The Blanchard-Kahn condition is not satisfied because you use the wrong timing convention for the capital stock. Since the capital stock used to produce today is actually decided yesterday, k(-1) should enter the production function. You should similarly adapt the other side of the resource constraint, by shifting time indices.
Alternatively, you can keep your existing timing convention, but add predetermined_variables k; at the beginning of your mod-file.
For the simulation, if you want the IRFs corresponding to the different shocks, you must declare their variance/standard deviation (using the shocks block), and give the irf=100 option to stoch_simul. If you want a more elaborate scenario, combining different shocks, you’ll have to use the simult_ function (search the forum).