A.5-polish-2 FLIP-to-back bug fixes — TDD. Two user-reported bugs from the first FLIP-to-back commit (1963ad4): (1) FLIPping made the entire card disappear instead of showing the back; (2) the FLIP btn stayed visible during the animation when it should hide just like the tarot-fan view's flip btn does. Root cause of (1): the back-image element was rendered with inline style="display:none" in the template. Inline styles beat CSS class rules in the cascade — my .sig-stage-card.is-flipped-to-back .sig-stage-card-back-img { display: block } rule was the right specificity but couldn't override an inline style attribute. So the toggle hid the front (CSS-controlled, no inline override) but failed to show the back (CSS blocked by inline). Net: empty stage. Fix: removed the inline style on the back-img element; default .sig-stage-card-back-img { display: none } rule (already present in SCSS) handles the hidden default, and the .is-flipped-to-back toggle now flips visibility cleanly. Both rules are pure-CSS so they cascade as expected. Root cause of (2): the non-polarized FLIP handler was a bare class toggle (no animation, no data-flipping attr), so there was no SCSS hook to hide the btn. Plus there was no equivalent SCSS rule even for the polarized _flipPolarityAnimated flow which DID set data-flipping — the polarized flip just animated without hiding the btn either. Fix: (a) added _flipToBackAnimated() JS function mirroring _flipPolarityAnimated's shape — rotateY 0→90→0 at 500ms ease, swap visual content at the halfway point (here: class toggle instead of revInput/polarity flip), set stageCard.dataset.flipping = '1' for the duration so SCSS has a hook. (b) New SCSS rule .my-sign-stage:has(.sig-stage-card[data-flipping]) .my-sign-flip-btn { opacity: 0; pointer-events: none } mirrors the tarot-fan view's pattern (_card-deck.scss:459 — .tarot-fan-wrap:has(.fan-card[data-flipping]) .fan-flip-btn). The :has() selector covers BOTH the polarized animation (which already sets data-flipping) AND the new non-polarized animation, so the btn hide-during-flip behavior now lands consistently across both flip modes — fixes a latent polished-flow gap not just the new code path. No new tests — the existing 3 ITs from 1963ad4 already verify the template/scaffold contract (data-deck-polarized attr + back-img element conditional render); the bug fixes here are CSS/JS-level behavior best caught by visual verify (no automated test would have caught the inline-style cascade issue since the IT asserted on element presence, not display state). 1303/1303 IT+UT total green (71s, unchanged from 1963ad4 since no new tests in this commit)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/static_src/scss/_card-deck.scss

@@ -887,6 +887,18 @@ html:has(.sig-backdrop) {
     display: inline-flex;
 }
 
+// Sprint A.5 — hide FLIP btn during the flip animation. `data-flipping="1"`
+// is set on .sig-stage-card by _flipPolarityAnimated (polarized) AND
+// _flipToBackAnimated (non-polarized) for the 500ms animation duration; CSS
+// :has() selects the parent .my-sign-stage when any child carries that attr
+// and zeros the btn so it doesn't visually interfere w. the rotateY mid-spin.
+// Mirrors the tarot-fan view's pattern (`_card-deck.scss:459` —
+// `.tarot-fan-wrap:has(.fan-card[data-flipping]) .fan-flip-btn`).
+.my-sign-stage:has(.sig-stage-card[data-flipping]) .my-sign-flip-btn {
+    opacity: 0;
+    pointer-events: none;
+}
+
 // ─── Mini card grid ───────────────────────────────────────────────────────────
 // flex: 0 0 auto — shrinks to card content; no background (backdrop blur).
 // align-content: start prevents CSS grid from distributing extra height between rows.