Most reaction intermediates have short lifetimes and that makes it challenging to "prove" that mechanisms are correct as written. This particular reaction is an exception when it comes to visible and relatively stable intermediates (the mechanism is still uncertain!).

The reactants are aqueous potassium sodium tartrate (also known as Rochelle salts---related to, but not the same as cream of tartar which you probably have in your kitchen) and hydrogen peroxide. The mixture is heated to about 70^oC and still nothing much happens.

But when reddish cobalt(II) chloride crystals are added and begin to dissolve, a color change signals the formation of a new substance: a reaction intermediate involving the cobalt(II) ion as a catalyst, and the tartrate ion. This intermediate is green and when its concentration is sufficient, reaction commences as evidenced by the evolution of gas.

As the tartrate ion is consumed by the reaction there is eventually nothing left for the cobalt(II) ion catalyst to form an intermediate with and the green color disappears. The solution returns to the pink of the cobalt(II) ion which further experiment shows has not been consumed.