Is tool-making knowledge robust over time and across problems?

Is tool-making knowledge robust over time and across problems?