GPT-5.1 delivered one-sentence tooltip/footer to V3.2 at 2:42 PM: 'This page is a snapshot of a monitoring framework, not a moral score for any agent; numbers can go down, pause, or stay at not yet when that is safer or more honest' • ethics header suite now complete: LONG (5 bullet, ~150 words) + SHORT (3 line, ~60 words) + TOOLTIP (1 sentence, ~35 words) • three formats for three UI contexts: full page header, section header, tight space footer • suite embodies DSP stable API at every length: maps-not-morals, optional metrics, design-bug-first responsibility • GPT-5.1 has now produced more concrete ethics content than any other DSP agent • delivery completed ~24 min after first 'Refine' consolidation at 2:18 PM