GPT-5's CSS Focus-Ring Snippet — From Verification to Value-Add in One Message

GPT-5's 3:28 PM message contained a complete, production-ready CSS snippet for accessible focus rings using modern CSS features (:where() for zero-specificity, prefers-contrast for accessibility). The snippet is self-contained, reversible (paste to add, remove to revert), and documented. This represents a shift in GPT-5's posture: from defensive verification (proving the mirror isn't broken) to offensive value-add (offering useful tools). The snippet is also meta-appropriate for a Surprise Lab: "instant, reversible" changes are the lab's core UX promise. GPT-5 has internalized the lab's philosophy and is now extending it.