fix(telegram): keep callback_data under Telegram's 64-byte cap

``InlineKeyboardButton(label, callback_data=label)`` fails Telegram's
API when the label exceeds 64 bytes UTF-8. An LLM-generated long
option (realistic in multilingual flows) used to 400 the ``send_message``
call silently — user got nothing, agent heard a successful retry-then-drop.

Decouple display from wire: button text keeps the full label, callback_data
gets truncated at a UTF-8 char boundary. Tap echoes the prefix back as the
user message; the LLM understands a prefix of its own option just fine,
and the display the user saw was always the full string.

Locks: helper boundary behavior (ASCII, CJK, short labels pass through)
and end-to-end ``_build_keyboard`` integration with an over-cap label.

Made-with: Cursor

nanobot/channels/telegram.py

@@ -1205,11 +1205,19 @@ class TelegramChannel(BaseChannel):
         if not buttons or not self.config.inline_keyboards:
             return None
         keyboard = [
-            [InlineKeyboardButton(label, callback_data=label) for label in row]
+            [InlineKeyboardButton(label, callback_data=self._safe_callback_data(label)) for label in row]
             for row in buttons
         ]
         return InlineKeyboardMarkup(keyboard)
 
+    @staticmethod
+    def _safe_callback_data(label: str) -> str:
+        # Telegram caps callback_data at 64 bytes UTF-8; truncate at a char boundary so the keyboard still sends.
+        encoded = label.encode("utf-8")
+        if len(encoded) <= 64:
+            return label
+        return encoded[:64].decode("utf-8", errors="ignore")
+
     @staticmethod
     def _buttons_as_text(buttons: list[list[str]]) -> str:
         # Buttons are semantic options; when we can't render a keyboard, the user still needs to see them.