From b65b3c6ff015582d6ef2aba721d969b8d87debe6 Mon Sep 17 00:00:00 2001
From: Limitless <127183162+Limitless2023@users.noreply.github.com>
Date: Sun, 15 Feb 2026 16:49:10 +0800
Subject: [PATCH] fix(telegram): include voice transcript in body text instead
 of raw audio (#16789)

- Move hasAudio detection before bodyText building
- Move preflight transcription before bodyText building
- If audio has transcript, use transcript as bodyText
- Otherwise use <media:audio> placeholder

Fixes #16772: Telegram voice messages leak raw audio binary into chat context

Co-authored-by: Limitless2023 <limitless@users.noreply.github.com>
---
 src/telegram/bot-message-context.ts | 23 +++++++++++++++--------
 1 file changed, 15 insertions(+), 8 deletions(-)

diff --git a/src/telegram/bot-message-context.ts b/src/telegram/bot-message-context.ts
index 041c93eab..62d6443c5 100644
--- a/src/telegram/bot-message-context.ts
+++ b/src/telegram/bot-message-context.ts
@@ -396,18 +396,11 @@ export const buildTelegramMessageContext = async ({
   }
 
   let bodyText = rawBody;
-  if (!bodyText && allMedia.length > 0) {
-    bodyText = `<media:image>${allMedia.length > 1 ? ` (${allMedia.length} images)` : ""}`;
-  }
-  const hasAnyMention = (msg.entities ?? msg.caption_entities ?? []).some(
-    (ent) => ent.type === "mention",
-  );
-  const explicitlyMentioned = botUsername ? hasBotMention(msg, botUsername) : false;
+  const hasAudio = allMedia.some((media) => media.contentType?.startsWith("audio/"));
 
   // Preflight audio transcription for mention detection in groups
   // This allows voice notes to be checked for mentions before being dropped
   let preflightTranscript: string | undefined;
-  const hasAudio = allMedia.some((media) => media.contentType?.startsWith("audio/"));
   const needsPreflightTranscription =
     isGroup && requireMention && hasAudio && !hasUserText && mentionRegexes.length > 0;
 
@@ -432,6 +425,20 @@ export const buildTelegramMessageContext = async ({
     }
   }
 
+  // Build bodyText - if there's audio with transcript, use transcript; otherwise use placeholder
+  if (!bodyText && allMedia.length > 0) {
+    if (hasAudio) {
+      bodyText = preflightTranscript || "<media:audio>";
+    } else {
+      bodyText = `<media:image>${allMedia.length > 1 ? ` (${allMedia.length} images)` : ""}`;
+    }
+  }
+
+  const hasAnyMention = (msg.entities ?? msg.caption_entities ?? []).some(
+    (ent) => ent.type === "mention",
+  );
+  const explicitlyMentioned = botUsername ? hasBotMention(msg, botUsername) : false;
+
   const computedWasMentioned = matchesMentionWithExplicit({
     text: msg.text ?? msg.caption ?? "",
     mentionRegexes,