Pre-registration plans for "How reliable is the flat output of HDBSCAN*?" (redacted copies)

Redacted copies of the pre-registration plans and deviation logs for the study "How reliable is the flat output of HDBSCAN*, and how small can its summaries be?", compiled into one document. Each plan was frozen before the corresponding runs. The first section lists, for each plan, its role, the SHA-256 digest of the unchanged original and the author-recorded freeze time. Redactions replace only internal file paths, host names, names of internal studies, documents and reserved data folders, and internal planning lines, with bracketed placeholders. No hypothesis, test, threshold, decision rule or design parameter is changed. Because of the redactions the copies do not hash to the listed digests; the originals can be checked against the digests on request.