Pardon the bastardization via my "armchair doctorate in philosophy", but your "tools and rules" model sounds like a classic "subjective/objective" model, where there are actors and there are things being acted upon.

While this is a useful model, it's just a model. And just like Newtonian Physics is a great model for predicting and altering most behavior at non-relativistic speeds and energies, the model breaks down when we start looking at the "extreme" situations, and so we have to upgrade to a larger more inclusive (albeit far clumsier) model: relativistic mechanics, quantum mechanics, and so on.

In particular, subjective/objective as a model keeps you from seeing that every observation you make as a subject also influences the object to a tiny degree. In Physics, this is parallel to the Heisenberg Uncertainty Principle. So you can't completely define a ruleset and a toolset. You can only define a toolset and ruleset that mostly work for any given system.

Keep your mind open to larger models as well, although again, you may not necessarily use them for everyday things.