fix(rules): handle multi-byte emoji in getLastChar via Array.from

getLastChar used t[t.length-1] which only gets the trailing surrogate
of emoji like 🔚 (U+1F51A, a surrogate pair in UTF-16). This meant
end symbol matching ALWAYS failed for emoji symbols, causing every
non-humanList message to hit rule_match_no_end_symbol -> no-reply.

Fix: use Array.from(t) to correctly split by Unicode code points.

plugin/rules.ts

@@ -43,7 +43,10 @@ function stripTrailingMetadata(input: string): string {
 
 function getLastChar(input: string): string {
   const t = stripTrailingMetadata(input).trim();
-  return t.length ? t[t.length - 1] : "";
+  if (!t.length) return "";
+  // Use Array.from to handle multi-byte characters (emoji, surrogate pairs)
+  const chars = Array.from(t);
+  return chars[chars.length - 1] || "";
 }
 
 export function resolvePolicy(config: WhisperGateConfig, channelId?: string, channelPolicies?: Record<string, ChannelPolicy>) {