another currently unanswered question in Muscle Mem is how to more cleanly express the targeting of named entities.

Currently, a user would have to explicitly have their @engine.tool call take an element ID as an argument to a click_element_by_name(id) in order for it to be reused. This works, but for Muscle Mem would lead to the codebase getting littered with hyper-specific functions that are there just to differentiate tools for Muscle Mem, which goes against the agent agnostic thesis of the project.

Still figuring out how to do this.