log(R)=(1-rho_r)*log(Rss)+rho_r*log(R(-1))+rho_rpi*log(pi/pihat)+rho_ry*log(y/yss)-eps_r;

Dear Professor, I have a question. If in the shock module of my random simulation, I set var eps_r; stderr 0.01; what should be the first-period change in interest rate R, and does that mean R decreases by 1% from the steady state?

No, that is only the exogenous change. But your rule includes contemporaneous feedback, which will alter the equilibrium response.

It means that eps_r increases from 0 to 0.01. The change in interest rates also takes into account the changes in inflation and output, but it is uncertain which variable the change in eps_r is added to？

I have a question about Dynare. In Dynare, is it the case that all variables are at their steady state at t=0, and then starting from t=1, IRF (Impulse Response Function) graphs are drawn? For example, is the format in the following picture correct?

At first order, the answer is yes. However, period 0 is not shown in graphs.

Is there any way to control the level of interest rate changes in the first period when such a shock occurs? For instance, I am considering a scenario where the interest rate is reduced by 1% in the first period.

You could write a loop or optimizer step but usually doing trial and error with the shock size is the quickest option.