Read the explanation
The local source contains eighteen catalog entries. Entering DseWiki in the native search returns two rows. At fifteen pixels per displayed row the bars shrink from two hundred seventy to thirty. The search checks name, operator, communication-vector prose, and platform type, so a match need not occur only in the title. This is filtering a hand-authored catalog, not live discovery or verification. Its claims about named sites, investigators, telemetry, or agent misconduct are uncorroborated here. No website is contacted and no allegation becomes true because a row matches a query. After the search reduces the visible table to two rows, the export still includes all eighteen entries under site catalog. It separately records the current query, filter values, filtered count, and selected dossier. At fifteen pixels per entry the visible-count bar spans thirty while the exported-catalog bar spans two hundred seventy. Filtering therefore changes presentation and context, not the size of the underlying exported catalog. The export also writes a fixed telemetry-confirmed field; that authored Boolean is not evidence of a checked log or independently verified event. Review provenance before using this JSON as a forensic report. The sandbox chooses among four prepared communication-vector branches and inserts their descriptions into a terminal panel. At seventy pixels per branch the first bar spans two hundred eighty. Actual network actions are zero, so the second bar remains zero. The terminal message says a breadcrumb is confirmed, but clicking the button only renders a string and a local timestamp. It does not send the pictured URL, write a remote message, test a vulnerability, or observe an agent bypass. Treat it as an illustration of claimed mechanisms, and keep that distinction when reading or exporting the dossier.