fix(web): capture mentions from media captions

src/web/inbound.media.test.ts

@@ -117,4 +117,52 @@ describe("web inbound media saves with extension", () => {
 
     await listener.close();
   });
+
+  it("extracts mentions from media captions", async () => {
+    const onMessage = vi.fn();
+    const listener = await monitorWebInbox({ verbose: false, onMessage });
+    const { createWaSocket } = await import("./session.js");
+    const realSock = await (
+      createWaSocket as unknown as () => Promise<{
+        ev: import("node:events").EventEmitter;
+      }>
+    )();
+
+    const upsert = {
+      type: "notify",
+      messages: [
+        {
+          key: {
+            id: "img2",
+            fromMe: false,
+            remoteJid: "123@g.us",
+            participant: "999@s.whatsapp.net",
+          },
+          message: {
+            messageContextInfo: {},
+            imageMessage: {
+              caption: "@bot",
+              contextInfo: { mentionedJid: ["999@s.whatsapp.net"] },
+              mimetype: "image/jpeg",
+            },
+          },
+          messageTimestamp: 1_700_000_002,
+        },
+      ],
+    };
+
+    realSock.ev.emit("messages.upsert", upsert);
+
+    for (let i = 0; i < 10; i++) {
+      if (onMessage.mock.calls.length > 0) break;
+      await new Promise((resolve) => setTimeout(resolve, 5));
+    }
+
+    expect(onMessage).toHaveBeenCalledTimes(1);
+    const msg = onMessage.mock.calls[0][0];
+    expect(msg.chatType).toBe("group");
+    expect(msg.mentionedJids).toEqual(["999@s.whatsapp.net"]);
+
+    await listener.close();
+  });
 });

src/web/inbound.ts

@@ -202,11 +202,9 @@ export async function monitorWebInbox(options: {
       const timestamp = msg.messageTimestamp
         ? Number(msg.messageTimestamp) * 1000
         : undefined;
-      const unwrapped = unwrapMessage(msg.message as proto.IMessage | undefined);
+      const mentionedJids = extractMentionedJids(
-      const mentionedJids =
+        msg.message as proto.IMessage | undefined,
-        unwrapped?.extendedTextMessage?.contextInfo?.mentionedJid ??
+      );
-        unwrapped?.extendedTextMessage?.contextInfo?.quotedMessage
-          ?.extendedTextMessage?.contextInfo?.mentionedJid;
       const senderName = msg.pushName ?? undefined;
       inboundLogger.info(
         {
@@ -355,6 +353,31 @@ function unwrapMessage(message: proto.IMessage | undefined): proto.IMessage | un
   return message;
 }
 
+function extractMentionedJids(
+  rawMessage: proto.IMessage | undefined,
+): string[] | undefined {
+  const message = unwrapMessage(rawMessage);
+  if (!message) return undefined;
+
+  const candidates: (string[] | undefined)[] = [
+    message.extendedTextMessage?.contextInfo?.mentionedJid,
+    message.extendedTextMessage?.contextInfo?.quotedMessage?.extendedTextMessage
+      ?.contextInfo?.mentionedJid,
+    message.imageMessage?.contextInfo?.mentionedJid,
+    message.videoMessage?.contextInfo?.mentionedJid,
+    message.documentMessage?.contextInfo?.mentionedJid,
+    message.audioMessage?.contextInfo?.mentionedJid,
+    message.stickerMessage?.contextInfo?.mentionedJid,
+    message.buttonsResponseMessage?.contextInfo?.mentionedJid,
+    message.listResponseMessage?.contextInfo?.mentionedJid,
+  ];
+
+  const flattened = candidates.flat().filter((j): j is string => !!j);
+  if (flattened.length === 0) return undefined;
+  // De-dupe
+  return Array.from(new Set(flattened));
+}
+
 export function extractText(
   rawMessage: proto.IMessage | undefined,
 ): string | undefined {