path, so technically speaking, you should be able to extend and gain additional magnification without significantly affecting IQ.

For experiments sake, lets say that hypothetical insect actually is our subject. Lets say we are shooting with an 18mp APS-C camera, at 1:1 magnification. Lets say the facets of our subjects eyes span 8x8 pixel areas (very small!)

If you are shooting 35mm 1:1 @ f/5.6, and slap on a 25mm extension tube. Magnification gain is extension/focalLength, so your adding 25mm/35mm, or 0.714x more magnification. Magnification affects the effective f-stop that you are shooting at. At 1.0x magnification, you are already experiencing some of the effects, and your effective aperture is f/11. With the additional magnification, your effective f-stop is f/5.6 * (1 + 1.714), or f/15. Your subjects eye facets are now about 26x26 pixels in size, and diffraction is affecting about 4 pixel areas.

Similarly, 50mm of extension would be 1.43x additional magnification (50/35), so effective f-stop is f/5.6 * (1 + 2.43), or f/19. Diffraction at that level is visible and will have a moderate impact on IQ, but not anywhere close to as bad as optical aberrations are going to be at f/2.8. It still isn't affecting the ultimate quality of your image, however...as your subject has also grown in detail. Your subject's eye facets are now about 43x43 pixels, and diffraction is affecting about 6 pixel areas.

Lets take the experiment farther...you have to stop down to f/22 to get enough DOF, and your extending by a whole 5x magnification. That gives you an effective aperture of f/22 * (1 + 5), or f/132. At this point, the effects of diffraction would span about a 150 pixel area for an 18mp APS-C sensor (which is VERY high resolution, about 116 lp/mm...line pairs/millimeter.) You might be inclined to think the effects of diffraction are now obliterating all the detail you worked so hard to get. That wouldn't necessarily be the case, though. Your at 5x magnification, almost three orders of magnitude greater than you were at 2.43x magnification before, where those fine details spanned 26x26 pixel areas. The same details should be spanning more than 250x250 pixel areas now. Diffraction has grown, and will likely blur out fine details, but is affecting about 50 pixel areas. You'll still be extracting more detail than you lose to diffraction.

To answer your fundamental question: How much can you magnify before you lose detail? The size of the airy disc will grow slightly faster than the size of the original detail at 1.0x magnification. This is due to the non-uniform nature of diffraction, and the way it will interfere with/amplify itself as its effect grow. Diffraction is also dependent on the wavelength of light...so while I have used the wavelength of yellow-green light (564nm) for my calculations so far, visible light spans the range from about 340nm violet to 790nm deep red. Deep red light will diffract more than other wavelengths, and will produce greater diffraction. You may eventually reach a limit, wherein diffraction affects IQ enough that you don't gain any further benefits. That limit is very far beyond the point where other mechanical limitations prevent you from extending any more.

In normal photography, the more you stop the aperture down, the more the effects of diffraction affect the image. Since the detail in the image is not getting larger as you stop down, the more detail you lose as airy discs grow. When it comes to macro photography, your magnifying the detail as you increase extension...and while your also increasing diffraction, the original details are always larger than the airy disc. You WILL lose some detail as you extend (you'll be bringing finer and finer detail to light, and beyond around 3x magnification, diffraction will start to affect the visibility of finer details than what you started out with at 1.0x.) Eventually the effects of diffraction will prevent you from continuing to make useful gains with additional magnification. But you can push magnification very far. In the general case, you are far more likely to run into the problem where your focal plane ends up too close or actually inside the lens before you actually run into problems with diffraction affecting IQ in a truly detrimental way.