Reproduce a Report

Hand-calculate the supplied count, total and topic map, then run your tool twice. Change one duration in a copy and predict the exact summary change. Compare input digests before and after. Evidence: two unchanged-input runs, one changed-input run, code identity and limitations of the digest. Pass when you can distinguish content identity from data validity and authorship.

Keep predictions before observations. Use the guided lab for commands and contracts; use calibration after your attempt.