Turn a run that came out well into a skill on the spot

A how-to that builds a research agent which runs on its own at a set time each week, using nothing but a scheduling feature. It likens building one to hiring a person and splits it into four stages — deciding what to ask for, building it, teaching it how to work, and opening up the material — then points out that most people skip the first stage and pile up nothing but useless results. It advises picking the first task where saving time and making money overlap, because work that only saves time shows no visible result and the support for it dries up. The most striking part is that writing one line under a run that came out well, asking for it to be made into a skill, registers that procedure as a reusable manual with a name and instructions. It is not perfect, so a few weeks are needed in which a person looks at the results and adds each wrong case to the instructions as a one-line exception.

This is a note: my own write-up after reading several sources. It has no source link; events checked against a source are kept as records.