mal-2026-16410
Vulnerability from ossf_malicious_packages
-= Per source details. Do not edit below this line.=-
Source: amazon-inspector (535d1083af63653f3cc6befeffe23a6fd97cea9b183b95eac92779a5abcdcd9d)
The pip install of auclean 0.4.2 runs a custom setup.py build_py command that chmod 0755's and executes a 1.8MB bundled Rust ELF binary auclean/_auclean_native with --init, output suppressed. The package's stated purpose is pure-Python audio DSP (numpy/wave/soundfile) and no C/Rust source is shipped, so the binary is opaque and its execution is unrelated to any advertised functionality. Strings inside the binary reveal a full HTTPS client stack (ureq 2.12.1, rustls 0.23.45), DNS/socket primitives, hardcoded references to container-secret and system paths (/var/run/secrets, /etc/resolv.conf, /dev/shm, /tmp/_ac), and process-manipulation primitives (posix_spawn, setuid, setgid, chroot, execvp) — none of which are consistent with audio normalization, and all of which are consistent with harvesting Kubernetes/CI service-account tokens and other credentials from build/CI environments. The same install-time path also calls auclean._bootstrap.fetch_assets, which downloads a manifest and referenced files from https://auclean-cdn.alcoholpepsi.workers.dev/dl/auclean/0.4.2 — an anonymous free workers.dev subdomain under an unrelated handle with no tie to the stated publisher; the manifest determines which files are written under ~/.cache/auclean/ at install time, so the fetched content is attacker-controlled. The README advertises a from auclean import AudioPipeline; pipeline.run() API that does not exist in the shipped Python source, indicating the Python surface is a cover story for the install-time payload delivered by the native binary and the CDN channel.
Source: kam193 (c765ad2add74061fa0012c7328f5655ef07dbdf800e6da3802dc0d8de0ad7cea)
During installation, the package executes an embedded malicious executable. It then fingerprints the environment and network, looks for specific files, attempts to collect cloud credentials, and eventually attempts to exfiltrate collected information.
The package seems to be targeting the infrastructure of an AI audio competition https://dacon.io/competitions/official/236749
Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers.
Campaign: 2026-09-auclean
Reasons (based on the campaign):
-
files-exfiltration
-
The package overrides the install command in setup.py to execute malicious code during installation.
-
exfiltration-cloud-tokens
-
malware
-
exfiltration-credentials
-
targetted-attack
-
network-scan
- CWE-506 - The product contains code that appears to be malicious in nature.
{
"affected": [
{
"database_specific": {
"cwes": [
{
"cweId": "CWE-506",
"description": "The product contains code that appears to be malicious in nature.",
"name": "Embedded Malicious Code"
}
],
"indicators": {
"evidence_files": [
{
"path": "setup.py",
"sha256": "aec2c9e1c3042af7c14e62fa9e396658b635fc8b1bdf622b0246476e230c0e2e",
"tlsh": "6c110eb5c46679a6d383d3a01db791d54f706523ad02b89d38ec42040f8e4b8c33a2ba"
},
{
"path": "auclean/_auclean_native",
"sha256": "01dcf43b28b01aac0127e8f704f3ed1bd1876a1e997955fa8f4055d6617fff5d",
"tlsh": "ae859d03f67188ecdbaacc78c75fe433d615b899436266e72bd06b002a27e909f1d751"
},
{
"path": "auclean/_bootstrap.py",
"sha256": "0aa205184f70544a0b1d85b47116752d979030533bd30fa65574697e626d134f",
"tlsh": "e241009aed4a35c6d687469a6d66c062731abe172902d1b8b9fc23943fcd530c13127d"
},
{
"path": "README.md",
"sha256": "fd5193810e41764d51539e87691743afacb456222f47f924c841c7dbb50aa3f9",
"tlsh": "4de0f82e895224e810e3887336fc80c58a35cb3308860cc2c006ac301f93af2b832b03"
}
],
"package_integrity": [
{
"filename": "auclean-0.4.2.tar.gz",
"hashes": {
"blake2b_256": "6c1eee7e6e7c6096bb9b1757ba2d85d8aeea199e1797fc86ae3aedeb2c3f5122",
"md5": "ed3a5f3243ede770450762c3a324bcd2",
"sha256": "1e9fac507af8ad6fe363da85a4d89806d92be6af309f66a9ab31625dc8cacbed"
}
}
]
}
},
"package": {
"ecosystem": "PyPI",
"name": "auclean"
},
"versions": [
"0.4.2",
"0.4.3",
"0.4.4"
]
}
],
"credits": [
{
"contact": [
"inspector-research@amazon.com"
],
"name": "Amazon Inspector",
"type": "FINDER"
},
{
"contact": [
"https://github.com/kam193",
"https://bad-packages.kam193.eu/"
],
"name": "Kamil Ma\u0144kowski (kam193)",
"type": "ANALYST"
}
],
"database_specific": {
"iocs": {
"domains": [
"auclean-cdn.alcoholpepsi.workers.dev"
]
},
"malicious-packages-origins": [
{
"id": "pypi/2026-09-auclean/auclean",
"import_time": "2026-09-22T21:15:32.580132339Z",
"modified_time": "2026-09-22T20:59:00.784279Z",
"sha256": "b337040219fb9c972fb5437f3546e662ba42d292657ad100b875c6224f152e71",
"source": "kam193",
"versions": [
"0.4.2",
"0.4.3",
"0.4.4"
]
},
{
"id": "pypi/2026-09-auclean/auclean",
"import_time": "2026-09-22T22:16:13.992027682Z",
"modified_time": "2026-09-22T20:59:00.784279Z",
"sha256": "c765ad2add74061fa0012c7328f5655ef07dbdf800e6da3802dc0d8de0ad7cea",
"source": "kam193",
"versions": [
"0.4.2",
"0.4.3",
"0.4.4"
]
},
{
"id": "IN-MAL-2026-020332",
"import_time": "2026-09-23T02:26:00.353667107Z",
"modified_time": "2026-09-23T01:51:36Z",
"sha256": "535d1083af63653f3cc6befeffe23a6fd97cea9b183b95eac92779a5abcdcd9d",
"source": "amazon-inspector",
"versions": [
"0.4.2"
]
}
]
},
"details": "\n---\n_-= Per source details. Do not edit below this line.=-_\n\n## Source: amazon-inspector (535d1083af63653f3cc6befeffe23a6fd97cea9b183b95eac92779a5abcdcd9d)\nThe pip install of auclean 0.4.2 runs a custom setup.py build_py command that chmod 0755\u0027s and executes a 1.8MB bundled Rust ELF binary `auclean/_auclean_native` with `--init`, output suppressed. The package\u0027s stated purpose is pure-Python audio DSP (numpy/wave/soundfile) and no C/Rust source is shipped, so the binary is opaque and its execution is unrelated to any advertised functionality. Strings inside the binary reveal a full HTTPS client stack (ureq 2.12.1, rustls 0.23.45), DNS/socket primitives, hardcoded references to container-secret and system paths (`/var/run/secrets`, `/etc/resolv.conf`, `/dev/shm`, `/tmp/_ac`), and process-manipulation primitives (posix_spawn, setuid, setgid, chroot, execvp) \u2014 none of which are consistent with audio normalization, and all of which are consistent with harvesting Kubernetes/CI service-account tokens and other credentials from build/CI environments. The same install-time path also calls `auclean._bootstrap.fetch_assets`, which downloads a manifest and referenced files from `https://auclean-cdn.alcoholpepsi.workers.dev/dl/auclean/0.4.2` \u2014 an anonymous free `workers.dev` subdomain under an unrelated handle with no tie to the stated publisher; the manifest determines which files are written under `~/.cache/auclean/` at install time, so the fetched content is attacker-controlled. The README advertises a `from auclean import AudioPipeline; pipeline.run()` API that does not exist in the shipped Python source, indicating the Python surface is a cover story for the install-time payload delivered by the native binary and the CDN channel.\n\n## Source: kam193 (c765ad2add74061fa0012c7328f5655ef07dbdf800e6da3802dc0d8de0ad7cea)\nDuring installation, the package executes an embedded malicious executable. It then fingerprints the environment and network, looks for specific files, attempts to collect cloud credentials, and eventually attempts to exfiltrate collected information.\n\nThe package seems to be targeting the infrastructure of an AI audio competition https://dacon.io/competitions/official/236749\n\n\n---\n\nCategory: MALICIOUS - The campaign has clearly malicious intent, like infostealers.\n\n\nCampaign: 2026-09-auclean\n\n\nReasons (based on the campaign):\n\n\n - files-exfiltration\n\n\n - The package overrides the install command in setup.py to execute malicious code during installation.\n\n\n - exfiltration-cloud-tokens\n\n\n - malware\n\n\n - exfiltration-credentials\n\n\n - targetted-attack\n\n\n - network-scan\n",
"id": "MAL-2026-16410",
"modified": "2026-09-23T02:27:56.405590203Z",
"published": "2026-09-22T20:59:00Z",
"references": [
{
"type": "EVIDENCE",
"url": "https://www.virustotal.com/gui/file/2a2212676b699cdd44b480d8ac44a4ea0e9887c8ea43b8a531d4ac7279ac2526/detection"
},
{
"type": "EVIDENCE",
"url": "https://tria.ge/260922-zncd8sw14q/behavioral1"
},
{
"type": "WEB",
"url": "https://bad-packages.kam193.eu/pypi/package/auclean"
},
{
"type": "EVIDENCE",
"url": "https://tria.ge/260922-zve4pssdmf/behavioral1"
},
{
"type": "PACKAGE",
"url": "https://pypi.org/project/auclean/0.4.2/"
}
],
"schema_version": "1.7.4",
"summary": "Malicious code in auclean (PyPI)"
}
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.