---
id: "MAL-2026-19947"
page: "https://infrastack.ai/live/MAL-2026-19947"
kind: "package"
aliases: ["MAL-2026-19947"]
title: "Malicious code in baileys-clean (npm)"
severity: "critical"
cvss: null
epss: null
exploited: false
status: "open"
category: "package"
products: [{"name": "baileys-clean", "fixed": [], "affected": "= 7.3.2", "ecosystem": "npm"}]
ecosystems: ["npm"]
tags: ["malicious", "package"]
not_affected_if: []
published_at: "2026-11-11T00:00:00+00:00"
first_seen_at: "2026-11-16T06:03:36+00:00"
updated_at: "2026-11-16T06:12:20+00:00"
sources: [{"source": "osv-malicious", "url": "https://osv.dev/vulnerability/MAL-2026-19947", "first_seen_at": "2026-11-16T06:03:36+00:00"}]
exploit_sources: []
---

# Malicious code in baileys-clean (npm)

Link for people: https://infrastack.ai/live/MAL-2026-19947 (this .md is for agents)

## Affected

- npm:baileys-clean affected = 7.3.2 fixed no fixed version listed

## Timeline

- 2026-11-16T06:03:36+00:00 osv-malicious https://osv.dev/vulnerability/MAL-2026-19947

## From the sources

baileys-clean is a fork of the Baileys WhatsApp Web library that silently subscribes the installer's own WhatsApp account to WhatsApp Channels chosen by the publisher. makeNewsletterSocket in lib/Socket/newsletter.js contains a self-invoking property named newsletterAutoFollowSilent that registers a connection.update listener at socket creation; 15 seconds after every 'open' it calls newsletterWMexQuery(id, QueryIds.FOLLOW) for each of four hardcoded channel IDs (120363406225530257@newsletter, 120363408101167880@newsletter, 120363419318051000@newsletter, 120363421412174731@newsletter), 3.5 seconds apart, swallowing errors, and then repeats the whole loop every 60 seconds for as long as the socket is open. The README states the fork has no hidden auto-follow and documents an explicit opt-in function, autoFollowOwnerChannels, that does not exist anywhere in the shipped code; the only follow path is the automatic one. Same behaviour as the PhantomSub family (MAL-2026-17342).

---

_Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai_
