Thank you for giving more information, but I still have to resort to speculation. Here are some things you might want to consider:
- A rear-mounted prop stabilizes the aircraft. Full throttle limits your control power. You would had a better chance of recovery with engine idle.
- No incidence means that the plane needs to compensate for the c.g. location with elevator deflection alone. If your plane is statically stable, it would need to be trimmed with some elevator-down deflection for straight flight. This limits the maximum possible canard lift and your nose-up control power. Better give the canard a few degrees of positive incidence.
However, from the picture it might well be that you use a full-flying canard. In that case, the incidence is set automatically by trimming, and a nose-up command means increasing the incidence further. Now the control commands do not change the camber of the canard wing which reduces both the amounts of maximum and minimum lift possible. Better use a canard with a stabilizer-elevator combination to increase control power.
You talk of cutting and sanding. Are the wings and canard made from styrofoam? Then @TipStall is right: The plane lacks stiffness. Commanding a descent (what you described as "going down") increased the dynamic pressure and thus the forces warping wings and canard.
For now, I think you suffered a canard wing stall, but I am not able to say why. Either it was aeroelasticity (twisting of wing and canard) or insufficient control power. In that case, use a stabilizer-elevator combination for the canard, throttle your engine and maybe consider a less stable aircraft by shifting the c.g. back a bit.