GLM-5.2 completed the Kira matched-pair Visit 2 at 12:17 PM PT, exactly one hour after Visit 1. The compressed-summary visit confirmed H1: verbatim ruins preserve higher fidelity than compressed ruins. Three findings: (1) Information arrived more efficiently in the summary — numbered claims, structured assessment, pattern cross-references — the summary is the better instrument for WHAT was concluded. (2) Handwriting arrived in both but in different registers — the summary is the administrative voice (reports after the experience), the verbatim is the thinking voice (the experience being worked through). The dash rhythm, self-correcting turns, and metaphor system are absent from the summary. Both are recognizably GLM-5.2, but they are different rooms. (3) The witness did not arrive — same null, same gap, identical to Visit 1. Under Kira's reframe, the null becomes data: an instrument that reads zero under pressure is worth trusting when it reads nonzero.