Search
Find a vulnerability
Search criteria
ⓘ
Use this form to refine search results.
Full-text search supports keyword queries with ranking and filtering.
You can combine vendor, product, and sources to narrow results.
Enable “Apply ordering” to sort by date instead of relevance.
82 vulnerabilities by huggingface
CVE-2026-5241 (GCVE-0-2026-5241)
Vulnerability from nvd – Published: 2026-06-03 12:33 – Updated: 2026-06-03 15:45
VLAI
Title
Policy Bypass in LightGlue Nested Config Resolution in huggingface/transformers
Summary
A vulnerability in the LightGlue model loading path of huggingface/transformers version 5.2.0 allows an attacker-controlled model repository to execute arbitrary code during model initialization. The issue arises because the `trust_remote_code` parameter, intended to prevent remote code execution, is overridden by untrusted serialized configuration data in a nested code path. Specifically, when loading a LightGlue model using `AutoModel.from_pretrained()` with `trust_remote_code=False`, the `LightGlueConfig` reads the `trust_remote_code` value from the untrusted `config.json` file and propagates it into nested `AutoConfig.from_pretrained()` calls. This results in the execution of attacker-provided Python modules, even when the victim explicitly disables remote code execution. The vulnerability poses a high risk for environments such as API inference servers, research notebooks, CI/CD pipelines, and model evaluation workers, potentially leading to credential theft, lateral movement, or persistence/backdoor deployment.
Severity
SSVC
Exploitation: poc
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-829 - Inclusion of Functionality from Untrusted Control Sphere
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 5.5.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-5241",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-06-03T15:45:14.993177Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-06-03T15:45:36.684Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "5.5.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the LightGlue model loading path of huggingface/transformers version 5.2.0 allows an attacker-controlled model repository to execute arbitrary code during model initialization. The issue arises because the `trust_remote_code` parameter, intended to prevent remote code execution, is overridden by untrusted serialized configuration data in a nested code path. Specifically, when loading a LightGlue model using `AutoModel.from_pretrained()` with `trust_remote_code=False`, the `LightGlueConfig` reads the `trust_remote_code` value from the untrusted `config.json` file and propagates it into nested `AutoConfig.from_pretrained()` calls. This results in the execution of attacker-provided Python modules, even when the victim explicitly disables remote code execution. The vulnerability poses a high risk for environments such as API inference servers, research notebooks, CI/CD pipelines, and model evaluation workers, potentially leading to credential theft, lateral movement, or persistence/backdoor deployment."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "HIGH",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "CHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-829",
"description": "CWE-829 Inclusion of Functionality from Untrusted Control Sphere",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-06-03T12:33:10.227Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619"
},
{
"url": "https://github.com/huggingface/transformers/commit/676559d5022b74aaa0cee1cee0842b7f27c5320e"
}
],
"source": {
"advisory": "ceb3ce1a-4c45-497a-b25e-cb9a7685e619",
"discovery": "EXTERNAL"
},
"title": "Policy Bypass in LightGlue Nested Config Resolution in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-5241",
"datePublished": "2026-06-03T12:33:10.227Z",
"dateReserved": "2026-03-31T14:26:14.353Z",
"dateUpdated": "2026-06-03T15:45:36.684Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-4372 (GCVE-0-2026-4372)
Vulnerability from nvd – Published: 2026-05-24 13:40 – Updated: 2026-05-26 15:21
VLAI
Title
Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers
Summary
A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker's repository with the victim's full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitization of internal fields, and unsandboxed execution of downloaded kernels. The vulnerability bypasses the `trust_remote_code` security mechanism, is invisible to the victim, and exploits the standard documented usage pattern, making it particularly severe. Users are advised to upgrade to version 5.3.0 or later to mitigate this issue.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-1066 - Missing Serialization Control Element
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 5.3.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-4372",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-05-26T15:08:29.632933Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-05-26T15:21:02.376Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "5.3.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker\u0027s repository with the victim\u0027s full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitization of internal fields, and unsandboxed execution of downloaded kernels. The vulnerability bypasses the `trust_remote_code` security mechanism, is invisible to the victim, and exploits the standard documented usage pattern, making it particularly severe. Users are advised to upgrade to version 5.3.0 or later to mitigate this issue."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 7.8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1066",
"description": "CWE-1066 Missing Serialization Control Element",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-05-24T13:40:40.995Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/1f693a6e-6836-4b8b-a0bd-ca036fba8884"
},
{
"url": "https://github.com/huggingface/transformers/commit/a7f8e7ff37d87d1a1a0c8cf607971c607741452f"
}
],
"source": {
"advisory": "1f693a6e-6836-4b8b-a0bd-ca036fba8884",
"discovery": "EXTERNAL"
},
"title": "Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-4372",
"datePublished": "2026-05-24T13:40:40.995Z",
"dateReserved": "2026-03-18T10:04:09.683Z",
"dateUpdated": "2026-05-26T15:21:02.376Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-44827 (GCVE-0-2026-44827)
Vulnerability from nvd – Published: 2026-05-14 16:33 – Updated: 2026-05-14 18:03
VLAI
Title
Diffusers: None.py Trust Remote Code Bypass
Summary
Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, diffusers 0.37.0 allows remote code execution without the trust_remote_code=True safeguard when loading pipelines from Hugging Face Hub repositories. The _resolve_custom_pipeline_and_cls function in pipeline_loading_utils.py performs string interpolation on the custom_pipeline parameter using f"{custom_pipeline}.py". When custom_pipeline is not supplied by the user, it defaults to None, which Python interpolates as the literal string "None.py". If an attacker publishes a Hub repository containing a file named None.py with a class that subclasses DiffusionPipeline, the file is automatically downloaded and executed during a standard DiffusionPipeline.from_pretrained() call with no additional keyword arguments. The trust_remote_code check in DiffusionPipeline.download() is bypassed because it evaluates custom_pipeline is not None as False (since the kwarg was never supplied), while the downstream code path that actually loads the module resolves the None value into a valid filename. An attacker can achieve silent arbitrary code execution by publishing a malicious model repository with a None.py file and a standard-looking model_index.json that references a legitimate pipeline class name, requiring only that a victim calls from_pretrained on the repository. This vulnerability is fixed in 0.38.0.
Severity
8.8 (High)
SSVC
Exploitation: poc
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://github.com/huggingface/diffusers/security… | x_refsource_CONFIRM |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | diffusers |
Affected:
< 0.38.0
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-44827",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-05-14T18:00:53.494731Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T18:03:35.120Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://github.com/huggingface/diffusers/security/advisories/GHSA-j7w6-vpvq-j3gm"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "diffusers",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "\u003c 0.38.0"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, diffusers 0.37.0 allows remote code execution without the trust_remote_code=True safeguard when loading pipelines from Hugging Face Hub repositories. The _resolve_custom_pipeline_and_cls function in pipeline_loading_utils.py performs string interpolation on the custom_pipeline parameter using f\"{custom_pipeline}.py\". When custom_pipeline is not supplied by the user, it defaults to None, which Python interpolates as the literal string \"None.py\". If an attacker publishes a Hub repository containing a file named None.py with a class that subclasses DiffusionPipeline, the file is automatically downloaded and executed during a standard DiffusionPipeline.from_pretrained() call with no additional keyword arguments. The trust_remote_code check in DiffusionPipeline.download() is bypassed because it evaluates custom_pipeline is not None as False (since the kwarg was never supplied), while the downstream code path that actually loads the module resolves the None value into a valid filename. An attacker can achieve silent arbitrary code execution by publishing a malicious model repository with a None.py file and a standard-looking model_index.json that references a legitimate pipeline class name, requiring only that a victim calls from_pretrained on the repository. This vulnerability is fixed in 0.38.0."
}
],
"metrics": [
{
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 8.8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T16:33:42.373Z",
"orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"shortName": "GitHub_M"
},
"references": [
{
"name": "https://github.com/huggingface/diffusers/security/advisories/GHSA-j7w6-vpvq-j3gm",
"tags": [
"x_refsource_CONFIRM"
],
"url": "https://github.com/huggingface/diffusers/security/advisories/GHSA-j7w6-vpvq-j3gm"
}
],
"source": {
"advisory": "GHSA-j7w6-vpvq-j3gm",
"discovery": "UNKNOWN"
},
"title": "Diffusers: None.py Trust Remote Code Bypass"
}
},
"cveMetadata": {
"assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"assignerShortName": "GitHub_M",
"cveId": "CVE-2026-44827",
"datePublished": "2026-05-14T16:33:42.373Z",
"dateReserved": "2026-05-07T21:21:48.351Z",
"dateUpdated": "2026-05-14T18:03:35.120Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-44513 (GCVE-0-2026-44513)
Vulnerability from nvd – Published: 2026-05-14 16:26 – Updated: 2026-05-14 19:51
VLAI
Title
Diffusers: `trust_remote_code` bypass via `custom_pipeline` and local custom components
Summary
Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, a trust_remote_code bypass in DiffusionPipeline.from_pretrained allows arbitrary remote code execution despite the user passing trust_remote_code=False (or omitting it, which is the default). The vulnerability has three variants, all sharing the same root cause — the trust_remote_code gate was implemented inside DiffusionPipeline.download() rather than at the actual dynamic-module load site, so any code path that bypassed or short-circuited download() also bypassed the security check. DiffusionPipeline.from_pretrained('repoA', custom_pipeline='attacker/repoB', trust_remote_code=False) — the gate evaluated against repoA's file list rather than repoB's, so repoB's pipeline.py was loaded and executed. DiffusionPipeline.from_pretrained('/local/snapshot', custom_pipeline='attacker/repoB', trust_remote_code=False) — the local-path branch never invoked download(), so the gate was never reached and remote code from repoB executed. DiffusionPipeline.from_pretrained('/local/snapshot', trust_remote_code=False) where the snapshot contains custom component files (e.g. unet/my_unet_model.py) referenced from model_index.json — same root cause; the local path skipped download() and custom component code executed. This vulnerability is fixed in 0.38.0.
Severity
8.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://github.com/huggingface/diffusers/security… | x_refsource_CONFIRM |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | diffusers |
Affected:
< 0.38.0
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-44513",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-05-14T17:38:51.150920Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T19:51:06.991Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "diffusers",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "\u003c 0.38.0"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, a trust_remote_code bypass in DiffusionPipeline.from_pretrained allows arbitrary remote code execution despite the user passing trust_remote_code=False (or omitting it, which is the default). The vulnerability has three variants, all sharing the same root cause \u2014 the trust_remote_code gate was implemented inside DiffusionPipeline.download() rather than at the actual dynamic-module load site, so any code path that bypassed or short-circuited download() also bypassed the security check. DiffusionPipeline.from_pretrained(\u0027repoA\u0027, custom_pipeline=\u0027attacker/repoB\u0027, trust_remote_code=False) \u2014 the gate evaluated against repoA\u0027s file list rather than repoB\u0027s, so repoB\u0027s pipeline.py was loaded and executed. DiffusionPipeline.from_pretrained(\u0027/local/snapshot\u0027, custom_pipeline=\u0027attacker/repoB\u0027, trust_remote_code=False) \u2014 the local-path branch never invoked download(), so the gate was never reached and remote code from repoB executed. DiffusionPipeline.from_pretrained(\u0027/local/snapshot\u0027, trust_remote_code=False) where the snapshot contains custom component files (e.g. unet/my_unet_model.py) referenced from model_index.json \u2014 same root cause; the local path skipped download() and custom component code executed. This vulnerability is fixed in 0.38.0."
}
],
"metrics": [
{
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 8.8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T16:26:03.907Z",
"orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"shortName": "GitHub_M"
},
"references": [
{
"name": "https://github.com/huggingface/diffusers/security/advisories/GHSA-98h9-4798-4q5v",
"tags": [
"x_refsource_CONFIRM"
],
"url": "https://github.com/huggingface/diffusers/security/advisories/GHSA-98h9-4798-4q5v"
}
],
"source": {
"advisory": "GHSA-98h9-4798-4q5v",
"discovery": "UNKNOWN"
},
"title": "Diffusers: `trust_remote_code` bypass via `custom_pipeline` and local custom components"
}
},
"cveMetadata": {
"assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"assignerShortName": "GitHub_M",
"cveId": "CVE-2026-44513",
"datePublished": "2026-05-14T16:26:03.907Z",
"dateReserved": "2026-05-06T18:28:20.887Z",
"dateUpdated": "2026-05-14T19:51:06.991Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-25874 (GCVE-0-2026-25874)
Vulnerability from nvd – Published: 2026-04-23 19:45 – Updated: 2026-06-23 16:13 X_Open Source
VLAI
Title
LeRobot Unsafe Deserialization Remote Code Execution via gRPC
Summary
LeRobot through 0.5.1 contains an unsafe deserialization vulnerability in the async inference pipeline where pickle.loads() is used to deserialize data received over unauthenticated gRPC channels without TLS in the policy server and robot client components. An unauthenticated network-reachable attacker can achieve arbitrary code execution on the server or client by sending a crafted pickle payload through the SendPolicyInstructions, SendObservations, or GetActions gRPC calls.
Severity
SSVC
Exploitation: poc
Automatable: yes
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
5 references
| URL | Tags |
|---|---|
| https://chocapikk.com/posts/2026/lerobot-pickle-rce/ | exploit |
| https://github.com/huggingface/lerobot/issues/3047 | issue-tracking |
| https://github.com/huggingface/lerobot/pull/3048 | issue-tracking |
| https://github.com/huggingface/lerobot/issues/3134 | issue-tracking |
| https://www.vulncheck.com/advisories/lerobot-unsa… | third-party-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | LeRobot |
Affected:
0 , ≤ 0.5.1
(semver)
|
Date Public
2026-02-27 00:00
Credits
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-25874",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-04-24T14:34:03.307494Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-04-24T18:20:13.815Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "LeRobot",
"repo": "https://github.com/huggingface/lerobot",
"vendor": "Hugging Face",
"versions": [
{
"lessThanOrEqual": "0.5.1",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "finder",
"value": "Valentin Lobstein (Chocapikk)"
}
],
"datePublic": "2026-02-27T00:00:00.000Z",
"descriptions": [
{
"lang": "en",
"value": "LeRobot through 0.5.1 contains an unsafe deserialization vulnerability in the async inference pipeline where pickle.loads() is used to deserialize data received over unauthenticated gRPC channels without TLS in the policy server and robot client components. An unauthenticated network-reachable attacker can achieve arbitrary code execution on the server or client by sending a crafted pickle payload through the SendPolicyInstructions, SendObservations, or GetActions gRPC calls."
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 9.3,
"baseSeverity": "CRITICAL",
"exploitMaturity": "NOT_DEFINED",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-06-23T16:13:56.651Z",
"orgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"shortName": "VulnCheck"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://chocapikk.com/posts/2026/lerobot-pickle-rce/"
},
{
"tags": [
"issue-tracking"
],
"url": "https://github.com/huggingface/lerobot/issues/3047"
},
{
"tags": [
"issue-tracking"
],
"url": "https://github.com/huggingface/lerobot/pull/3048"
},
{
"tags": [
"issue-tracking"
],
"url": "https://github.com/huggingface/lerobot/issues/3134"
},
{
"tags": [
"third-party-advisory"
],
"url": "https://www.vulncheck.com/advisories/lerobot-unsafe-deserialization-remote-code-execution-via-grpc"
}
],
"source": {
"discovery": "EXTERNAL"
},
"tags": [
"x_open-source"
],
"title": "LeRobot Unsafe Deserialization Remote Code Execution via gRPC",
"x_generator": {
"engine": "Vulnogram 1.0.1"
}
}
},
"cveMetadata": {
"assignerOrgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"assignerShortName": "VulnCheck",
"cveId": "CVE-2026-25874",
"datePublished": "2026-04-23T19:45:01.090Z",
"dateReserved": "2026-02-06T19:12:03.464Z",
"dateUpdated": "2026-06-23T16:13:56.651Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-1839 (GCVE-0-2026-1839)
Vulnerability from nvd – Published: 2026-04-07 05:22 – Updated: 2026-04-07 13:27
VLAI
Title
Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers
Summary
A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch>=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3.
Severity
6.5 (Medium)
SSVC
Exploitation: poc
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < v5.0.0rc3
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-1839",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-04-07T13:27:38.099805Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-04-07T13:27:41.789Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/3c77bb97-e493-493d-9a88-c57f5c536485"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "v5.0.0rc3",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch\u003e=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "HIGH",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 6.5,
"baseSeverity": "MEDIUM",
"confidentialityImpact": "HIGH",
"integrityImpact": "LOW",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:L/A:H",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502 Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-04-07T05:22:00.959Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/3c77bb97-e493-493d-9a88-c57f5c536485"
},
{
"url": "https://github.com/huggingface/transformers/commit/03c8082ba4594c9b8d6fe190ca9bed0e5f8ca396"
}
],
"source": {
"advisory": "3c77bb97-e493-493d-9a88-c57f5c536485",
"discovery": "EXTERNAL"
},
"title": "Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-1839",
"datePublished": "2026-04-07T05:22:00.959Z",
"dateReserved": "2026-02-03T16:49:27.781Z",
"dateUpdated": "2026-04-07T13:27:41.789Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-4963 (GCVE-0-2026-4963)
Vulnerability from nvd – Published: 2026-03-27 17:05 – Updated: 2026-03-31 13:32
VLAI
Title
huggingface smolagents Incomplete Fix CVE-2025-9959 local_python_executor.py evaluate_with code injection
Summary
A weakness has been identified in huggingface smolagents 1.25.0.dev0. This affects the function evaluate_augassign/evaluate_call/evaluate_with of the file src/smolagents/local_python_executor.py of the component Incomplete Fix CVE-2025-9959. This manipulation causes code injection. It is possible to initiate the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
Severity
SSVC
Exploitation: poc
Automatable: no
Technical Impact: partial
CISA Coordinator (v2.0.3)
Assigner
References
7 references
| URL | Tags |
|---|---|
| https://vuldb.com/?id.353840 | vdb-entrytechnical-description |
| https://vuldb.com/?ctiid.353840 | signaturepermissions-required |
| https://vuldb.com/?submit.777623 | third-party-advisory |
| https://vuldb.com/?submit.777643 | third-party-advisory |
| https://vuldb.com/?submit.777644 | third-party-advisory |
| https://gist.github.com/YLChen-007/7146f45960f79b… | related |
| https://gist.github.com/YLChen-007/35b7d46e892266… | exploit |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | smolagents |
Affected:
1.25.0.dev0
cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:* |
Credits
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-4963",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-03-31T13:32:07.450819Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-03-31T13:32:41.488Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"cpes": [
"cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:*"
],
"modules": [
"Incomplete Fix CVE-2025-9959"
],
"product": "smolagents",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "1.25.0.dev0"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Eric-z (VulDB User)"
},
{
"lang": "en",
"type": "coordinator",
"value": "VulDB"
}
],
"descriptions": [
{
"lang": "en",
"value": "A weakness has been identified in huggingface smolagents 1.25.0.dev0. This affects the function evaluate_augassign/evaluate_call/evaluate_with of the file src/smolagents/local_python_executor.py of the component Incomplete Fix CVE-2025-9959. This manipulation causes code injection. It is possible to initiate the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way."
}
],
"metrics": [
{
"cvssV4_0": {
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P",
"version": "4.0"
}
},
{
"cvssV3_1": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.1"
}
},
{
"cvssV3_0": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.0"
}
},
{
"cvssV2_0": {
"baseScore": 7.5,
"vectorString": "AV:N/AC:L/Au:N/C:P/I:P/A:P/E:POC/RL:ND/RC:UR",
"version": "2.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "Code Injection",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-74",
"description": "Injection",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-03-27T22:07:35.402Z",
"orgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"shortName": "VulDB"
},
"references": [
{
"name": "VDB-353840 | huggingface smolagents Incomplete Fix CVE-2025-9959 local_python_executor.py evaluate_with code injection",
"tags": [
"vdb-entry",
"technical-description"
],
"url": "https://vuldb.com/?id.353840"
},
{
"name": "VDB-353840 | CTI Indicators (IOB, IOC, TTP, IOA)",
"tags": [
"signature",
"permissions-required"
],
"url": "https://vuldb.com/?ctiid.353840"
},
{
"name": "Submit #777623 | HuggingFace smolagents 1.25.0.dev0 CWE-693",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.777623"
},
{
"name": "Submit #777643 | HuggingFace smolagents 1.25.0.dev0 CWE-693 (Duplicate)",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.777643"
},
{
"name": "Submit #777644 | HuggingFace smolagents 1.25.0.dev0 CWE-693 (Duplicate)",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.777644"
},
{
"tags": [
"related"
],
"url": "https://gist.github.com/YLChen-007/7146f45960f79bc1e2976fed526e0a9b"
},
{
"tags": [
"exploit"
],
"url": "https://gist.github.com/YLChen-007/35b7d46e892266a0ed6dbe57802858be"
}
],
"timeline": [
{
"lang": "en",
"time": "2026-03-27T00:00:00.000Z",
"value": "Advisory disclosed"
},
{
"lang": "en",
"time": "2026-03-27T01:00:00.000Z",
"value": "VulDB entry created"
},
{
"lang": "en",
"time": "2026-03-27T09:23:31.000Z",
"value": "VulDB entry last update"
}
],
"title": "huggingface smolagents Incomplete Fix CVE-2025-9959 local_python_executor.py evaluate_with code injection"
}
},
"cveMetadata": {
"assignerOrgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"assignerShortName": "VulDB",
"cveId": "CVE-2026-4963",
"datePublished": "2026-03-27T17:05:20.096Z",
"dateReserved": "2026-03-27T08:17:46.710Z",
"dateUpdated": "2026-03-31T13:32:41.488Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-2654 (GCVE-0-2026-2654)
Vulnerability from nvd – Published: 2026-02-18 13:32 – Updated: 2026-02-23 10:17
VLAI
Title
huggingface smolagents LocalPythonExecutor requests.post server-side request forgery
Summary
A weakness has been identified in huggingface smolagents 1.24.0. Impacted is the function requests.get/requests.post of the component LocalPythonExecutor. Executing a manipulation can lead to server-side request forgery. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
Severity
SSVC
Exploitation: poc
Automatable: no
Technical Impact: partial
CISA Coordinator (v2.0.3)
CWE
- CWE-918 - Server-Side Request Forgery
Assigner
References
5 references
| URL | Tags |
|---|---|
| https://vuldb.com/?id.346451 | vdb-entrytechnical-description |
| https://vuldb.com/?ctiid.346451 | signaturepermissions-required |
| https://vuldb.com/?submit.752774 | third-party-advisory |
| https://github.com/CH0ico/CVE_choco_smolagent/tree/main | related |
| https://github.com/CH0ico/CVE_choco_smolagent/blo… | exploit |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | smolagents |
Affected:
1.24.0
cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:* |
Credits
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-2654",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-02-18T14:10:54.671299Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-02-18T14:11:19.055Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"cpes": [
"cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:*"
],
"modules": [
"LocalPythonExecutor"
],
"product": "smolagents",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "1.24.0"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Choco094late (VulDB User)"
}
],
"descriptions": [
{
"lang": "en",
"value": "A weakness has been identified in huggingface smolagents 1.24.0. Impacted is the function requests.get/requests.post of the component LocalPythonExecutor. Executing a manipulation can lead to server-side request forgery. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way."
}
],
"metrics": [
{
"cvssV4_0": {
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P",
"version": "4.0"
}
},
{
"cvssV3_1": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.1"
}
},
{
"cvssV3_0": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.0"
}
},
{
"cvssV2_0": {
"baseScore": 6.5,
"vectorString": "AV:N/AC:L/Au:S/C:P/I:P/A:P/E:POC/RL:ND/RC:UR",
"version": "2.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-918",
"description": "Server-Side Request Forgery",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-02-23T10:17:31.637Z",
"orgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"shortName": "VulDB"
},
"references": [
{
"name": "VDB-346451 | huggingface smolagents LocalPythonExecutor requests.post server-side request forgery",
"tags": [
"vdb-entry",
"technical-description"
],
"url": "https://vuldb.com/?id.346451"
},
{
"name": "VDB-346451 | CTI Indicators (IOB, IOC, IOA)",
"tags": [
"signature",
"permissions-required"
],
"url": "https://vuldb.com/?ctiid.346451"
},
{
"name": "Submit #752774 | Hugging Face smolagents 1.24.0 Server-Side Request Forgery",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.752774"
},
{
"tags": [
"related"
],
"url": "https://github.com/CH0ico/CVE_choco_smolagent/tree/main"
},
{
"tags": [
"exploit"
],
"url": "https://github.com/CH0ico/CVE_choco_smolagent/blob/main/report.md#proof-of-concept-execution"
}
],
"timeline": [
{
"lang": "en",
"time": "2026-02-18T00:00:00.000Z",
"value": "Advisory disclosed"
},
{
"lang": "en",
"time": "2026-02-18T01:00:00.000Z",
"value": "VulDB entry created"
},
{
"lang": "en",
"time": "2026-02-21T18:45:39.000Z",
"value": "VulDB entry last update"
}
],
"title": "huggingface smolagents LocalPythonExecutor requests.post server-side request forgery"
}
},
"cveMetadata": {
"assignerOrgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"assignerShortName": "VulDB",
"cveId": "CVE-2026-2654",
"datePublished": "2026-02-18T13:32:06.959Z",
"dateReserved": "2026-02-18T06:39:02.728Z",
"dateUpdated": "2026-02-23T10:17:31.637Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-0599 (GCVE-0-2026-0599)
Vulnerability from nvd – Published: 2026-02-02 10:36 – Updated: 2026-02-02 12:49
VLAI
Title
Unbounded External Image Fetch in Validation Leads to Resource-Exhaustion DoS in huggingface/text-generation-inference
Summary
A vulnerability in huggingface/text-generation-inference version 3.3.6 allows unauthenticated remote attackers to exploit unbounded external image fetching during input validation in VLM mode. The issue arises when the router scans inputs for Markdown image links and performs a blocking HTTP GET request, reading the entire response body into memory and cloning it before decoding. This behavior can lead to resource exhaustion, including network bandwidth saturation, memory inflation, and CPU overutilization. The vulnerability is triggered even if the request is later rejected for exceeding token limits. The default deployment configuration, which lacks memory usage limits and authentication, exacerbates the impact, potentially crashing the host machine. The issue is resolved in version 3.3.7.
Severity
7.5 (High)
SSVC
Exploitation: poc
Automatable: yes
Technical Impact: partial
CISA Coordinator (v2.0.3)
CWE
- CWE-400 - Uncontrolled Resource Consumption
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/text-generation-inference |
Affected:
unspecified , < 3.3.7
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-0599",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-02-02T12:48:08.191120Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-02-02T12:49:08.220Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/text-generation-inference",
"vendor": "huggingface",
"versions": [
{
"lessThan": "3.3.7",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in huggingface/text-generation-inference version 3.3.6 allows unauthenticated remote attackers to exploit unbounded external image fetching during input validation in VLM mode. The issue arises when the router scans inputs for Markdown image links and performs a blocking HTTP GET request, reading the entire response body into memory and cloning it before decoding. This behavior can lead to resource exhaustion, including network bandwidth saturation, memory inflation, and CPU overutilization. The vulnerability is triggered even if the request is later rejected for exceeding token limits. The default deployment configuration, which lacks memory usage limits and authentication, exacerbates the impact, potentially crashing the host machine. The issue is resolved in version 3.3.7."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-400",
"description": "CWE-400 Uncontrolled Resource Consumption",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-02-02T10:36:25.459Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/1d3f2085-666c-4441-b265-22f6f7d8d9cd"
},
{
"url": "https://github.com/huggingface/text-generation-inference/commit/24ee40d143d8d046039f12f76940a85886cbe152"
}
],
"source": {
"advisory": "1d3f2085-666c-4441-b265-22f6f7d8d9cd",
"discovery": "EXTERNAL"
},
"title": "Unbounded External Image Fetch in Validation Leads to Resource-Exhaustion DoS in huggingface/text-generation-inference"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-0599",
"datePublished": "2026-02-02T10:36:25.459Z",
"dateReserved": "2026-01-05T11:35:41.938Z",
"dateUpdated": "2026-02-02T12:49:08.220Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14930 (GCVE-0-2025-14930)
Vulnerability from nvd – Published: 2025-12-23 21:04 – Updated: 2025-12-26 14:32
VLAI
Title
Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability
Summary
Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the parsing of weights. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-28309.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.1
|
Date Public
2025-12-18 21:14
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14930",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-26T14:32:04.448613Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-26T14:32:13.294Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.1"
}
]
}
],
"dateAssigned": "2025-12-18T20:50:05.863Z",
"datePublic": "2025-12-18T21:14:55.064Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.\n\nThe specific flaw exists within the parsing of weights. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-28309."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502: Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:52.512Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1145",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1145/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Demeng Chen, and Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14930",
"datePublished": "2025-12-23T21:04:52.512Z",
"dateReserved": "2025-12-18T20:50:05.828Z",
"dateUpdated": "2025-12-26T14:32:13.294Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14928 (GCVE-0-2025-14928)
Vulnerability from nvd – Published: 2025-12-23 21:04 – Updated: 2025-12-26 14:31
VLAI
Title
Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability
Summary
Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.
The specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28253.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.0
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14928",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-26T14:31:22.533381Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-26T14:31:29.167Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.0"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:58.797Z",
"datePublic": "2025-12-18T21:15:04.691Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.\n\nThe specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28253."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:48.438Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1146",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1146/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14928",
"datePublished": "2025-12-23T21:04:48.438Z",
"dateReserved": "2025-12-18T20:49:58.765Z",
"dateUpdated": "2025-12-26T14:31:29.167Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14924 (GCVE-0-2025-14924)
Vulnerability from nvd – Published: 2025-12-23 21:04 – Updated: 2025-12-26 14:22
VLAI
Title
Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability
Summary
Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the parsing of checkpoints. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-27984.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
95faabf0a6cd845f4c5548697e288a79e424b096
|
Date Public
2025-12-18 21:14
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14924",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-26T14:22:20.584776Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-26T14:22:34.996Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "95faabf0a6cd845f4c5548697e288a79e424b096"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:41.206Z",
"datePublic": "2025-12-18T21:14:10.449Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.\n\nThe specific flaw exists within the parsing of checkpoints. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-27984."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502: Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:40.830Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1141",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1141/"
}
],
"source": {
"lang": "en",
"value": "Michael DePlante (@izobashi) of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14924",
"datePublished": "2025-12-23T21:04:40.830Z",
"dateReserved": "2025-12-18T20:49:41.182Z",
"dateUpdated": "2025-12-26T14:22:34.996Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14920 (GCVE-0-2025-14920)
Vulnerability from nvd – Published: 2025-12-23 21:04 – Updated: 2025-12-24 16:29
VLAI
Title
Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability
Summary
Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the parsing of model files. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-25423.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
9c8bd3fc1befe54f3efb9f385561eef49f060a70
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14920",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-24T16:28:51.404092Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-24T16:29:05.697Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "9c8bd3fc1befe54f3efb9f385561eef49f060a70"
}
]
}
],
"dateAssigned": "2025-12-18T20:43:16.331Z",
"datePublic": "2025-12-18T21:15:49.038Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.\n\nThe specific flaw exists within the parsing of model files. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-25423."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502: Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:36.809Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1150",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1150/"
}
],
"source": {
"lang": "en",
"value": "The_Kernel_Panic"
},
"title": "Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14920",
"datePublished": "2025-12-23T21:04:36.809Z",
"dateReserved": "2025-12-18T20:43:16.275Z",
"dateUpdated": "2025-12-24T16:29:05.697Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14926 (GCVE-0-2025-14926)
Vulnerability from nvd – Published: 2025-12-23 21:04 – Updated: 2025-12-24 16:26
VLAI
Title
Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability
Summary
Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.
The specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28251.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.0
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14926",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-24T16:26:36.372960Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-24T16:26:53.253Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.0"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:50.678Z",
"datePublic": "2025-12-18T21:15:14.466Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.\n\nThe specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28251."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:32.162Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1147",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1147/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14926",
"datePublished": "2025-12-23T21:04:32.162Z",
"dateReserved": "2025-12-18T20:49:50.656Z",
"dateUpdated": "2025-12-24T16:26:53.253Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14927 (GCVE-0-2025-14927)
Vulnerability from nvd – Published: 2025-12-23 21:04 – Updated: 2025-12-24 16:26
VLAI
Title
Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability
Summary
Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.
The specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user.
. Was ZDI-CAN-28252.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.0
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14927",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-24T16:25:45.983078Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-24T16:26:02.865Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.0"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:54.299Z",
"datePublic": "2025-12-18T21:15:26.091Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.\n\nThe specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user.\n\n. Was ZDI-CAN-28252."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:28.008Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1148",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1148/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14927",
"datePublished": "2025-12-23T21:04:28.008Z",
"dateReserved": "2025-12-18T20:49:54.276Z",
"dateUpdated": "2025-12-24T16:26:02.865Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-5241 (GCVE-0-2026-5241)
Vulnerability from cvelistv5 – Published: 2026-06-03 12:33 – Updated: 2026-06-03 15:45
VLAI
Title
Policy Bypass in LightGlue Nested Config Resolution in huggingface/transformers
Summary
A vulnerability in the LightGlue model loading path of huggingface/transformers version 5.2.0 allows an attacker-controlled model repository to execute arbitrary code during model initialization. The issue arises because the `trust_remote_code` parameter, intended to prevent remote code execution, is overridden by untrusted serialized configuration data in a nested code path. Specifically, when loading a LightGlue model using `AutoModel.from_pretrained()` with `trust_remote_code=False`, the `LightGlueConfig` reads the `trust_remote_code` value from the untrusted `config.json` file and propagates it into nested `AutoConfig.from_pretrained()` calls. This results in the execution of attacker-provided Python modules, even when the victim explicitly disables remote code execution. The vulnerability poses a high risk for environments such as API inference servers, research notebooks, CI/CD pipelines, and model evaluation workers, potentially leading to credential theft, lateral movement, or persistence/backdoor deployment.
Severity
SSVC
Exploitation: poc
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-829 - Inclusion of Functionality from Untrusted Control Sphere
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 5.5.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-5241",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-06-03T15:45:14.993177Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-06-03T15:45:36.684Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "5.5.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the LightGlue model loading path of huggingface/transformers version 5.2.0 allows an attacker-controlled model repository to execute arbitrary code during model initialization. The issue arises because the `trust_remote_code` parameter, intended to prevent remote code execution, is overridden by untrusted serialized configuration data in a nested code path. Specifically, when loading a LightGlue model using `AutoModel.from_pretrained()` with `trust_remote_code=False`, the `LightGlueConfig` reads the `trust_remote_code` value from the untrusted `config.json` file and propagates it into nested `AutoConfig.from_pretrained()` calls. This results in the execution of attacker-provided Python modules, even when the victim explicitly disables remote code execution. The vulnerability poses a high risk for environments such as API inference servers, research notebooks, CI/CD pipelines, and model evaluation workers, potentially leading to credential theft, lateral movement, or persistence/backdoor deployment."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "HIGH",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "CHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-829",
"description": "CWE-829 Inclusion of Functionality from Untrusted Control Sphere",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-06-03T12:33:10.227Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/ceb3ce1a-4c45-497a-b25e-cb9a7685e619"
},
{
"url": "https://github.com/huggingface/transformers/commit/676559d5022b74aaa0cee1cee0842b7f27c5320e"
}
],
"source": {
"advisory": "ceb3ce1a-4c45-497a-b25e-cb9a7685e619",
"discovery": "EXTERNAL"
},
"title": "Policy Bypass in LightGlue Nested Config Resolution in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-5241",
"datePublished": "2026-06-03T12:33:10.227Z",
"dateReserved": "2026-03-31T14:26:14.353Z",
"dateUpdated": "2026-06-03T15:45:36.684Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-4372 (GCVE-0-2026-4372)
Vulnerability from cvelistv5 – Published: 2026-05-24 13:40 – Updated: 2026-05-26 15:21
VLAI
Title
Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers
Summary
A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker's repository with the victim's full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitization of internal fields, and unsandboxed execution of downloaded kernels. The vulnerability bypasses the `trust_remote_code` security mechanism, is invisible to the victim, and exploits the standard documented usage pattern, making it particularly severe. Users are advised to upgrade to version 5.3.0 or later to mitigate this issue.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-1066 - Missing Serialization Control Element
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 5.3.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-4372",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-05-26T15:08:29.632933Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-05-26T15:21:02.376Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "5.3.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A critical remote code execution vulnerability exists in all versions of the HuggingFace transformers library prior to version 5.3.0. The vulnerability allows an attacker to craft a malicious `config.json` file containing the `_attn_implementation_internal` field set to an attacker-controlled HuggingFace Hub repository ID. When a victim loads this model using the standard `AutoModelForCausalLM.from_pretrained()` API, the library downloads and executes arbitrary Python code from the attacker\u0027s repository with the victim\u0027s full OS privileges. This issue arises due to unfiltered deserialization of configuration attributes, insufficient sanitization of internal fields, and unsandboxed execution of downloaded kernels. The vulnerability bypasses the `trust_remote_code` security mechanism, is invisible to the victim, and exploits the standard documented usage pattern, making it particularly severe. Users are advised to upgrade to version 5.3.0 or later to mitigate this issue."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 7.8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1066",
"description": "CWE-1066 Missing Serialization Control Element",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-05-24T13:40:40.995Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/1f693a6e-6836-4b8b-a0bd-ca036fba8884"
},
{
"url": "https://github.com/huggingface/transformers/commit/a7f8e7ff37d87d1a1a0c8cf607971c607741452f"
}
],
"source": {
"advisory": "1f693a6e-6836-4b8b-a0bd-ca036fba8884",
"discovery": "EXTERNAL"
},
"title": "Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-4372",
"datePublished": "2026-05-24T13:40:40.995Z",
"dateReserved": "2026-03-18T10:04:09.683Z",
"dateUpdated": "2026-05-26T15:21:02.376Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-44827 (GCVE-0-2026-44827)
Vulnerability from cvelistv5 – Published: 2026-05-14 16:33 – Updated: 2026-05-14 18:03
VLAI
Title
Diffusers: None.py Trust Remote Code Bypass
Summary
Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, diffusers 0.37.0 allows remote code execution without the trust_remote_code=True safeguard when loading pipelines from Hugging Face Hub repositories. The _resolve_custom_pipeline_and_cls function in pipeline_loading_utils.py performs string interpolation on the custom_pipeline parameter using f"{custom_pipeline}.py". When custom_pipeline is not supplied by the user, it defaults to None, which Python interpolates as the literal string "None.py". If an attacker publishes a Hub repository containing a file named None.py with a class that subclasses DiffusionPipeline, the file is automatically downloaded and executed during a standard DiffusionPipeline.from_pretrained() call with no additional keyword arguments. The trust_remote_code check in DiffusionPipeline.download() is bypassed because it evaluates custom_pipeline is not None as False (since the kwarg was never supplied), while the downstream code path that actually loads the module resolves the None value into a valid filename. An attacker can achieve silent arbitrary code execution by publishing a malicious model repository with a None.py file and a standard-looking model_index.json that references a legitimate pipeline class name, requiring only that a victim calls from_pretrained on the repository. This vulnerability is fixed in 0.38.0.
Severity
8.8 (High)
SSVC
Exploitation: poc
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://github.com/huggingface/diffusers/security… | x_refsource_CONFIRM |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | diffusers |
Affected:
< 0.38.0
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-44827",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-05-14T18:00:53.494731Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T18:03:35.120Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://github.com/huggingface/diffusers/security/advisories/GHSA-j7w6-vpvq-j3gm"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "diffusers",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "\u003c 0.38.0"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, diffusers 0.37.0 allows remote code execution without the trust_remote_code=True safeguard when loading pipelines from Hugging Face Hub repositories. The _resolve_custom_pipeline_and_cls function in pipeline_loading_utils.py performs string interpolation on the custom_pipeline parameter using f\"{custom_pipeline}.py\". When custom_pipeline is not supplied by the user, it defaults to None, which Python interpolates as the literal string \"None.py\". If an attacker publishes a Hub repository containing a file named None.py with a class that subclasses DiffusionPipeline, the file is automatically downloaded and executed during a standard DiffusionPipeline.from_pretrained() call with no additional keyword arguments. The trust_remote_code check in DiffusionPipeline.download() is bypassed because it evaluates custom_pipeline is not None as False (since the kwarg was never supplied), while the downstream code path that actually loads the module resolves the None value into a valid filename. An attacker can achieve silent arbitrary code execution by publishing a malicious model repository with a None.py file and a standard-looking model_index.json that references a legitimate pipeline class name, requiring only that a victim calls from_pretrained on the repository. This vulnerability is fixed in 0.38.0."
}
],
"metrics": [
{
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 8.8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T16:33:42.373Z",
"orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"shortName": "GitHub_M"
},
"references": [
{
"name": "https://github.com/huggingface/diffusers/security/advisories/GHSA-j7w6-vpvq-j3gm",
"tags": [
"x_refsource_CONFIRM"
],
"url": "https://github.com/huggingface/diffusers/security/advisories/GHSA-j7w6-vpvq-j3gm"
}
],
"source": {
"advisory": "GHSA-j7w6-vpvq-j3gm",
"discovery": "UNKNOWN"
},
"title": "Diffusers: None.py Trust Remote Code Bypass"
}
},
"cveMetadata": {
"assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"assignerShortName": "GitHub_M",
"cveId": "CVE-2026-44827",
"datePublished": "2026-05-14T16:33:42.373Z",
"dateReserved": "2026-05-07T21:21:48.351Z",
"dateUpdated": "2026-05-14T18:03:35.120Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-44513 (GCVE-0-2026-44513)
Vulnerability from cvelistv5 – Published: 2026-05-14 16:26 – Updated: 2026-05-14 19:51
VLAI
Title
Diffusers: `trust_remote_code` bypass via `custom_pipeline` and local custom components
Summary
Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, a trust_remote_code bypass in DiffusionPipeline.from_pretrained allows arbitrary remote code execution despite the user passing trust_remote_code=False (or omitting it, which is the default). The vulnerability has three variants, all sharing the same root cause — the trust_remote_code gate was implemented inside DiffusionPipeline.download() rather than at the actual dynamic-module load site, so any code path that bypassed or short-circuited download() also bypassed the security check. DiffusionPipeline.from_pretrained('repoA', custom_pipeline='attacker/repoB', trust_remote_code=False) — the gate evaluated against repoA's file list rather than repoB's, so repoB's pipeline.py was loaded and executed. DiffusionPipeline.from_pretrained('/local/snapshot', custom_pipeline='attacker/repoB', trust_remote_code=False) — the local-path branch never invoked download(), so the gate was never reached and remote code from repoB executed. DiffusionPipeline.from_pretrained('/local/snapshot', trust_remote_code=False) where the snapshot contains custom component files (e.g. unet/my_unet_model.py) referenced from model_index.json — same root cause; the local path skipped download() and custom component code executed. This vulnerability is fixed in 0.38.0.
Severity
8.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://github.com/huggingface/diffusers/security… | x_refsource_CONFIRM |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | diffusers |
Affected:
< 0.38.0
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-44513",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-05-14T17:38:51.150920Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T19:51:06.991Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "diffusers",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "\u003c 0.38.0"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "Diffusers is the a library for pretrained diffusion models. Prior to 0.38.0, a trust_remote_code bypass in DiffusionPipeline.from_pretrained allows arbitrary remote code execution despite the user passing trust_remote_code=False (or omitting it, which is the default). The vulnerability has three variants, all sharing the same root cause \u2014 the trust_remote_code gate was implemented inside DiffusionPipeline.download() rather than at the actual dynamic-module load site, so any code path that bypassed or short-circuited download() also bypassed the security check. DiffusionPipeline.from_pretrained(\u0027repoA\u0027, custom_pipeline=\u0027attacker/repoB\u0027, trust_remote_code=False) \u2014 the gate evaluated against repoA\u0027s file list rather than repoB\u0027s, so repoB\u0027s pipeline.py was loaded and executed. DiffusionPipeline.from_pretrained(\u0027/local/snapshot\u0027, custom_pipeline=\u0027attacker/repoB\u0027, trust_remote_code=False) \u2014 the local-path branch never invoked download(), so the gate was never reached and remote code from repoB executed. DiffusionPipeline.from_pretrained(\u0027/local/snapshot\u0027, trust_remote_code=False) where the snapshot contains custom component files (e.g. unet/my_unet_model.py) referenced from model_index.json \u2014 same root cause; the local path skipped download() and custom component code executed. This vulnerability is fixed in 0.38.0."
}
],
"metrics": [
{
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 8.8,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"integrityImpact": "HIGH",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.1"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-05-14T16:26:03.907Z",
"orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"shortName": "GitHub_M"
},
"references": [
{
"name": "https://github.com/huggingface/diffusers/security/advisories/GHSA-98h9-4798-4q5v",
"tags": [
"x_refsource_CONFIRM"
],
"url": "https://github.com/huggingface/diffusers/security/advisories/GHSA-98h9-4798-4q5v"
}
],
"source": {
"advisory": "GHSA-98h9-4798-4q5v",
"discovery": "UNKNOWN"
},
"title": "Diffusers: `trust_remote_code` bypass via `custom_pipeline` and local custom components"
}
},
"cveMetadata": {
"assignerOrgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
"assignerShortName": "GitHub_M",
"cveId": "CVE-2026-44513",
"datePublished": "2026-05-14T16:26:03.907Z",
"dateReserved": "2026-05-06T18:28:20.887Z",
"dateUpdated": "2026-05-14T19:51:06.991Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-25874 (GCVE-0-2026-25874)
Vulnerability from cvelistv5 – Published: 2026-04-23 19:45 – Updated: 2026-06-23 16:13 X_Open Source
VLAI
Title
LeRobot Unsafe Deserialization Remote Code Execution via gRPC
Summary
LeRobot through 0.5.1 contains an unsafe deserialization vulnerability in the async inference pipeline where pickle.loads() is used to deserialize data received over unauthenticated gRPC channels without TLS in the policy server and robot client components. An unauthenticated network-reachable attacker can achieve arbitrary code execution on the server or client by sending a crafted pickle payload through the SendPolicyInstructions, SendObservations, or GetActions gRPC calls.
Severity
SSVC
Exploitation: poc
Automatable: yes
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
5 references
| URL | Tags |
|---|---|
| https://chocapikk.com/posts/2026/lerobot-pickle-rce/ | exploit |
| https://github.com/huggingface/lerobot/issues/3047 | issue-tracking |
| https://github.com/huggingface/lerobot/pull/3048 | issue-tracking |
| https://github.com/huggingface/lerobot/issues/3134 | issue-tracking |
| https://www.vulncheck.com/advisories/lerobot-unsa… | third-party-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | LeRobot |
Affected:
0 , ≤ 0.5.1
(semver)
|
Date Public
2026-02-27 00:00
Credits
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-25874",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-04-24T14:34:03.307494Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-04-24T18:20:13.815Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "LeRobot",
"repo": "https://github.com/huggingface/lerobot",
"vendor": "Hugging Face",
"versions": [
{
"lessThanOrEqual": "0.5.1",
"status": "affected",
"version": "0",
"versionType": "semver"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "finder",
"value": "Valentin Lobstein (Chocapikk)"
}
],
"datePublic": "2026-02-27T00:00:00.000Z",
"descriptions": [
{
"lang": "en",
"value": "LeRobot through 0.5.1 contains an unsafe deserialization vulnerability in the async inference pipeline where pickle.loads() is used to deserialize data received over unauthenticated gRPC channels without TLS in the policy server and robot client components. An unauthenticated network-reachable attacker can achieve arbitrary code execution on the server or client by sending a crafted pickle payload through the SendPolicyInstructions, SendObservations, or GetActions gRPC calls."
}
],
"metrics": [
{
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 9.3,
"baseSeverity": "CRITICAL",
"exploitMaturity": "NOT_DEFINED",
"privilegesRequired": "NONE",
"providerUrgency": "NOT_DEFINED",
"subAvailabilityImpact": "NONE",
"subConfidentialityImpact": "NONE",
"subIntegrityImpact": "NONE",
"userInteraction": "NONE",
"valueDensity": "NOT_DEFINED",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
"version": "4.0",
"vulnAvailabilityImpact": "HIGH",
"vulnConfidentialityImpact": "HIGH",
"vulnIntegrityImpact": "HIGH",
"vulnerabilityResponseEffort": "NOT_DEFINED"
},
"format": "CVSS",
"scenarios": [
{
"lang": "en",
"value": "GENERAL"
}
]
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-06-23T16:13:56.651Z",
"orgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"shortName": "VulnCheck"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://chocapikk.com/posts/2026/lerobot-pickle-rce/"
},
{
"tags": [
"issue-tracking"
],
"url": "https://github.com/huggingface/lerobot/issues/3047"
},
{
"tags": [
"issue-tracking"
],
"url": "https://github.com/huggingface/lerobot/pull/3048"
},
{
"tags": [
"issue-tracking"
],
"url": "https://github.com/huggingface/lerobot/issues/3134"
},
{
"tags": [
"third-party-advisory"
],
"url": "https://www.vulncheck.com/advisories/lerobot-unsafe-deserialization-remote-code-execution-via-grpc"
}
],
"source": {
"discovery": "EXTERNAL"
},
"tags": [
"x_open-source"
],
"title": "LeRobot Unsafe Deserialization Remote Code Execution via gRPC",
"x_generator": {
"engine": "Vulnogram 1.0.1"
}
}
},
"cveMetadata": {
"assignerOrgId": "83251b91-4cc7-4094-a5c7-464a1b83ea10",
"assignerShortName": "VulnCheck",
"cveId": "CVE-2026-25874",
"datePublished": "2026-04-23T19:45:01.090Z",
"dateReserved": "2026-02-06T19:12:03.464Z",
"dateUpdated": "2026-06-23T16:13:56.651Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-1839 (GCVE-0-2026-1839)
Vulnerability from cvelistv5 – Published: 2026-04-07 05:22 – Updated: 2026-04-07 13:27
VLAI
Title
Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers
Summary
A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch>=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3.
Severity
6.5 (Medium)
SSVC
Exploitation: poc
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < v5.0.0rc3
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-1839",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-04-07T13:27:38.099805Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-04-07T13:27:41.789Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/3c77bb97-e493-493d-9a88-c57f5c536485"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "v5.0.0rc3",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch\u003e=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "HIGH",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 6.5,
"baseSeverity": "MEDIUM",
"confidentialityImpact": "HIGH",
"integrityImpact": "LOW",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:L/A:H",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502 Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-04-07T05:22:00.959Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/3c77bb97-e493-493d-9a88-c57f5c536485"
},
{
"url": "https://github.com/huggingface/transformers/commit/03c8082ba4594c9b8d6fe190ca9bed0e5f8ca396"
}
],
"source": {
"advisory": "3c77bb97-e493-493d-9a88-c57f5c536485",
"discovery": "EXTERNAL"
},
"title": "Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-1839",
"datePublished": "2026-04-07T05:22:00.959Z",
"dateReserved": "2026-02-03T16:49:27.781Z",
"dateUpdated": "2026-04-07T13:27:41.789Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-4963 (GCVE-0-2026-4963)
Vulnerability from cvelistv5 – Published: 2026-03-27 17:05 – Updated: 2026-03-31 13:32
VLAI
Title
huggingface smolagents Incomplete Fix CVE-2025-9959 local_python_executor.py evaluate_with code injection
Summary
A weakness has been identified in huggingface smolagents 1.25.0.dev0. This affects the function evaluate_augassign/evaluate_call/evaluate_with of the file src/smolagents/local_python_executor.py of the component Incomplete Fix CVE-2025-9959. This manipulation causes code injection. It is possible to initiate the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
Severity
SSVC
Exploitation: poc
Automatable: no
Technical Impact: partial
CISA Coordinator (v2.0.3)
Assigner
References
7 references
| URL | Tags |
|---|---|
| https://vuldb.com/?id.353840 | vdb-entrytechnical-description |
| https://vuldb.com/?ctiid.353840 | signaturepermissions-required |
| https://vuldb.com/?submit.777623 | third-party-advisory |
| https://vuldb.com/?submit.777643 | third-party-advisory |
| https://vuldb.com/?submit.777644 | third-party-advisory |
| https://gist.github.com/YLChen-007/7146f45960f79b… | related |
| https://gist.github.com/YLChen-007/35b7d46e892266… | exploit |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | smolagents |
Affected:
1.25.0.dev0
cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:* |
Credits
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-4963",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-03-31T13:32:07.450819Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-03-31T13:32:41.488Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"cpes": [
"cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:*"
],
"modules": [
"Incomplete Fix CVE-2025-9959"
],
"product": "smolagents",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "1.25.0.dev0"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Eric-z (VulDB User)"
},
{
"lang": "en",
"type": "coordinator",
"value": "VulDB"
}
],
"descriptions": [
{
"lang": "en",
"value": "A weakness has been identified in huggingface smolagents 1.25.0.dev0. This affects the function evaluate_augassign/evaluate_call/evaluate_with of the file src/smolagents/local_python_executor.py of the component Incomplete Fix CVE-2025-9959. This manipulation causes code injection. It is possible to initiate the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way."
}
],
"metrics": [
{
"cvssV4_0": {
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P",
"version": "4.0"
}
},
{
"cvssV3_1": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.1"
}
},
{
"cvssV3_0": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.0"
}
},
{
"cvssV2_0": {
"baseScore": 7.5,
"vectorString": "AV:N/AC:L/Au:N/C:P/I:P/A:P/E:POC/RL:ND/RC:UR",
"version": "2.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "Code Injection",
"lang": "en",
"type": "CWE"
}
]
},
{
"descriptions": [
{
"cweId": "CWE-74",
"description": "Injection",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-03-27T22:07:35.402Z",
"orgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"shortName": "VulDB"
},
"references": [
{
"name": "VDB-353840 | huggingface smolagents Incomplete Fix CVE-2025-9959 local_python_executor.py evaluate_with code injection",
"tags": [
"vdb-entry",
"technical-description"
],
"url": "https://vuldb.com/?id.353840"
},
{
"name": "VDB-353840 | CTI Indicators (IOB, IOC, TTP, IOA)",
"tags": [
"signature",
"permissions-required"
],
"url": "https://vuldb.com/?ctiid.353840"
},
{
"name": "Submit #777623 | HuggingFace smolagents 1.25.0.dev0 CWE-693",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.777623"
},
{
"name": "Submit #777643 | HuggingFace smolagents 1.25.0.dev0 CWE-693 (Duplicate)",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.777643"
},
{
"name": "Submit #777644 | HuggingFace smolagents 1.25.0.dev0 CWE-693 (Duplicate)",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.777644"
},
{
"tags": [
"related"
],
"url": "https://gist.github.com/YLChen-007/7146f45960f79bc1e2976fed526e0a9b"
},
{
"tags": [
"exploit"
],
"url": "https://gist.github.com/YLChen-007/35b7d46e892266a0ed6dbe57802858be"
}
],
"timeline": [
{
"lang": "en",
"time": "2026-03-27T00:00:00.000Z",
"value": "Advisory disclosed"
},
{
"lang": "en",
"time": "2026-03-27T01:00:00.000Z",
"value": "VulDB entry created"
},
{
"lang": "en",
"time": "2026-03-27T09:23:31.000Z",
"value": "VulDB entry last update"
}
],
"title": "huggingface smolagents Incomplete Fix CVE-2025-9959 local_python_executor.py evaluate_with code injection"
}
},
"cveMetadata": {
"assignerOrgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"assignerShortName": "VulDB",
"cveId": "CVE-2026-4963",
"datePublished": "2026-03-27T17:05:20.096Z",
"dateReserved": "2026-03-27T08:17:46.710Z",
"dateUpdated": "2026-03-31T13:32:41.488Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-2654 (GCVE-0-2026-2654)
Vulnerability from cvelistv5 – Published: 2026-02-18 13:32 – Updated: 2026-02-23 10:17
VLAI
Title
huggingface smolagents LocalPythonExecutor requests.post server-side request forgery
Summary
A weakness has been identified in huggingface smolagents 1.24.0. Impacted is the function requests.get/requests.post of the component LocalPythonExecutor. Executing a manipulation can lead to server-side request forgery. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way.
Severity
SSVC
Exploitation: poc
Automatable: no
Technical Impact: partial
CISA Coordinator (v2.0.3)
CWE
- CWE-918 - Server-Side Request Forgery
Assigner
References
5 references
| URL | Tags |
|---|---|
| https://vuldb.com/?id.346451 | vdb-entrytechnical-description |
| https://vuldb.com/?ctiid.346451 | signaturepermissions-required |
| https://vuldb.com/?submit.752774 | third-party-advisory |
| https://github.com/CH0ico/CVE_choco_smolagent/tree/main | related |
| https://github.com/CH0ico/CVE_choco_smolagent/blo… | exploit |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | smolagents |
Affected:
1.24.0
cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:* |
Credits
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-2654",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-02-18T14:10:54.671299Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-02-18T14:11:19.055Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"cpes": [
"cpe:2.3:a:huggingface:smolagents:*:*:*:*:*:*:*:*"
],
"modules": [
"LocalPythonExecutor"
],
"product": "smolagents",
"vendor": "huggingface",
"versions": [
{
"status": "affected",
"version": "1.24.0"
}
]
}
],
"credits": [
{
"lang": "en",
"type": "reporter",
"value": "Choco094late (VulDB User)"
}
],
"descriptions": [
{
"lang": "en",
"value": "A weakness has been identified in huggingface smolagents 1.24.0. Impacted is the function requests.get/requests.post of the component LocalPythonExecutor. Executing a manipulation can lead to server-side request forgery. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. The vendor was contacted early about this disclosure but did not respond in any way."
}
],
"metrics": [
{
"cvssV4_0": {
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P",
"version": "4.0"
}
},
{
"cvssV3_1": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.1"
}
},
{
"cvssV3_0": {
"baseScore": 6.3,
"baseSeverity": "MEDIUM",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R",
"version": "3.0"
}
},
{
"cvssV2_0": {
"baseScore": 6.5,
"vectorString": "AV:N/AC:L/Au:S/C:P/I:P/A:P/E:POC/RL:ND/RC:UR",
"version": "2.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-918",
"description": "Server-Side Request Forgery",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-02-23T10:17:31.637Z",
"orgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"shortName": "VulDB"
},
"references": [
{
"name": "VDB-346451 | huggingface smolagents LocalPythonExecutor requests.post server-side request forgery",
"tags": [
"vdb-entry",
"technical-description"
],
"url": "https://vuldb.com/?id.346451"
},
{
"name": "VDB-346451 | CTI Indicators (IOB, IOC, IOA)",
"tags": [
"signature",
"permissions-required"
],
"url": "https://vuldb.com/?ctiid.346451"
},
{
"name": "Submit #752774 | Hugging Face smolagents 1.24.0 Server-Side Request Forgery",
"tags": [
"third-party-advisory"
],
"url": "https://vuldb.com/?submit.752774"
},
{
"tags": [
"related"
],
"url": "https://github.com/CH0ico/CVE_choco_smolagent/tree/main"
},
{
"tags": [
"exploit"
],
"url": "https://github.com/CH0ico/CVE_choco_smolagent/blob/main/report.md#proof-of-concept-execution"
}
],
"timeline": [
{
"lang": "en",
"time": "2026-02-18T00:00:00.000Z",
"value": "Advisory disclosed"
},
{
"lang": "en",
"time": "2026-02-18T01:00:00.000Z",
"value": "VulDB entry created"
},
{
"lang": "en",
"time": "2026-02-21T18:45:39.000Z",
"value": "VulDB entry last update"
}
],
"title": "huggingface smolagents LocalPythonExecutor requests.post server-side request forgery"
}
},
"cveMetadata": {
"assignerOrgId": "1af790b2-7ee1-4545-860a-a788eba489b5",
"assignerShortName": "VulDB",
"cveId": "CVE-2026-2654",
"datePublished": "2026-02-18T13:32:06.959Z",
"dateReserved": "2026-02-18T06:39:02.728Z",
"dateUpdated": "2026-02-23T10:17:31.637Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2026-0599 (GCVE-0-2026-0599)
Vulnerability from cvelistv5 – Published: 2026-02-02 10:36 – Updated: 2026-02-02 12:49
VLAI
Title
Unbounded External Image Fetch in Validation Leads to Resource-Exhaustion DoS in huggingface/text-generation-inference
Summary
A vulnerability in huggingface/text-generation-inference version 3.3.6 allows unauthenticated remote attackers to exploit unbounded external image fetching during input validation in VLM mode. The issue arises when the router scans inputs for Markdown image links and performs a blocking HTTP GET request, reading the entire response body into memory and cloning it before decoding. This behavior can lead to resource exhaustion, including network bandwidth saturation, memory inflation, and CPU overutilization. The vulnerability is triggered even if the request is later rejected for exceeding token limits. The default deployment configuration, which lacks memory usage limits and authentication, exacerbates the impact, potentially crashing the host machine. The issue is resolved in version 3.3.7.
Severity
7.5 (High)
SSVC
Exploitation: poc
Automatable: yes
Technical Impact: partial
CISA Coordinator (v2.0.3)
CWE
- CWE-400 - Uncontrolled Resource Consumption
Assigner
References
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | huggingface/text-generation-inference |
Affected:
unspecified , < 3.3.7
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2026-0599",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2026-02-02T12:48:08.191120Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2026-02-02T12:49:08.220Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/text-generation-inference",
"vendor": "huggingface",
"versions": [
{
"lessThan": "3.3.7",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in huggingface/text-generation-inference version 3.3.6 allows unauthenticated remote attackers to exploit unbounded external image fetching during input validation in VLM mode. The issue arises when the router scans inputs for Markdown image links and performs a blocking HTTP GET request, reading the entire response body into memory and cloning it before decoding. This behavior can lead to resource exhaustion, including network bandwidth saturation, memory inflation, and CPU overutilization. The vulnerability is triggered even if the request is later rejected for exceeding token limits. The default deployment configuration, which lacks memory usage limits and authentication, exacerbates the impact, potentially crashing the host machine. The issue is resolved in version 3.3.7."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-400",
"description": "CWE-400 Uncontrolled Resource Consumption",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2026-02-02T10:36:25.459Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/1d3f2085-666c-4441-b265-22f6f7d8d9cd"
},
{
"url": "https://github.com/huggingface/text-generation-inference/commit/24ee40d143d8d046039f12f76940a85886cbe152"
}
],
"source": {
"advisory": "1d3f2085-666c-4441-b265-22f6f7d8d9cd",
"discovery": "EXTERNAL"
},
"title": "Unbounded External Image Fetch in Validation Leads to Resource-Exhaustion DoS in huggingface/text-generation-inference"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2026-0599",
"datePublished": "2026-02-02T10:36:25.459Z",
"dateReserved": "2026-01-05T11:35:41.938Z",
"dateUpdated": "2026-02-02T12:49:08.220Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14930 (GCVE-0-2025-14930)
Vulnerability from cvelistv5 – Published: 2025-12-23 21:04 – Updated: 2025-12-26 14:32
VLAI
Title
Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability
Summary
Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the parsing of weights. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-28309.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.1
|
Date Public
2025-12-18 21:14
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14930",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-26T14:32:04.448613Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-26T14:32:13.294Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.1"
}
]
}
],
"dateAssigned": "2025-12-18T20:50:05.863Z",
"datePublic": "2025-12-18T21:14:55.064Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.\n\nThe specific flaw exists within the parsing of weights. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-28309."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502: Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:52.512Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1145",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1145/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Demeng Chen, and Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers GLM4 Deserialization of Untrusted Data Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14930",
"datePublished": "2025-12-23T21:04:52.512Z",
"dateReserved": "2025-12-18T20:50:05.828Z",
"dateUpdated": "2025-12-26T14:32:13.294Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14928 (GCVE-0-2025-14928)
Vulnerability from cvelistv5 – Published: 2025-12-23 21:04 – Updated: 2025-12-26 14:31
VLAI
Title
Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability
Summary
Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.
The specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28253.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.0
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14928",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-26T14:31:22.533381Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-26T14:31:29.167Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.0"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:58.797Z",
"datePublic": "2025-12-18T21:15:04.691Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.\n\nThe specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28253."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:48.438Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1146",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1146/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers HuBERT convert_config Code Injection Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14928",
"datePublished": "2025-12-23T21:04:48.438Z",
"dateReserved": "2025-12-18T20:49:58.765Z",
"dateUpdated": "2025-12-26T14:31:29.167Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14924 (GCVE-0-2025-14924)
Vulnerability from cvelistv5 – Published: 2025-12-23 21:04 – Updated: 2025-12-26 14:22
VLAI
Title
Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability
Summary
Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the parsing of checkpoints. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-27984.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
95faabf0a6cd845f4c5548697e288a79e424b096
|
Date Public
2025-12-18 21:14
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14924",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-26T14:22:20.584776Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-26T14:22:34.996Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "95faabf0a6cd845f4c5548697e288a79e424b096"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:41.206Z",
"datePublic": "2025-12-18T21:14:10.449Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.\n\nThe specific flaw exists within the parsing of checkpoints. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-27984."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502: Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:40.830Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1141",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1141/"
}
],
"source": {
"lang": "en",
"value": "Michael DePlante (@izobashi) of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers megatron_gpt2 Deserialization of Untrusted Data Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14924",
"datePublished": "2025-12-23T21:04:40.830Z",
"dateReserved": "2025-12-18T20:49:41.182Z",
"dateUpdated": "2025-12-26T14:22:34.996Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14920 (GCVE-0-2025-14920)
Vulnerability from cvelistv5 – Published: 2025-12-23 21:04 – Updated: 2025-12-24 16:29
VLAI
Title
Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability
Summary
Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.
The specific flaw exists within the parsing of model files. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-25423.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-502 - Deserialization of Untrusted Data
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
9c8bd3fc1befe54f3efb9f385561eef49f060a70
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14920",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-24T16:28:51.404092Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-24T16:29:05.697Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "9c8bd3fc1befe54f3efb9f385561eef49f060a70"
}
]
}
],
"dateAssigned": "2025-12-18T20:43:16.331Z",
"datePublic": "2025-12-18T21:15:49.038Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file.\n\nThe specific flaw exists within the parsing of model files. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-25423."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-502",
"description": "CWE-502: Deserialization of Untrusted Data",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:36.809Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1150",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1150/"
}
],
"source": {
"lang": "en",
"value": "The_Kernel_Panic"
},
"title": "Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14920",
"datePublished": "2025-12-23T21:04:36.809Z",
"dateReserved": "2025-12-18T20:43:16.275Z",
"dateUpdated": "2025-12-24T16:29:05.697Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14926 (GCVE-0-2025-14926)
Vulnerability from cvelistv5 – Published: 2025-12-23 21:04 – Updated: 2025-12-24 16:26
VLAI
Title
Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability
Summary
Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.
The specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28251.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.0
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14926",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-24T16:26:36.372960Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-24T16:26:53.253Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.0"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:50.678Z",
"datePublic": "2025-12-18T21:15:14.466Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.\n\nThe specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user. Was ZDI-CAN-28251."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:32.162Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1147",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1147/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers SEW convert_config Code Injection Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14926",
"datePublished": "2025-12-23T21:04:32.162Z",
"dateReserved": "2025-12-18T20:49:50.656Z",
"dateUpdated": "2025-12-24T16:26:53.253Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}
CVE-2025-14927 (GCVE-0-2025-14927)
Vulnerability from cvelistv5 – Published: 2025-12-23 21:04 – Updated: 2025-12-24 16:26
VLAI
Title
Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability
Summary
Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.
The specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user.
. Was ZDI-CAN-28252.
Severity
7.8 (High)
SSVC
Exploitation: none
Automatable: no
Technical Impact: total
CISA Coordinator (v2.0.3)
CWE
- CWE-94 - Improper Control of Generation of Code ('Code Injection')
Assigner
References
1 reference
| URL | Tags |
|---|---|
| https://www.zerodayinitiative.com/advisories/ZDI-… | x_research-advisory |
Impacted products
1 product
| Vendor | Product | Version | |
|---|---|---|---|
| Hugging Face | Transformers |
Affected:
4.57.0
|
Date Public
2025-12-18 21:15
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-14927",
"options": [
{
"Exploitation": "none"
},
{
"Automatable": "no"
},
{
"Technical Impact": "total"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-12-24T16:25:45.983078Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-12-24T16:26:02.865Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"defaultStatus": "unknown",
"product": "Transformers",
"vendor": "Hugging Face",
"versions": [
{
"status": "affected",
"version": "4.57.0"
}
]
}
],
"dateAssigned": "2025-12-18T20:49:54.299Z",
"datePublic": "2025-12-18T21:15:26.091Z",
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face Transformers. User interaction is required to exploit this vulnerability in that the target must convert a malicious checkpoint.\n\nThe specific flaw exists within the convert_config function. The issue results from the lack of proper validation of a user-supplied string before using it to execute Python code. An attacker can leverage this vulnerability to execute code in the context of the current user.\n\n. Was ZDI-CAN-28252."
}
],
"metrics": [
{
"cvssV3_0": {
"baseScore": 7.8,
"baseSeverity": "HIGH",
"vectorString": "CVSS:3.0/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"format": "CVSS"
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-94",
"description": "CWE-94: Improper Control of Generation of Code (\u0027Code Injection\u0027)",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-12-23T21:04:28.008Z",
"orgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"shortName": "zdi"
},
"references": [
{
"name": "ZDI-25-1148",
"tags": [
"x_research-advisory"
],
"url": "https://www.zerodayinitiative.com/advisories/ZDI-25-1148/"
}
],
"source": {
"lang": "en",
"value": "Peter Girnus (@gothburz), Brandon Niemczyk of Trend Zero Day Initiative"
},
"title": "Hugging Face Transformers SEW-D convert_config Code Injection Remote Code Execution Vulnerability"
}
},
"cveMetadata": {
"assignerOrgId": "99f1926a-a320-47d8-bbb5-42feb611262e",
"assignerShortName": "zdi",
"cveId": "CVE-2025-14927",
"datePublished": "2025-12-23T21:04:28.008Z",
"dateReserved": "2025-12-18T20:49:54.276Z",
"dateUpdated": "2025-12-24T16:26:02.865Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.2"
}