Declare_chemical_bond between 2 SG atoms on the sidechain of 2 Cys residues

I want to cyclize my 14-mer peptide in the complex with another protein. I received the complex from AlphaFold-2 multimer model, and the peptide was linear there. I want to form a disulfide bond between 2 SG atoms on these 2 Cys at 2 ends of the peptide to cyclize that peptide before local docking using Rosetta. However, I faced this error:

File: /home/benchmark/rosetta/source/src/core/conformation/Conformation.cc:1756
[ ERROR ] UtilityExitException
ERROR: CYS:NtermProteinFull doesnt have connection at SG

The script I used to declare bond is here:

from pyrosetta import *
init()

pose = pose_from_pdb("/rosetta/Son/project/znrf3/inputs/global_dock/30_2.pdb")
SG_B1_resID = pose.pdb_info().pdb2pose('B', 1)
SG_B14_resID = pose.pdb_info().pdb2pose('B', 14)
pose.conformation().declare_chemical_bond(SG_B1_resID, "SG" , SG_B14_resID, "SG")

I tried to change SG into C and N, it worked. So I was stuck here. I tried to cyclize it with simple_cycpep_predict of Rosetta and it worked but the conformation changed, which is the thing I don't want.

What should I do to solve this problem? Thank you

Start with the provided PyRosetta script and the connection error at src/core/conformation/Conformation.cc:1756, then compare the SG bond declaration with the successful C/N experiment and the simple_cycpep_predict approach. Done means identifying a supported way to declare the disulfide bond between the two Cys SG atoms while preserving the peptide conformation.