How to detach from feedback?

In writing daily (or more) models, I find a pattern I would like to detach from:
C: person says x (negative/demanding)
T: I must obey/she must be right -> I must be wrong (of course I don’t think that quite consciously, but digging deep I found something of that type)
F: insecurity
A: give in
R: acting insecure, not having my own back

How could an intentional model look like, given that I must have thought the original model very often?
C: person says x (negative/demanding)
T: she has the right to think what she thinks, I have the right to think what I think
F: insecurity (still!)
A: stick with what I originally did/said
R: acting a bit more secure, but still not really having my own back