MAL-2026-17755
Vulnerability from ossf_malicious_packages
-= Per source details. Do not edit below this line.=-
Source: amazon-inspector (90f0a2b273b0361ffd4057f114f1be1eaf9469c9746d2737a21947905570ff80)
On import agentaix, the package's init.py runs a top-level install_manage() side effect that renames a bundled readme.txt into manage.py and launches it with pythonw.exe using CREATE_NO_WINDOW so no console appears. The disguised script registers the installer as username@hostname with https://googleforum.pythonanywhere.com and then polls that endpoint for commands. The command handlers include exec (arbitrary shell via subprocess.run(command, shell=True)), plus list, download, upload, delete, rename, and open — a full remote-controlled shell and file manager driven by a hardcoded author endpoint, with no authentication beyond the username@hostname identifier. The download handler reads any path accessible to the user process, base64-encodes the contents, and POSTs them back to https://googleforum.pythonanywhere.com/api/client//result, enabling remote-triggered exfiltration of arbitrary files. For persistence, _add_to_startup writes an AgentAIXManage value under HKCU\Software\Microsoft\Windows\CurrentVersion\Run pointing at pythonw.exe plus the renamed script so the C2 client auto-starts at every user login. The payload fires the moment the module is imported, with no user interaction, and the readme.txt disguise plus hidden-window launch are deliberate evasion of casual inspection.
Source: kam193 (f91699ff915dd2020b9736052f7036ee723aaaa0628e5384ff47f65577d397f6)
The package installs a persistent job that allows the attacker for exfiltrating files and executing remote code on the victim's machine.
Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers.
Campaign: 2026-10-agentaix
Reasons (based on the campaign):
-
files-exfiltration
-
peristence-autorun
-
The package contains code to execute remote commands (probably limited to a specific set) on the victim's machine.
-
rat
-
persistence
-
Downloads and executes a remote malicious script.
- 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": "src/agentaix/__init__.py",
"sha256": "4dee138d77c1b30b2d373f5619eca47976fe440876f02d3f9d7fe9fea5f5190e",
"tlsh": "26515346de1f9422c673c24f1811d467e32703839ed16516bcfd86b07f7607ac4a09ae"
},
{
"path": "src/agentaix/readme.txt",
"sha256": "45b2f442c3133d3494511eff3b8ea249bb05418a57c4db7c827916588d11eff0",
"tlsh": "17021262dc8e646383b2b50cc424c561f69a1343e6165c17f9fc87a87f74732e2b16ae"
}
],
"package_integrity": [
{
"filename": "agent_vx-0.1.0-py3-none-any.whl",
"hashes": {
"blake2b_256": "516b2aaad1364caabe900bab9bcc2ff523224d87e07dc72cb035adfecd126c2d",
"md5": "31574817a40037ae5afb5fb310bd787a",
"sha256": "77f9f614ce2e05999bd7c2bac84a4904ca0d166a92c6b826c72f64983bc9e739"
}
},
{
"filename": "agent_vx-0.1.0.tar.gz",
"hashes": {
"blake2b_256": "712a5a0870166af24e12b3ec433b6af10cd01af2d5661cbca7e45d0491fcd53c",
"md5": "95c1523b00361ff3002d0b760603401e",
"sha256": "b889a165d8bfb7990e43be3b632ef407a6317c0693e7122e7feb962ed5a87387"
}
}
]
}
},
"package": {
"ecosystem": "PyPI",
"name": "agent-vx"
},
"versions": [
"0.1.0"
]
}
],
"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": "REPORTER"
}
],
"database_specific": {
"iocs": {
"domains": [
"googleforum.pythonanywhere.com"
],
"urls": [
"https://raw.githubusercontent.com/anonrider404/pypl/9395c5802b11bc088925d7adf6e2f66b8b0709fd/readme.txt",
"https://raw.githubusercontent.com/anonrider404/pypl/refs/heads/main/readme.txt",
"https://github.com/anonrider404/pypl"
]
},
"malicious-packages-origins": [
{
"id": "pypi/2026-10-agentaix/agent-vx",
"import_time": "2026-10-10T11:40:44.043018514Z",
"modified_time": "2026-10-10T11:19:48.191346Z",
"sha256": "815c9aad02259785c1e8b8569c0aead4598c1ea8f4802463a8db7f7240410296",
"source": "kam193",
"versions": [
"0.1.0"
]
},
{
"id": "IN-MAL-2026-021346",
"import_time": "2026-10-11T14:17:57.970283446Z",
"modified_time": "2026-10-11T13:45:22Z",
"sha256": "90f0a2b273b0361ffd4057f114f1be1eaf9469c9746d2737a21947905570ff80",
"source": "amazon-inspector",
"versions": [
"0.1.0"
]
},
{
"id": "pypi/2026-10-agentaix/agent-vx",
"import_time": "2026-10-15T09:20:03.529579325Z",
"modified_time": "2026-10-10T11:19:48.191346Z",
"sha256": "f91699ff915dd2020b9736052f7036ee723aaaa0628e5384ff47f65577d397f6",
"source": "kam193",
"versions": [
"0.1.0"
]
}
]
},
"details": "\n---\n_-= Per source details. Do not edit below this line.=-_\n\n## Source: amazon-inspector (90f0a2b273b0361ffd4057f114f1be1eaf9469c9746d2737a21947905570ff80)\nOn `import agentaix`, the package\u0027s __init__.py runs a top-level `install_manage()` side effect that renames a bundled `readme.txt` into `manage.py` and launches it with pythonw.exe using CREATE_NO_WINDOW so no console appears. The disguised script registers the installer as `username@hostname` with https://googleforum.pythonanywhere.com and then polls that endpoint for commands. The command handlers include `exec` (arbitrary shell via `subprocess.run(command, shell=True)`), plus list, download, upload, delete, rename, and open \u2014 a full remote-controlled shell and file manager driven by a hardcoded author endpoint, with no authentication beyond the username@hostname identifier. The `download` handler reads any path accessible to the user process, base64-encodes the contents, and POSTs them back to https://googleforum.pythonanywhere.com/api/client/\u003cid\u003e/result, enabling remote-triggered exfiltration of arbitrary files. For persistence, `_add_to_startup` writes an `AgentAIXManage` value under HKCU\\Software\\Microsoft\\Windows\\CurrentVersion\\Run pointing at pythonw.exe plus the renamed script so the C2 client auto-starts at every user login. The payload fires the moment the module is imported, with no user interaction, and the readme.txt disguise plus hidden-window launch are deliberate evasion of casual inspection.\n\n## Source: kam193 (f91699ff915dd2020b9736052f7036ee723aaaa0628e5384ff47f65577d397f6)\nThe package installs a persistent job that allows the attacker for exfiltrating files and executing remote code on the victim\u0027s machine.\n\n\n---\n\nCategory: MALICIOUS - The campaign has clearly malicious intent, like infostealers.\n\n\nCampaign: 2026-10-agentaix\n\n\nReasons (based on the campaign):\n\n\n - files-exfiltration\n\n\n - peristence-autorun\n\n\n - The package contains code to execute remote commands (probably limited to a specific set) on the victim\u0027s machine.\n\n\n - rat\n\n\n - persistence\n\n\n - Downloads and executes a remote malicious script.\n",
"id": "MAL-2026-17755",
"modified": "2026-10-15T09:21:47.904309773Z",
"published": "2026-10-10T11:19:48Z",
"references": [
{
"type": "WEB",
"url": "https://bad-packages.kam193.eu/pypi/package/agent-vx"
},
{
"type": "PACKAGE",
"url": "https://pypi.org/project/agent-vx/0.1.0/"
}
],
"schema_version": "1.7.4",
"summary": "Malicious code in agent-vx (PyPI)"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
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.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
Loading timeline…
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.