Confusion Regarding Semidefinite Constraints

[JoshPilipovsky]

Hello, I am trying to run MOSEK using the following code (with inputs: X_0T: [n, T], X_1T: [n, T], U_0T: [m, T], Qx >= 0, R > 0, N, muf, Sigmaf >= 0 of appropriate dimension):

% Get dimensions
n = size(X_0T, 1);
m = size(U_0T, 1);
T = size(X_0T, 2);

% Define optimization variables
H     = sdpvar(T, N + 1);
Sigma = sdpvar(n, n, N + 1);
Pi    = sdpvar(T, n, N + 1);
Z     = sdpvar(m, m, N);

% Define cost function
meanCost = 0;
covCost  = 0;
Qtilde = X_0T' * Qx * X_0T;
Rtilde = U_0T' * R * U_0T;
Stilde = Qtilde + Rtilde;
for k = 1 : N + 1
    meanCost = meanCost + H(:, k)' * Stilde * H(:, k);
end
for k = 1 : N
    covCost = covCost + trace(Qx * Sigma(:, :, k)) + trace(Z(:, :, k));
end
objective = meanCost + covCost;

% Define constraints
Constraints = [];

% Initial condition
Constraints = [Constraints, Sigma(:, :, 1) >= eye(n)];

% Terminal conditions
terminalMeanConstraint = X_0T * H(:, N + 1) == muf;
terminalCovConstraint  = -Sigma(:, :, N + 1) + Sigmaf >= 0;
Constraints = [Constraints, terminalMeanConstraint];
Constraints = [Constraints, terminalCovConstraint];

% Covariance dynamics constraints
for k = 1 : N
    cov_dynamics_k = [Sigma(:, :, k + 1) - eye(n)  X_1T * Pi(:, :, k);
                      Pi(:, :, k)' * X_1T'         Sigma(:, :, k)    ] >= 0;
    Constraints = [Constraints, cov_dynamics_k];
end

% Mean dynamics constraints
for k = 1 : N
    mean_dynamics_k = X_0T * H(:, k + 1) == X_1T * H(:, k);
    Constraints = [Constraints, mean_dynamics_k];
end

% Cost condition
for k = 1 : N
    
    LMI_cost_k = [Z(:, :, k)                       sqrtm(R) * U_0T * Pi(:, :, k);
                  Pi(:, :, k)' * U_0T' * sqrtm(R)  Sigma(:, :,k)                ] >= 0;
    Constraints = [Constraints, LMI_cost_k];
    
end

% Equality constraints
for k = 1 : N + 1
    Constraints = [Constraints, Sigma(:, :, k) == X_0T * Pi(:, :, k)];
end

% Solve the Problem
options = sdpsettings('solver','mosek');
sol = optimize(Constraints, objective, options);

I am running into an issue where the optimizer says 'Solver not applicable (mosek does not support semidefinite constraints)'. I am a bit confused by this because all of my inequality constraints are written as LMI's, which MOSEK can handle (I have a similar optimization problem that is dual and feasible, and many of the constraints are exactly the same). So, my question is in particular which constraints that I have in this optimization problem are causing problems for MOSEK? Thank you!