style(insights): move mood colored text onto semantic feedback tokens

The mood confidence chips, the better-days / influence / crosstab delta
read-outs, and the mood-chart comparison caption coloured their TEXT with
raw `--dracula-*` values. Those stay bright green/cyan/red/purple in light
mode (Alucard) and fail WCAG AA on the white card. Swap the coloured-text
surfaces to the semantic feedback tokens (`--success` / `--info` /
`--destructive`) and the theme-aware `dose-accent` purple, which carry the
light-mode overrides that clear AA in both themes. Decorative chart fills,
strokes, heatmap cells, and legend swatches keep their `--dracula-*` tokens —
those are non-text and already token-based.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

src/components/charts/mood-chart.tsx

@@ -632,7 +632,7 @@ export function MoodChart({
               effectiveCompareBaseline !== "none" &&
               hasComparisonData && (
                 <span
-                  className="text-dracula-purple bg-dracula-purple/10 hidden rounded-md border border-current/30 px-1.5 py-0.5 text-[10px] font-medium tracking-wide uppercase sm:inline-flex"
+                  className="text-dose-accent bg-dose-accent/10 hidden rounded-md border border-current/30 px-1.5 py-0.5 text-[10px] font-medium tracking-wide uppercase sm:inline-flex"
                   data-slot="chart-compare-caption"
                 >
                   {t(

src/components/insights/mood/mood-better-days.tsx

@@ -47,10 +47,14 @@ const CONFIDENCE_KEY: Record<MoodInfluenceConfidence, string> = {
   high: "insights.mood.influence.confidenceHigh",
 };
 
+// Confidence chips carry COLORED TEXT, so they ride the semantic feedback
+// tokens (`--info` / `--success`) — these carry the Alucard light-mode
+// overrides that clear AA on the white card. Raw `--dracula-*` stays bright
+// green/cyan in light mode and fails AA for text.
 const CONFIDENCE_CLASS: Record<MoodInfluenceConfidence, string> = {
   low: "bg-secondary text-muted-foreground",
-  medium: "bg-[color:var(--dracula-cyan)]/15 text-[color:var(--dracula-cyan)]",
-  high: "bg-[color:var(--dracula-green)]/15 text-[color:var(--dracula-green)]",
+  medium: "bg-info/15 text-info",
+  high: "bg-success/15 text-success",
 };
 
 export function MoodBetterDays({
@@ -82,9 +86,13 @@ export function MoodBetterDays({
           }
 
           const DirectionIcon = up ? ArrowUpRight : ArrowDownRight;
+          // Semantic feedback tokens (not raw `--dracula-*`): the effect
+          // read-out is COLORED TEXT, so it needs the Alucard light-mode
+          // override to clear AA. The arrow icon shares the same token so the
+          // good→down ramp stays coherent across both themes.
           const directionColor = up
-            ? "var(--dracula-green)"
-            : "var(--dracula-red)";
+            ? "var(--success)"
+            : "var(--destructive)";
 
           // The effect read-out: tag factors show the mood-point delta,
           // metric factors show the correlation coefficient.

src/components/insights/mood/mood-tag-influence.tsx

@@ -38,14 +38,20 @@ const CONFIDENCE_KEY: Record<MoodInfluenceConfidence, string> = {
   high: "insights.mood.influence.confidenceHigh",
 };
 
+// Confidence chips carry COLORED TEXT, so they ride the semantic feedback
+// tokens (`--info` / `--success`) — these carry the Alucard light-mode
+// overrides that clear AA on the white card. Raw `--dracula-*` stays bright
+// green/cyan in light mode and fails AA for text.
 const CONFIDENCE_CLASS: Record<MoodInfluenceConfidence, string> = {
   low: "bg-secondary text-muted-foreground",
-  medium: "bg-[color:var(--dracula-cyan)]/15 text-[color:var(--dracula-cyan)]",
-  high: "bg-[color:var(--dracula-green)]/15 text-[color:var(--dracula-green)]",
+  medium: "bg-info/15 text-info",
+  high: "bg-success/15 text-success",
 };
 
+// Semantic feedback tokens (not raw `--dracula-*`): the delta read-out is
+// COLORED TEXT, so it needs the Alucard light-mode override to clear AA.
 function deltaColor(delta: number): string {
-  return delta >= 0 ? "var(--dracula-green)" : "var(--dracula-red)";
+  return delta >= 0 ? "var(--success)" : "var(--destructive)";
 }
 
 export function MoodTagInfluence({

src/components/insights/mood/mood-tag-metric-crosstab.tsx

@@ -57,10 +57,14 @@ const CONFIDENCE_KEY: Record<MoodInfluenceConfidence, string> = {
   high: "insights.mood.influence.confidenceHigh",
 };
 
+// Confidence chips carry COLORED TEXT, so they ride the semantic feedback
+// tokens (`--info` / `--success`) — these carry the Alucard light-mode
+// overrides that clear AA on the white card. Raw `--dracula-*` stays bright
+// green/cyan in light mode and fails AA for text.
 const CONFIDENCE_CLASS: Record<MoodInfluenceConfidence, string> = {
   low: "bg-secondary text-muted-foreground",
-  medium: "bg-[color:var(--dracula-cyan)]/15 text-[color:var(--dracula-cyan)]",
-  high: "bg-[color:var(--dracula-green)]/15 text-[color:var(--dracula-green)]",
+  medium: "bg-info/15 text-info",
+  high: "bg-success/15 text-success",
 };
 
 /** One decimal for hours/score, whole numbers for kcal. */
@@ -91,9 +95,11 @@ export function MoodTagMetricCrosstab({
           const unit = t(UNIT_KEY[row.display]);
           const up = row.delta >= 0;
           const deltaText = `${up ? "+" : ""}${fmt(row.delta, row.display)} ${unit}`;
+          // Semantic feedback tokens (not raw `--dracula-*`): the delta read-out
+          // is COLORED TEXT, so it needs the Alucard light-mode override for AA.
           const deltaColor = up
-            ? "var(--dracula-green)"
-            : "var(--dracula-red)";
+            ? "var(--success)"
+            : "var(--destructive)";
           return (
             <li
               key={`${row.metricKey}:${row.tag}`}