Unable to Apply Operators from a List of Operator Names in `cudaq.register_operation()`

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

I am trying to apply several general unitary matrices to a circuit. I can successfully use cudaq.register_operation(...) by explicitly specifying the operation name. However, I have a use case where I need to store these operator names in a list and retrieve them within a loop. While no error is raised, nothing gets added to the kernel when I attempt this approach. Below is a minimal working example that demonstrates the issue.

Steps to reproduce the bug

Working Example (Using Operator Name Directly):

import cudaq
import math
from cudaq import spin
import numpy as np

N = 2

# Define the observable in spin operator form
observable = 0.5 * spin.z(0) * spin.z(1)

U = np.array([
    [0, 0, 0.92387953, -0.38268343j],
    [0, 0, -0.38268343j, 0.92387953],
    [-0.38268343j, 0.92387953, 0, 0],
    [0.92387953, -0.38268343j, 0, 0]
]) 

cudaq.register_operation("spa", U)
op = ["spa"]

##############################
###   BASE FUNCTIONALITY   ###
##############################
@cudaq.kernel
def cudaq_circuit():
    qubits = cudaq.qvector(N)
    spa(qubits[0], qubits[1])

# Observe the circuit with the specified observable
result = cudaq.observe(cudaq_circuit, observable)
print("Using directly: ", result.expectation())    ### Output :  0.3535533770918846

Expected behavior

Desired Implementation (Using Operator from List):

@cudaq.kernel
def cudaq_circuit():
    qubits = cudaq.qvector(N)
    op[0](qubits[0], qubits[1])

# Observe the circuit with the specified observable
result = cudaq.observe(cudaq_circuit, observable)
print("Trying to use from vector of operator names: ", result.expectation())    ### Output:  0.5

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

• CUDA-Q version:
• Python version:
• C++ compiler:
• Operating system:

Suggestions

No response

Start by running the minimal example around cudaq.register_operation(), cudaq.kernel, and cudaq.observe, comparing the direct operator call with op0. Trace where the operator name is resolved during kernel construction. Done means the list-based call adds the registered operation and produces the expected observation result.