I ran a QST3 calculation under an expanded basis set:
S 0
6-311+G(2df,p)
****
H C N O 0
6-311+G(d,p)
****
(this is the same expanded basis set I used for optimizations)
When the calculation finished, the output provided a geometry that was nearly identical to my product but it contained 1 imaginary frequency at -255. I don't believe this to be the transition state. How can I refine the calculation to get a better result?