Mediation analysis Conflict: Baron & Kenny vs. only indirect effect

Hello I have a question

My independent variable is a dummy coding in which two groups are compared with each other. Expectations (t1-t0) were used as mediator and adherence behavior (t1-t0) as AV.

If I would proceed according to Baron and Kenny, it shows that there is neither a total effect nor a direct effect.

Likewise, there is no significant result from the condition on the mediator from t0 to t1.

However, the result between the mediator and the AV is highly significant.

According to Baron and Kenny, there would be no mediation.

However, bootstrapping with AMOS (BcA; N = 50.000) shows that there is a significant indirect effect.However, the patially standardized regression weight is only 0.09. The result is very close. ((CI95%) =[- ∞ ; -0.005]).

I am puzzled. Some authors write that only the indirect effect must be present for mediation. Others write that only a part of Baron and Kenny and the indirect effect must be significant.

The question arises how I can now answer the hypothesis.

Many greetings