mal-2026-17444
Vulnerability from ossf_malicious_packages
This package is a fork of the Baileys WhatsApp Web library (npm) and belongs to the "PhantomSub" family described by OX Security in September 2026: the publisher injected code that, without the installer asking, uses the installer's own authenticated WhatsApp session to subscribe that account to WhatsApp Channels (newsletters) the publisher chose, by sending the WhatsApp w:mex FOLLOW query (query_id 7871414976211147) for each target channel JID.
package/lib/Socket/newsletter.js lines 62-90 (cited line 68: QueryIds.FOLLOW) run in the body of makeNewsletterSocket, so they fire whenever a socket is created: 10 seconds later, and then at 3-second intervals, the code sends FOLLOW queries for three channel JIDs, each stored as a base64 literal decoded with Buffer.from(..., "base64"): 120363424403369453@newsletter, 120363423038562425@newsletter and 120363427530519865@newsletter. There is no option to turn this off and the README does not mention it.
Separately, package/lib/Socket/danz-auto-follow.js follows the channel behind invite code 0029VbD7H6m9MF9ALdskro07 on connect. That one is documented in the 1.1.0 README with a danzAutoFollowChannel: false opt-out; the base64-hidden follow of the three JIDs above is not.
Line numbers refer to version 1.1.0; the same code is present in every published version listed under affected.
Static review of the published tarball(s) found no code that sends WhatsApp credentials or session keys off-host, no install-time payload (the only install hook is Baileys' stock Node.js version check in engine-requirements.js), and no persistence; the payload runs when an application creates a WhatsApp socket with the library. The harm is the covert use of the victim's account to inflate the publisher's channel subscribers.
SHA-256 of the published npm tarball(s): 1.0.0: 1b568e46aa2ffbb186319e72b9fed52dde060f99319a82003d9d2ac15410dbbd; 1.1.0: 7370d73e2e5a0434fe41fc033ba9fc91d624a8a9461188fda82e04e3873e6d26.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "danz-bails"
},
"ranges": [
{
"events": [
{
"introduced": "0"
}
],
"type": "SEMVER"
}
],
"versions": [
"1.0.0",
"1.1.0"
]
}
],
"credits": [
{
"contact": [
"https://github.com/jaym2026"
],
"name": "Rising Sun Solutions LLC",
"type": "FINDER"
}
],
"database_specific": {
"iocs": {
"files": [
{
"digests": {
"sha256": "f638ea09996c7c80cbf6456963f3b4792295234036c2067d3db3106a5fbb9693"
},
"note": "File containing the injected WhatsApp Channel follow code.",
"paths": [
"package/lib/Socket/newsletter.js"
],
"source": "PACKAGE_ARCHIVE"
}
]
}
},
"details": "This package is a fork of the Baileys WhatsApp Web library (npm) and belongs to the \"PhantomSub\" family described by OX Security in September 2026: the publisher injected code that, without the installer asking, uses the installer\u0027s own authenticated WhatsApp session to subscribe that account to WhatsApp Channels (newsletters) the publisher chose, by sending the WhatsApp `w:mex` FOLLOW query (query_id 7871414976211147) for each target channel JID.\n\n`package/lib/Socket/newsletter.js` lines 62-90 (cited line 68: `QueryIds.FOLLOW`) run in the body of `makeNewsletterSocket`, so they fire whenever a socket is created: 10 seconds later, and then at 3-second intervals, the code sends FOLLOW queries for three channel JIDs, each stored as a base64 literal decoded with `Buffer.from(..., \"base64\")`: `120363424403369453@newsletter`, `120363423038562425@newsletter` and `120363427530519865@newsletter`. There is no option to turn this off and the README does not mention it.\n\nSeparately, `package/lib/Socket/danz-auto-follow.js` follows the channel behind invite code `0029VbD7H6m9MF9ALdskro07` on connect. That one is documented in the 1.1.0 README with a `danzAutoFollowChannel: false` opt-out; the base64-hidden follow of the three JIDs above is not.\n\nLine numbers refer to version 1.1.0; the same code is present in every published version listed under `affected`.\n\nStatic review of the published tarball(s) found no code that sends WhatsApp credentials or session keys off-host, no install-time payload (the only install hook is Baileys\u0027 stock Node.js version check in `engine-requirements.js`), and no persistence; the payload runs when an application creates a WhatsApp socket with the library. The harm is the covert use of the victim\u0027s account to inflate the publisher\u0027s channel subscribers.\n\nSHA-256 of the published npm tarball(s): 1.0.0: `1b568e46aa2ffbb186319e72b9fed52dde060f99319a82003d9d2ac15410dbbd`; 1.1.0: `7370d73e2e5a0434fe41fc033ba9fc91d624a8a9461188fda82e04e3873e6d26`.",
"id": "MAL-2026-17444",
"modified": "2026-10-01T23:05:37Z",
"references": [
{
"type": "ARTICLE",
"url": "https://www.ox.security/blog/phantomsub-malicious-npm-campaign-secretly-adds-users-to-whatsapp-spam-channels/"
},
{
"type": "PACKAGE",
"url": "https://www.npmjs.com/package/danz-bails"
}
],
"schema_version": "1.7.4",
"summary": "Malicious code in danz-bails (npm)"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.