Some research tasks are skill-shaped, most aren't. By the end of this station you'll be able to tell them apart in ten seconds and stop wasting your time on the ones that aren't ready.
A noun phrase. An XRD pattern. A list of papers. A reviewer comment + manuscript section.
A single transformation. Identify. Classify. Fit. Not three at once.
Something you could schema. A phase + confidence. A gap list. A classified comment.
If you can't write the input as a noun phrase, the verb as a single word, and the output as something you could draw on a piece of paper, the task isn't skill-shaped yet. That's fine — it just means it needs more thought before you wrap it in a skill.
Given a powder XRD pattern and a shortlist of plausible phases, identify the dominant crystal phase with a confidence score and flag any peaks that don't match.
No input. No defined transformation. No checkable output. A great conversation, but not a skill.
Not a grand project. The 5-minute thing you do 50 times a year. The XRD pattern you eyeball. The literature gap analysis you keep half-finishing. The reviewer comment you avoid drafting a response to. Hold it in your head. We'll come back to it at every station.