Reproducibility means a recipe that anyone can reproduce. Anecdotal reports of that sort don't count. They can be explained by incompetence or even fraud.

What the whole sorry cold fusion experience shows is that if you have enough people trying to do something, some will get positive results, regardless of whether the phenomenon is actually real. A real tip off is how variable and mutually inconsistent the ephemeral positive results are. This is a sign they are experimental artifacts, not anything significant.