Search criteria
60 vulnerabilities found for transformers by huggingface
FKIE_CVE-2025-6921
Vulnerability from fkie_nvd - Published: 2025-09-23 14:15 - Updated: 2025-10-10 21:31
Severity ?
Summary
The huggingface/transformers library, versions prior to 4.53.0, is vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer. The vulnerability arises from the _do_use_weight_decay method, which processes user-controlled regular expressions in the include_in_weight_decay and exclude_from_weight_decay lists. Malicious regular expressions can cause catastrophic backtracking during the re.search call, leading to 100% CPU utilization and a denial of service. This issue can be exploited by attackers who can control the patterns in these lists, potentially causing the machine learning task to hang and rendering services unresponsive.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "028A5B9B-4F6A-41A1-8A94-8F904334B237",
"versionEndExcluding": "4.53.0",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "The huggingface/transformers library, versions prior to 4.53.0, is vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer. The vulnerability arises from the _do_use_weight_decay method, which processes user-controlled regular expressions in the include_in_weight_decay and exclude_from_weight_decay lists. Malicious regular expressions can cause catastrophic backtracking during the re.search call, leading to 100% CPU utilization and a denial of service. This issue can be exploited by attackers who can control the patterns in these lists, potentially causing the machine learning task to hang and rendering services unresponsive."
}
],
"id": "CVE-2025-6921",
"lastModified": "2025-10-10T21:31:30.230",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
],
"cvssMetricV31": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"version": "3.1"
},
"exploitabilityScore": 3.9,
"impactScore": 3.6,
"source": "nvd@nist.gov",
"type": "Primary"
}
]
},
"published": "2025-09-23T14:15:41.387",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/287d15a7-6e7c-45d2-8c05-11e305776f1f"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-400"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
FKIE_CVE-2025-6051
Vulnerability from fkie_nvd - Published: 2025-09-14 17:15 - Updated: 2025-10-21 14:16
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method's handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities.
References
| URL | Tags | ||
|---|---|---|---|
| security@huntr.dev | https://github.com/huggingface/transformers/commit/ba8eaba9865618253f997784aa565b96206426f0 | Patch | |
| security@huntr.dev | https://huntr.com/bounties/af929523-7b59-418a-bf55-301830b2ac9d | Exploit, Third Party Advisory, Issue Tracking, Patch |
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | 4.52.4 |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:4.52.4:*:*:*:*:*:*:*",
"matchCriteriaId": "BA362206-8A09-4249-A522-427C8793D241",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method\u0027s handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities."
}
],
"id": "CVE-2025-6051",
"lastModified": "2025-10-21T14:16:24.770",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
]
},
"published": "2025-09-14T17:15:34.210",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/ba8eaba9865618253f997784aa565b96206426f0"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory",
"Issue Tracking",
"Patch"
],
"url": "https://huntr.com/bounties/af929523-7b59-418a-bf55-301830b2ac9d"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
FKIE_CVE-2025-6638
Vulnerability from fkie_nvd - Published: 2025-09-12 11:15 - Updated: 2025-10-21 13:33
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer's `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | 4.52.4 |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:4.52.4:*:*:*:*:*:*:*",
"matchCriteriaId": "BA362206-8A09-4249-A522-427C8793D241",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer\u0027s `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service."
}
],
"id": "CVE-2025-6638",
"lastModified": "2025-10-21T13:33:08.580",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
],
"cvssMetricV31": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"version": "3.1"
},
"exploitabilityScore": 3.9,
"impactScore": 3.6,
"source": "nvd@nist.gov",
"type": "Primary"
}
]
},
"published": "2025-09-12T11:15:31.770",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/6a6c933f-9ce8-4ded-8b3b-2c1444c61f36"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
FKIE_CVE-2025-5197
Vulnerability from fkie_nvd - Published: 2025-08-06 12:15 - Updated: 2025-10-21 16:46
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats.
References
| URL | Tags | ||
|---|---|---|---|
| security@huntr.dev | https://github.com/huggingface/transformers/commit/944b56000be5e9b61af8301aa340838770ad8a0b | Patch | |
| security@huntr.dev | https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf | Exploit, Third Party Advisory, Issue Tracking, Patch | |
| 134c704f-9b21-4f2e-91b3-4a467353bcc0 | https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf | Exploit, Third Party Advisory, Issue Tracking, Patch |
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "028A5B9B-4F6A-41A1-8A94-8F904334B237",
"versionEndExcluding": "4.53.0",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats."
},
{
"lang": "es",
"value": "Existe una vulnerabilidad de denegaci\u00f3n de servicio de expresiones regulares (ReDoS) en la librer\u00eda Hugging Face Transformers, espec\u00edficamente en la funci\u00f3n `convert_tf_weight_name_to_pt_weight_name()`. Esta funci\u00f3n, responsable de convertir los nombres de peso de TensorFlow al formato de PyTorch, utiliza un patr\u00f3n de expresiones regulares `/[^/]*___([^/]*)/` que puede explotarse para causar un consumo excesivo de CPU mediante cadenas de entrada manipuladas debido a un retroceso catastr\u00f3fico. La vulnerabilidad afecta a las versiones hasta la 4.51.3 y se corrigi\u00f3 en la versi\u00f3n 4.53.0. Este problema puede provocar interrupciones del servicio, agotamiento de recursos y posibles vulnerabilidades del servicio API, lo que afecta los procesos de conversi\u00f3n de modelos entre los formatos de TensorFlow y PyTorch."
}
],
"id": "CVE-2025-5197",
"lastModified": "2025-10-21T16:46:13.380",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
]
},
"published": "2025-08-06T12:15:26.837",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/944b56000be5e9b61af8301aa340838770ad8a0b"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory",
"Issue Tracking",
"Patch"
],
"url": "https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf"
},
{
"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"tags": [
"Exploit",
"Third Party Advisory",
"Issue Tracking",
"Patch"
],
"url": "https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Secondary"
}
]
}
FKIE_CVE-2025-3933
Vulnerability from fkie_nvd - Published: 2025-07-11 10:15 - Updated: 2025-08-07 01:01
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the DonutProcessor class's `token2json()` method. This vulnerability affects versions 4.50.3 and earlier, and is fixed in version 4.52.1. The issue arises from the regex pattern `<s_(.*?)>` which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting document processing tasks using the Donut model.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "603CA36F-1ABE-4950-8FC3-CCB444106044",
"versionEndExcluding": "4.52.1",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the DonutProcessor class\u0027s `token2json()` method. This vulnerability affects versions 4.50.3 and earlier, and is fixed in version 4.52.1. The issue arises from the regex pattern `\u003cs_(.*?)\u003e` which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting document processing tasks using the Donut model."
},
{
"lang": "es",
"value": "Se descubri\u00f3 una vulnerabilidad de denegaci\u00f3n de servicio por expresi\u00f3n regular (ReDoS) en la librer\u00eda Hugging Face Transformers, espec\u00edficamente en el m\u00e9todo `token2json()` de la clase DonutProcessor. Esta vulnerabilidad afecta a las versiones 4.50.3 y anteriores, y se corrigi\u00f3 en la versi\u00f3n 4.52.1. El problema surge del patr\u00f3n de expresiones regulares ``, que puede explotarse para causar un consumo excesivo de CPU mediante cadenas de entrada manipuladas debido a un retroceso catastr\u00f3fico. Esta vulnerabilidad puede provocar interrupciones del servicio, agotamiento de recursos y posibles vulnerabilidades del servicio API, lo que afecta las tareas de procesamiento de documentos que utilizan el modelo Donut."
}
],
"id": "CVE-2025-3933",
"lastModified": "2025-08-07T01:01:46.780",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
],
"cvssMetricV31": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L",
"version": "3.1"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "nvd@nist.gov",
"type": "Primary"
}
]
},
"published": "2025-07-11T10:15:22.293",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/ebbe9b12dd75b69f92100d684c47f923ee262a93"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/25282953-5827-4384-bb6f-5790d275721b"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
FKIE_CVE-2025-3777
Vulnerability from fkie_nvd - Published: 2025-07-07 10:15 - Updated: 2025-08-07 00:54
Severity ?
Summary
Hugging Face Transformers versions up to 4.49.0 are affected by an improper input validation vulnerability in the `image_utils.py` file. The vulnerability arises from insecure URL validation using the `startswith()` method, which can be bypassed through URL username injection. This allows attackers to craft URLs that appear to be from YouTube but resolve to malicious domains, potentially leading to phishing attacks, malware distribution, or data exfiltration. The issue is fixed in version 4.52.1.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "603CA36F-1ABE-4950-8FC3-CCB444106044",
"versionEndExcluding": "4.52.1",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers versions up to 4.49.0 are affected by an improper input validation vulnerability in the `image_utils.py` file. The vulnerability arises from insecure URL validation using the `startswith()` method, which can be bypassed through URL username injection. This allows attackers to craft URLs that appear to be from YouTube but resolve to malicious domains, potentially leading to phishing attacks, malware distribution, or data exfiltration. The issue is fixed in version 4.52.1."
},
{
"lang": "es",
"value": "Las versiones de Hugging Face Transformers hasta la 4.49.0 se ven afectadas por una vulnerabilidad de validaci\u00f3n de entrada incorrecta en el archivo `image_utils.py`. Esta vulnerabilidad se debe a una validaci\u00f3n de URL insegura mediante el m\u00e9todo `startswith()`, que puede eludirse mediante la inyecci\u00f3n de nombres de usuario en la URL. Esto permite a los atacantes manipular URL que parecen provenir de YouTube, pero que redirigen a dominios maliciosos, lo que podr\u00eda provocar ataques de phishing, distribuci\u00f3n de malware o exfiltraci\u00f3n de datos. El problema se ha corregido en la versi\u00f3n 4.52.1."
}
],
"id": "CVE-2025-3777",
"lastModified": "2025-08-07T00:54:16.957",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 3.5,
"baseSeverity": "LOW",
"confidentialityImpact": "LOW",
"integrityImpact": "NONE",
"privilegesRequired": "LOW",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:L/UI:R/S:U/C:L/I:N/A:N",
"version": "3.0"
},
"exploitabilityScore": 2.1,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
]
},
"published": "2025-07-07T10:15:28.297",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/4dda5f71b35fb70cf602187eef84bb17a50b9082"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/ccba0730-9248-4853-b7ff-5c20e6364f09"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-20"
}
],
"source": "security@huntr.dev",
"type": "Primary"
},
{
"description": [
{
"lang": "en",
"value": "NVD-CWE-noinfo"
}
],
"source": "nvd@nist.gov",
"type": "Primary"
}
]
}
FKIE_CVE-2025-3262
Vulnerability from fkie_nvd - Published: 2025-07-07 10:15 - Updated: 2025-08-02 01:20
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the huggingface/transformers repository, specifically in version 4.49.0. The vulnerability is due to inefficient regular expression complexity in the `SETTING_RE` variable within the `transformers/commands/chat.py` file. The regex contains repetition groups and non-optimized quantifiers, leading to exponential backtracking when processing 'almost matching' payloads. This can degrade application performance and potentially result in a denial-of-service (DoS) when handling specially crafted input strings. The issue is fixed in version 4.51.0.
References
| URL | Tags | ||
|---|---|---|---|
| security@huntr.dev | https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76 | Patch | |
| security@huntr.dev | https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184 | Exploit, Third Party Advisory | |
| 134c704f-9b21-4f2e-91b3-4a467353bcc0 | https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184 | Exploit, Third Party Advisory |
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "EE8F40B9-E52D-4DD7-9E09-EADE25B7788C",
"versionEndExcluding": "4.51.0",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the huggingface/transformers repository, specifically in version 4.49.0. The vulnerability is due to inefficient regular expression complexity in the `SETTING_RE` variable within the `transformers/commands/chat.py` file. The regex contains repetition groups and non-optimized quantifiers, leading to exponential backtracking when processing \u0027almost matching\u0027 payloads. This can degrade application performance and potentially result in a denial-of-service (DoS) when handling specially crafted input strings. The issue is fixed in version 4.51.0."
},
{
"lang": "es",
"value": "Se descubri\u00f3 una vulnerabilidad de denegaci\u00f3n de servicio por expresi\u00f3n regular (ReDoS) en el repositorio huggingface/transformers, concretamente en la versi\u00f3n 4.49.0. La vulnerabilidad se debe a la complejidad ineficiente de la expresi\u00f3n regular en la variable `SETTING_RE` del archivo `transformers/commands/chat.py`. La expresi\u00f3n regular contiene grupos de repetici\u00f3n y cuantificadores no optimizados, lo que provoca un retroceso exponencial al procesar payloads casi coincidentes. Esto puede reducir el rendimiento de la aplicaci\u00f3n y potencialmente provocar una denegaci\u00f3n de servicio (DoS) al procesar cadenas de entrada especialmente manipuladas. El problema se ha corregido en la versi\u00f3n 4.51.0."
}
],
"id": "CVE-2025-3262",
"lastModified": "2025-08-02T01:20:02.673",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
],
"cvssMetricV31": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"version": "3.1"
},
"exploitabilityScore": 3.9,
"impactScore": 3.6,
"source": "nvd@nist.gov",
"type": "Primary"
}
]
},
"published": "2025-07-07T10:15:27.200",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184"
},
{
"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Secondary"
}
]
}
FKIE_CVE-2025-3263
Vulnerability from fkie_nvd - Published: 2025-07-07 10:15 - Updated: 2025-08-07 01:03
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_configuration_file()` function within the `transformers.configuration_utils` module. The affected version is 4.49.0, and the issue is resolved in version 4.51.0. The vulnerability arises from the use of a regular expression pattern `config\.(.*)\.json` that can be exploited to cause excessive CPU consumption through crafted input strings, leading to catastrophic backtracking. This can result in model serving disruption, resource exhaustion, and increased latency in applications using the library.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "C286893D-37B2-495A-A91F-F7CB2CFE872C",
"versionEndExcluding": "4.51.0",
"versionStartIncluding": "4.49.0",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_configuration_file()` function within the `transformers.configuration_utils` module. The affected version is 4.49.0, and the issue is resolved in version 4.51.0. The vulnerability arises from the use of a regular expression pattern `config\\.(.*)\\.json` that can be exploited to cause excessive CPU consumption through crafted input strings, leading to catastrophic backtracking. This can result in model serving disruption, resource exhaustion, and increased latency in applications using the library."
},
{
"lang": "es",
"value": "Se descubri\u00f3 una vulnerabilidad de denegaci\u00f3n de servicio por expresi\u00f3n regular (ReDoS) en la librer\u00eda Hugging Face Transformers, espec\u00edficamente en la funci\u00f3n `get_configuration_file()` del m\u00f3dulo `transformers.configuration_utils`. La versi\u00f3n afectada es la 4.49.0 y el problema se solucion\u00f3 en la versi\u00f3n 4.51.0. La vulnerabilidad surge del uso de un patr\u00f3n de expresi\u00f3n regular `config\\.(.*)\\.json`, que puede explotarse para causar un consumo excesivo de CPU mediante cadenas de entrada manipuladas, lo que provoca un retroceso catastr\u00f3fico. Esto puede provocar interrupciones en el servicio del modelo, agotamiento de recursos y mayor latencia en las aplicaciones que utilizan la librer\u00eda."
}
],
"id": "CVE-2025-3263",
"lastModified": "2025-08-07T01:03:17.990",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
]
},
"published": "2025-07-07T10:15:27.350",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/c7a69150-54f8-4e81-8094-791e7a2a0f29"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
FKIE_CVE-2025-3264
Vulnerability from fkie_nvd - Published: 2025-07-07 10:15 - Updated: 2025-08-07 01:02
Severity ?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_imports()` function within `dynamic_module_utils.py`. This vulnerability affects versions 4.49.0 and is fixed in version 4.51.0. The issue arises from a regular expression pattern `\s*try\s*:.*?except.*?:` used to filter out try/except blocks from Python code, which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to remote code loading disruption, resource exhaustion in model serving, supply chain attack vectors, and development pipeline disruption.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "C286893D-37B2-495A-A91F-F7CB2CFE872C",
"versionEndExcluding": "4.51.0",
"versionStartIncluding": "4.49.0",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_imports()` function within `dynamic_module_utils.py`. This vulnerability affects versions 4.49.0 and is fixed in version 4.51.0. The issue arises from a regular expression pattern `\\s*try\\s*:.*?except.*?:` used to filter out try/except blocks from Python code, which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to remote code loading disruption, resource exhaustion in model serving, supply chain attack vectors, and development pipeline disruption."
},
{
"lang": "es",
"value": "Se descubri\u00f3 una vulnerabilidad de denegaci\u00f3n de servicio por expresi\u00f3n regular (ReDoS) en la librer\u00eda Hugging Face Transformers, espec\u00edficamente en la funci\u00f3n `get_imports()` de `dynamic_module_utils.py`. Esta vulnerabilidad afecta a la versi\u00f3n 4.49.0 y se corrigi\u00f3 en la versi\u00f3n 4.51.0. El problema surge de un patr\u00f3n de expresi\u00f3n regular `\\s*try\\s*:.*?except.*?:` utilizado para filtrar bloques try/except del c\u00f3digo Python, que puede explotarse para causar un consumo excesivo de CPU mediante cadenas de entrada manipuladas debido a un retroceso catastr\u00f3fico. Esta vulnerabilidad puede provocar interrupciones en la carga remota de c\u00f3digo, agotamiento de recursos en la entrega de modelos, vectores de ataque a la cadena de suministro e interrupciones en el flujo de trabajo de desarrollo."
}
],
"id": "CVE-2025-3264",
"lastModified": "2025-08-07T01:02:30.700",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
]
},
"published": "2025-07-07T10:15:27.500",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/3c6f7822-9992-476d-8cf0-b0b1623427df"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
FKIE_CVE-2025-2099
Vulnerability from fkie_nvd - Published: 2025-05-19 12:15 - Updated: 2025-05-21 17:43
Severity ?
Summary
A vulnerability in the `preprocess_string()` function of the `transformers.testing_utils` module in huggingface/transformers version v4.48.3 allows for a Regular Expression Denial of Service (ReDoS) attack. The regular expression used to process code blocks in docstrings contains nested quantifiers, leading to exponential backtracking when processing input with a large number of newline characters. An attacker can exploit this by providing a specially crafted payload, causing high CPU usage and potential application downtime, effectively resulting in a Denial of Service (DoS) scenario.
References
Impacted products
| Vendor | Product | Version | |
|---|---|---|---|
| huggingface | transformers | * |
{
"configurations": [
{
"nodes": [
{
"cpeMatch": [
{
"criteria": "cpe:2.3:a:huggingface:transformers:*:*:*:*:*:*:*:*",
"matchCriteriaId": "8C6BB08B-59CD-4524-AC16-55130113AE69",
"versionEndIncluding": "4.48.3",
"vulnerable": true
}
],
"negate": false,
"operator": "OR"
}
]
}
],
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the `preprocess_string()` function of the `transformers.testing_utils` module in huggingface/transformers version v4.48.3 allows for a Regular Expression Denial of Service (ReDoS) attack. The regular expression used to process code blocks in docstrings contains nested quantifiers, leading to exponential backtracking when processing input with a large number of newline characters. An attacker can exploit this by providing a specially crafted payload, causing high CPU usage and potential application downtime, effectively resulting in a Denial of Service (DoS) scenario."
},
{
"lang": "es",
"value": "Una vulnerabilidad en la funci\u00f3n `preprocess_string()` del m\u00f3dulo `transformers.testing_utils` en huggingface/transformers versi\u00f3n v4.48.3 permite un ataque de denegaci\u00f3n de servicio (ReDoS) mediante expresiones regulares. La expresi\u00f3n regular utilizada para procesar bloques de c\u00f3digo en cadenas de documentaci\u00f3n contiene cuantificadores anidados, lo que provoca un retroceso exponencial al procesar entradas con un gran n\u00famero de caracteres de nueva l\u00ednea. Un atacante puede explotar esto proporcionando un payload especialmente manipulado, lo que provoca un alto consumo de CPU y un posible tiempo de inactividad de la aplicaci\u00f3n, lo que resulta en un escenario de denegaci\u00f3n de servicio (DoS)."
}
],
"id": "CVE-2025-2099",
"lastModified": "2025-05-21T17:43:15.080",
"metrics": {
"cvssMetricV30": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
},
"exploitabilityScore": 3.9,
"impactScore": 1.4,
"source": "security@huntr.dev",
"type": "Secondary"
}
],
"cvssMetricV31": [
{
"cvssData": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "NONE",
"integrityImpact": "NONE",
"privilegesRequired": "NONE",
"scope": "UNCHANGED",
"userInteraction": "NONE",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"version": "3.1"
},
"exploitabilityScore": 3.9,
"impactScore": 3.6,
"source": "nvd@nist.gov",
"type": "Primary"
}
]
},
"published": "2025-05-19T12:15:19.640",
"references": [
{
"source": "security@huntr.dev",
"tags": [
"Patch"
],
"url": "https://github.com/huggingface/transformers/commit/8cb522b4190bd556ce51be04942720650b1a3e57"
},
{
"source": "security@huntr.dev",
"tags": [
"Exploit",
"Third Party Advisory"
],
"url": "https://huntr.com/bounties/97b780f3-ffca-424f-ad5d-0e1c57a5bde4"
}
],
"sourceIdentifier": "security@huntr.dev",
"vulnStatus": "Analyzed",
"weaknesses": [
{
"description": [
{
"lang": "en",
"value": "CWE-1333"
}
],
"source": "security@huntr.dev",
"type": "Primary"
}
]
}
CVE-2025-6921 (GCVE-0-2025-6921)
Vulnerability from cvelistv5 – Published: 2025-09-23 13:56 – Updated: 2025-09-23 14:56
VLAI?
Summary
The huggingface/transformers library, versions prior to 4.53.0, is vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer. The vulnerability arises from the _do_use_weight_decay method, which processes user-controlled regular expressions in the include_in_weight_decay and exclude_from_weight_decay lists. Malicious regular expressions can cause catastrophic backtracking during the re.search call, leading to 100% CPU utilization and a denial of service. This issue can be exploited by attackers who can control the patterns in these lists, potentially causing the machine learning task to hang and rendering services unresponsive.
Severity ?
5.3 (Medium)
CWE
- CWE-400 - Uncontrolled Resource Consumption
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-6921",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-09-23T14:56:14.042617Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-09-23T14:56:29.038Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "The huggingface/transformers library, versions prior to 4.53.0, is vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer. The vulnerability arises from the _do_use_weight_decay method, which processes user-controlled regular expressions in the include_in_weight_decay and exclude_from_weight_decay lists. Malicious regular expressions can cause catastrophic backtracking during the re.search call, leading to 100% CPU utilization and a denial of service. This issue can be exploited by attackers who can control the patterns in these lists, potentially causing the machine learning task to hang and rendering services unresponsive."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-400",
"description": "CWE-400 Uncontrolled Resource Consumption",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-09-23T13:56:15.636Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/287d15a7-6e7c-45d2-8c05-11e305776f1f"
},
{
"url": "https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be"
}
],
"source": {
"advisory": "287d15a7-6e7c-45d2-8c05-11e305776f1f",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-6921",
"datePublished": "2025-09-23T13:56:15.636Z",
"dateReserved": "2025-06-30T09:44:12.092Z",
"dateUpdated": "2025-09-23T14:56:29.038Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-6051 (GCVE-0-2025-6051)
Vulnerability from cvelistv5 – Published: 2025-09-14 17:03 – Updated: 2025-09-15 15:59
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method's handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-6051",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-09-15T15:59:46.280744Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-09-15T15:59:54.371Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method\u0027s handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-09-14T17:03:02.953Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/af929523-7b59-418a-bf55-301830b2ac9d"
},
{
"url": "https://github.com/huggingface/transformers/commit/ba8eaba9865618253f997784aa565b96206426f0"
}
],
"source": {
"advisory": "af929523-7b59-418a-bf55-301830b2ac9d",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-6051",
"datePublished": "2025-09-14T17:03:02.953Z",
"dateReserved": "2025-06-13T10:39:33.128Z",
"dateUpdated": "2025-09-15T15:59:54.371Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-6638 (GCVE-0-2025-6638)
Vulnerability from cvelistv5 – Published: 2025-09-12 10:46 – Updated: 2025-09-12 11:52
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer's `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-6638",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-09-12T11:52:42.923436Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-09-12T11:52:53.854Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer\u0027s `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-09-12T10:46:07.934Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/6a6c933f-9ce8-4ded-8b3b-2c1444c61f36"
},
{
"url": "https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be"
}
],
"source": {
"advisory": "6a6c933f-9ce8-4ded-8b3b-2c1444c61f36",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-6638",
"datePublished": "2025-09-12T10:46:07.934Z",
"dateReserved": "2025-06-25T14:07:29.841Z",
"dateUpdated": "2025-09-12T11:52:53.854Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-5197 (GCVE-0-2025-5197)
Vulnerability from cvelistv5 – Published: 2025-08-06 11:53 – Updated: 2025-08-06 13:03
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-5197",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-08-06T13:02:53.771161Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-08-06T13:03:10.518Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-08-06T11:53:37.232Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf"
},
{
"url": "https://github.com/huggingface/transformers/commit/944b56000be5e9b61af8301aa340838770ad8a0b"
}
],
"source": {
"advisory": "3f8b3fd0-166b-46e7-b60f-60dd9d2678bf",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-5197",
"datePublished": "2025-08-06T11:53:37.232Z",
"dateReserved": "2025-05-26T09:26:53.172Z",
"dateUpdated": "2025-08-06T13:03:10.518Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3933 (GCVE-0-2025-3933)
Vulnerability from cvelistv5 – Published: 2025-07-11 09:22 – Updated: 2025-07-11 15:07
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the DonutProcessor class's `token2json()` method. This vulnerability affects versions 4.50.3 and earlier, and is fixed in version 4.52.1. The issue arises from the regex pattern `<s_(.*?)>` which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting document processing tasks using the Donut model.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.52.1
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3933",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-11T14:34:20.168976Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-11T15:07:34.839Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.52.1",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the DonutProcessor class\u0027s `token2json()` method. This vulnerability affects versions 4.50.3 and earlier, and is fixed in version 4.52.1. The issue arises from the regex pattern `\u003cs_(.*?)\u003e` which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting document processing tasks using the Donut model."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-11T09:22:27.103Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/25282953-5827-4384-bb6f-5790d275721b"
},
{
"url": "https://github.com/huggingface/transformers/commit/ebbe9b12dd75b69f92100d684c47f923ee262a93"
}
],
"source": {
"advisory": "25282953-5827-4384-bb6f-5790d275721b",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3933",
"datePublished": "2025-07-11T09:22:27.103Z",
"dateReserved": "2025-04-25T13:37:56.821Z",
"dateUpdated": "2025-07-11T15:07:34.839Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3777 (GCVE-0-2025-3777)
Vulnerability from cvelistv5 – Published: 2025-07-07 09:55 – Updated: 2025-07-07 12:26
VLAI?
Summary
Hugging Face Transformers versions up to 4.49.0 are affected by an improper input validation vulnerability in the `image_utils.py` file. The vulnerability arises from insecure URL validation using the `startswith()` method, which can be bypassed through URL username injection. This allows attackers to craft URLs that appear to be from YouTube but resolve to malicious domains, potentially leading to phishing attacks, malware distribution, or data exfiltration. The issue is fixed in version 4.52.1.
Severity ?
CWE
- CWE-20 - Improper Input Validation
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.52.1
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3777",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T12:24:47.677205Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T12:26:15.099Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.52.1",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers versions up to 4.49.0 are affected by an improper input validation vulnerability in the `image_utils.py` file. The vulnerability arises from insecure URL validation using the `startswith()` method, which can be bypassed through URL username injection. This allows attackers to craft URLs that appear to be from YouTube but resolve to malicious domains, potentially leading to phishing attacks, malware distribution, or data exfiltration. The issue is fixed in version 4.52.1."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 3.5,
"baseSeverity": "LOW",
"confidentialityImpact": "LOW",
"integrityImpact": "NONE",
"privilegesRequired": "LOW",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:L/UI:R/S:U/C:L/I:N/A:N",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-20",
"description": "CWE-20 Improper Input Validation",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:55:38.190Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/ccba0730-9248-4853-b7ff-5c20e6364f09"
},
{
"url": "https://github.com/huggingface/transformers/commit/4dda5f71b35fb70cf602187eef84bb17a50b9082"
}
],
"source": {
"advisory": "ccba0730-9248-4853-b7ff-5c20e6364f09",
"discovery": "EXTERNAL"
},
"title": "Improper Input Validation in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3777",
"datePublished": "2025-07-07T09:55:38.190Z",
"dateReserved": "2025-04-17T17:35:27.010Z",
"dateUpdated": "2025-07-07T12:26:15.099Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3264 (GCVE-0-2025-3264)
Vulnerability from cvelistv5 – Published: 2025-07-07 09:55 – Updated: 2025-07-07 14:40
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_imports()` function within `dynamic_module_utils.py`. This vulnerability affects versions 4.49.0 and is fixed in version 4.51.0. The issue arises from a regular expression pattern `\s*try\s*:.*?except.*?:` used to filter out try/except blocks from Python code, which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to remote code loading disruption, resource exhaustion in model serving, supply chain attack vectors, and development pipeline disruption.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.51.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3264",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T14:37:34.878471Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T14:40:12.531Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.51.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_imports()` function within `dynamic_module_utils.py`. This vulnerability affects versions 4.49.0 and is fixed in version 4.51.0. The issue arises from a regular expression pattern `\\s*try\\s*:.*?except.*?:` used to filter out try/except blocks from Python code, which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to remote code loading disruption, resource exhaustion in model serving, supply chain attack vectors, and development pipeline disruption."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:55:11.030Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/3c6f7822-9992-476d-8cf0-b0b1623427df"
},
{
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
}
],
"source": {
"advisory": "3c6f7822-9992-476d-8cf0-b0b1623427df",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3264",
"datePublished": "2025-07-07T09:55:11.030Z",
"dateReserved": "2025-04-04T12:41:38.443Z",
"dateUpdated": "2025-07-07T14:40:12.531Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3263 (GCVE-0-2025-3263)
Vulnerability from cvelistv5 – Published: 2025-07-07 09:54 – Updated: 2025-07-07 14:51
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_configuration_file()` function within the `transformers.configuration_utils` module. The affected version is 4.49.0, and the issue is resolved in version 4.51.0. The vulnerability arises from the use of a regular expression pattern `config\.(.*)\.json` that can be exploited to cause excessive CPU consumption through crafted input strings, leading to catastrophic backtracking. This can result in model serving disruption, resource exhaustion, and increased latency in applications using the library.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.51.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3263",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T14:49:04.994384Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T14:51:54.953Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.51.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_configuration_file()` function within the `transformers.configuration_utils` module. The affected version is 4.49.0, and the issue is resolved in version 4.51.0. The vulnerability arises from the use of a regular expression pattern `config\\.(.*)\\.json` that can be exploited to cause excessive CPU consumption through crafted input strings, leading to catastrophic backtracking. This can result in model serving disruption, resource exhaustion, and increased latency in applications using the library."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:54:59.670Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/c7a69150-54f8-4e81-8094-791e7a2a0f29"
},
{
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
}
],
"source": {
"advisory": "c7a69150-54f8-4e81-8094-791e7a2a0f29",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3263",
"datePublished": "2025-07-07T09:54:59.670Z",
"dateReserved": "2025-04-04T12:40:57.226Z",
"dateUpdated": "2025-07-07T14:51:54.953Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3262 (GCVE-0-2025-3262)
Vulnerability from cvelistv5 – Published: 2025-07-07 09:54 – Updated: 2025-07-07 15:19
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the huggingface/transformers repository, specifically in version 4.49.0. The vulnerability is due to inefficient regular expression complexity in the `SETTING_RE` variable within the `transformers/commands/chat.py` file. The regex contains repetition groups and non-optimized quantifiers, leading to exponential backtracking when processing 'almost matching' payloads. This can degrade application performance and potentially result in a denial-of-service (DoS) when handling specially crafted input strings. The issue is fixed in version 4.51.0.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.51.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3262",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T15:19:35.031775Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T15:19:42.413Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.51.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the huggingface/transformers repository, specifically in version 4.49.0. The vulnerability is due to inefficient regular expression complexity in the `SETTING_RE` variable within the `transformers/commands/chat.py` file. The regex contains repetition groups and non-optimized quantifiers, leading to exponential backtracking when processing \u0027almost matching\u0027 payloads. This can degrade application performance and potentially result in a denial-of-service (DoS) when handling specially crafted input strings. The issue is fixed in version 4.51.0."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:54:39.079Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184"
},
{
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
}
],
"source": {
"advisory": "ecf5ccc4-39e7-4fb3-b547-14a41d31a184",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3262",
"datePublished": "2025-07-07T09:54:39.079Z",
"dateReserved": "2025-04-04T12:40:28.677Z",
"dateUpdated": "2025-07-07T15:19:42.413Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-2099 (GCVE-0-2025-2099)
Vulnerability from cvelistv5 – Published: 2025-05-19 11:22 – Updated: 2025-05-19 13:45
VLAI?
Summary
A vulnerability in the `preprocess_string()` function of the `transformers.testing_utils` module in huggingface/transformers version v4.48.3 allows for a Regular Expression Denial of Service (ReDoS) attack. The regular expression used to process code blocks in docstrings contains nested quantifiers, leading to exponential backtracking when processing input with a large number of newline characters. An attacker can exploit this by providing a specially crafted payload, causing high CPU usage and potential application downtime, effectively resulting in a Denial of Service (DoS) scenario.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.50.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-2099",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-05-19T13:38:03.242262Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-05-19T13:45:13.385Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.50.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the `preprocess_string()` function of the `transformers.testing_utils` module in huggingface/transformers version v4.48.3 allows for a Regular Expression Denial of Service (ReDoS) attack. The regular expression used to process code blocks in docstrings contains nested quantifiers, leading to exponential backtracking when processing input with a large number of newline characters. An attacker can exploit this by providing a specially crafted payload, causing high CPU usage and potential application downtime, effectively resulting in a Denial of Service (DoS) scenario."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-05-19T11:22:36.908Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/97b780f3-ffca-424f-ad5d-0e1c57a5bde4"
},
{
"url": "https://github.com/huggingface/transformers/commit/8cb522b4190bd556ce51be04942720650b1a3e57"
}
],
"source": {
"advisory": "97b780f3-ffca-424f-ad5d-0e1c57a5bde4",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-2099",
"datePublished": "2025-05-19T11:22:36.908Z",
"dateReserved": "2025-03-07T17:39:16.856Z",
"dateUpdated": "2025-05-19T13:45:13.385Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-6921 (GCVE-0-2025-6921)
Vulnerability from nvd – Published: 2025-09-23 13:56 – Updated: 2025-09-23 14:56
VLAI?
Summary
The huggingface/transformers library, versions prior to 4.53.0, is vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer. The vulnerability arises from the _do_use_weight_decay method, which processes user-controlled regular expressions in the include_in_weight_decay and exclude_from_weight_decay lists. Malicious regular expressions can cause catastrophic backtracking during the re.search call, leading to 100% CPU utilization and a denial of service. This issue can be exploited by attackers who can control the patterns in these lists, potentially causing the machine learning task to hang and rendering services unresponsive.
Severity ?
5.3 (Medium)
CWE
- CWE-400 - Uncontrolled Resource Consumption
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-6921",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-09-23T14:56:14.042617Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-09-23T14:56:29.038Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "The huggingface/transformers library, versions prior to 4.53.0, is vulnerable to Regular Expression Denial of Service (ReDoS) in the AdamWeightDecay optimizer. The vulnerability arises from the _do_use_weight_decay method, which processes user-controlled regular expressions in the include_in_weight_decay and exclude_from_weight_decay lists. Malicious regular expressions can cause catastrophic backtracking during the re.search call, leading to 100% CPU utilization and a denial of service. This issue can be exploited by attackers who can control the patterns in these lists, potentially causing the machine learning task to hang and rendering services unresponsive."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-400",
"description": "CWE-400 Uncontrolled Resource Consumption",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-09-23T13:56:15.636Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/287d15a7-6e7c-45d2-8c05-11e305776f1f"
},
{
"url": "https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be"
}
],
"source": {
"advisory": "287d15a7-6e7c-45d2-8c05-11e305776f1f",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-6921",
"datePublished": "2025-09-23T13:56:15.636Z",
"dateReserved": "2025-06-30T09:44:12.092Z",
"dateUpdated": "2025-09-23T14:56:29.038Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-6051 (GCVE-0-2025-6051)
Vulnerability from nvd – Published: 2025-09-14 17:03 – Updated: 2025-09-15 15:59
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method's handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-6051",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-09-15T15:59:46.280744Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-09-15T15:59:54.371Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the `normalize_numbers()` method of the `EnglishNormalizer` class. This vulnerability affects versions up to 4.52.4 and is fixed in version 4.53.0. The issue arises from the method\u0027s handling of numeric strings, which can be exploited using crafted input strings containing long sequences of digits, leading to excessive CPU consumption. This vulnerability impacts text-to-speech and number normalization tasks, potentially causing service disruption, resource exhaustion, and API vulnerabilities."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-09-14T17:03:02.953Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/af929523-7b59-418a-bf55-301830b2ac9d"
},
{
"url": "https://github.com/huggingface/transformers/commit/ba8eaba9865618253f997784aa565b96206426f0"
}
],
"source": {
"advisory": "af929523-7b59-418a-bf55-301830b2ac9d",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-6051",
"datePublished": "2025-09-14T17:03:02.953Z",
"dateReserved": "2025-06-13T10:39:33.128Z",
"dateUpdated": "2025-09-15T15:59:54.371Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-6638 (GCVE-0-2025-6638)
Vulnerability from nvd – Published: 2025-09-12 10:46 – Updated: 2025-09-12 11:52
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer's `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-6638",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-09-12T11:52:42.923436Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-09-12T11:52:53.854Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically affecting the MarianTokenizer\u0027s `remove_language_code()` method. This vulnerability is present in version 4.52.4 and has been fixed in version 4.53.0. The issue arises from inefficient regex processing, which can be exploited by crafted input strings containing malformed language code patterns, leading to excessive CPU consumption and potential denial of service."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-09-12T10:46:07.934Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/6a6c933f-9ce8-4ded-8b3b-2c1444c61f36"
},
{
"url": "https://github.com/huggingface/transformers/commit/47c34fba5c303576560cb29767efb452ff12b8be"
}
],
"source": {
"advisory": "6a6c933f-9ce8-4ded-8b3b-2c1444c61f36",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-6638",
"datePublished": "2025-09-12T10:46:07.934Z",
"dateReserved": "2025-06-25T14:07:29.841Z",
"dateUpdated": "2025-09-12T11:52:53.854Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-5197 (GCVE-0-2025-5197)
Vulnerability from nvd – Published: 2025-08-06 11:53 – Updated: 2025-08-06 13:03
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.53.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-5197",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-08-06T13:02:53.771161Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-08-06T13:03:10.518Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.53.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability exists in the Hugging Face Transformers library, specifically in the `convert_tf_weight_name_to_pt_weight_name()` function. This function, responsible for converting TensorFlow weight names to PyTorch format, uses a regex pattern `/[^/]*___([^/]*)/` that can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. The vulnerability affects versions up to 4.51.3 and is fixed in version 4.53.0. This issue can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting model conversion processes between TensorFlow and PyTorch formats."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-08-06T11:53:37.232Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/3f8b3fd0-166b-46e7-b60f-60dd9d2678bf"
},
{
"url": "https://github.com/huggingface/transformers/commit/944b56000be5e9b61af8301aa340838770ad8a0b"
}
],
"source": {
"advisory": "3f8b3fd0-166b-46e7-b60f-60dd9d2678bf",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-5197",
"datePublished": "2025-08-06T11:53:37.232Z",
"dateReserved": "2025-05-26T09:26:53.172Z",
"dateUpdated": "2025-08-06T13:03:10.518Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3933 (GCVE-0-2025-3933)
Vulnerability from nvd – Published: 2025-07-11 09:22 – Updated: 2025-07-11 15:07
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the DonutProcessor class's `token2json()` method. This vulnerability affects versions 4.50.3 and earlier, and is fixed in version 4.52.1. The issue arises from the regex pattern `<s_(.*?)>` which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting document processing tasks using the Donut model.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.52.1
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3933",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-11T14:34:20.168976Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-11T15:07:34.839Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.52.1",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically within the DonutProcessor class\u0027s `token2json()` method. This vulnerability affects versions 4.50.3 and earlier, and is fixed in version 4.52.1. The issue arises from the regex pattern `\u003cs_(.*?)\u003e` which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to service disruption, resource exhaustion, and potential API service vulnerabilities, impacting document processing tasks using the Donut model."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-11T09:22:27.103Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/25282953-5827-4384-bb6f-5790d275721b"
},
{
"url": "https://github.com/huggingface/transformers/commit/ebbe9b12dd75b69f92100d684c47f923ee262a93"
}
],
"source": {
"advisory": "25282953-5827-4384-bb6f-5790d275721b",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3933",
"datePublished": "2025-07-11T09:22:27.103Z",
"dateReserved": "2025-04-25T13:37:56.821Z",
"dateUpdated": "2025-07-11T15:07:34.839Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3777 (GCVE-0-2025-3777)
Vulnerability from nvd – Published: 2025-07-07 09:55 – Updated: 2025-07-07 12:26
VLAI?
Summary
Hugging Face Transformers versions up to 4.49.0 are affected by an improper input validation vulnerability in the `image_utils.py` file. The vulnerability arises from insecure URL validation using the `startswith()` method, which can be bypassed through URL username injection. This allows attackers to craft URLs that appear to be from YouTube but resolve to malicious domains, potentially leading to phishing attacks, malware distribution, or data exfiltration. The issue is fixed in version 4.52.1.
Severity ?
CWE
- CWE-20 - Improper Input Validation
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.52.1
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3777",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "no"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T12:24:47.677205Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T12:26:15.099Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.52.1",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "Hugging Face Transformers versions up to 4.49.0 are affected by an improper input validation vulnerability in the `image_utils.py` file. The vulnerability arises from insecure URL validation using the `startswith()` method, which can be bypassed through URL username injection. This allows attackers to craft URLs that appear to be from YouTube but resolve to malicious domains, potentially leading to phishing attacks, malware distribution, or data exfiltration. The issue is fixed in version 4.52.1."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 3.5,
"baseSeverity": "LOW",
"confidentialityImpact": "LOW",
"integrityImpact": "NONE",
"privilegesRequired": "LOW",
"scope": "UNCHANGED",
"userInteraction": "REQUIRED",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:L/UI:R/S:U/C:L/I:N/A:N",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-20",
"description": "CWE-20 Improper Input Validation",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:55:38.190Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/ccba0730-9248-4853-b7ff-5c20e6364f09"
},
{
"url": "https://github.com/huggingface/transformers/commit/4dda5f71b35fb70cf602187eef84bb17a50b9082"
}
],
"source": {
"advisory": "ccba0730-9248-4853-b7ff-5c20e6364f09",
"discovery": "EXTERNAL"
},
"title": "Improper Input Validation in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3777",
"datePublished": "2025-07-07T09:55:38.190Z",
"dateReserved": "2025-04-17T17:35:27.010Z",
"dateUpdated": "2025-07-07T12:26:15.099Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3264 (GCVE-0-2025-3264)
Vulnerability from nvd – Published: 2025-07-07 09:55 – Updated: 2025-07-07 14:40
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_imports()` function within `dynamic_module_utils.py`. This vulnerability affects versions 4.49.0 and is fixed in version 4.51.0. The issue arises from a regular expression pattern `\s*try\s*:.*?except.*?:` used to filter out try/except blocks from Python code, which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to remote code loading disruption, resource exhaustion in model serving, supply chain attack vectors, and development pipeline disruption.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.51.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3264",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T14:37:34.878471Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T14:40:12.531Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.51.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_imports()` function within `dynamic_module_utils.py`. This vulnerability affects versions 4.49.0 and is fixed in version 4.51.0. The issue arises from a regular expression pattern `\\s*try\\s*:.*?except.*?:` used to filter out try/except blocks from Python code, which can be exploited to cause excessive CPU consumption through crafted input strings due to catastrophic backtracking. This vulnerability can lead to remote code loading disruption, resource exhaustion in model serving, supply chain attack vectors, and development pipeline disruption."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:55:11.030Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/3c6f7822-9992-476d-8cf0-b0b1623427df"
},
{
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
}
],
"source": {
"advisory": "3c6f7822-9992-476d-8cf0-b0b1623427df",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3264",
"datePublished": "2025-07-07T09:55:11.030Z",
"dateReserved": "2025-04-04T12:41:38.443Z",
"dateUpdated": "2025-07-07T14:40:12.531Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3263 (GCVE-0-2025-3263)
Vulnerability from nvd – Published: 2025-07-07 09:54 – Updated: 2025-07-07 14:51
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_configuration_file()` function within the `transformers.configuration_utils` module. The affected version is 4.49.0, and the issue is resolved in version 4.51.0. The vulnerability arises from the use of a regular expression pattern `config\.(.*)\.json` that can be exploited to cause excessive CPU consumption through crafted input strings, leading to catastrophic backtracking. This can result in model serving disruption, resource exhaustion, and increased latency in applications using the library.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.51.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3263",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T14:49:04.994384Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T14:51:54.953Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.51.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the Hugging Face Transformers library, specifically in the `get_configuration_file()` function within the `transformers.configuration_utils` module. The affected version is 4.49.0, and the issue is resolved in version 4.51.0. The vulnerability arises from the use of a regular expression pattern `config\\.(.*)\\.json` that can be exploited to cause excessive CPU consumption through crafted input strings, leading to catastrophic backtracking. This can result in model serving disruption, resource exhaustion, and increased latency in applications using the library."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:54:59.670Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/c7a69150-54f8-4e81-8094-791e7a2a0f29"
},
{
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
}
],
"source": {
"advisory": "c7a69150-54f8-4e81-8094-791e7a2a0f29",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3263",
"datePublished": "2025-07-07T09:54:59.670Z",
"dateReserved": "2025-04-04T12:40:57.226Z",
"dateUpdated": "2025-07-07T14:51:54.953Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-3262 (GCVE-0-2025-3262)
Vulnerability from nvd – Published: 2025-07-07 09:54 – Updated: 2025-07-07 15:19
VLAI?
Summary
A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the huggingface/transformers repository, specifically in version 4.49.0. The vulnerability is due to inefficient regular expression complexity in the `SETTING_RE` variable within the `transformers/commands/chat.py` file. The regex contains repetition groups and non-optimized quantifiers, leading to exponential backtracking when processing 'almost matching' payloads. This can degrade application performance and potentially result in a denial-of-service (DoS) when handling specially crafted input strings. The issue is fixed in version 4.51.0.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.51.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-3262",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-07-07T15:19:35.031775Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T15:19:42.413Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"references": [
{
"tags": [
"exploit"
],
"url": "https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184"
}
],
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.51.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A Regular Expression Denial of Service (ReDoS) vulnerability was discovered in the huggingface/transformers repository, specifically in version 4.49.0. The vulnerability is due to inefficient regular expression complexity in the `SETTING_RE` variable within the `transformers/commands/chat.py` file. The regex contains repetition groups and non-optimized quantifiers, leading to exponential backtracking when processing \u0027almost matching\u0027 payloads. This can degrade application performance and potentially result in a denial-of-service (DoS) when handling specially crafted input strings. The issue is fixed in version 4.51.0."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-07-07T09:54:39.079Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/ecf5ccc4-39e7-4fb3-b547-14a41d31a184"
},
{
"url": "https://github.com/huggingface/transformers/commit/0720e206c6ba28887e4d60ef60a6a089f6c1cc76"
}
],
"source": {
"advisory": "ecf5ccc4-39e7-4fb3-b547-14a41d31a184",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-3262",
"datePublished": "2025-07-07T09:54:39.079Z",
"dateReserved": "2025-04-04T12:40:28.677Z",
"dateUpdated": "2025-07-07T15:19:42.413Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}
CVE-2025-2099 (GCVE-0-2025-2099)
Vulnerability from nvd – Published: 2025-05-19 11:22 – Updated: 2025-05-19 13:45
VLAI?
Summary
A vulnerability in the `preprocess_string()` function of the `transformers.testing_utils` module in huggingface/transformers version v4.48.3 allows for a Regular Expression Denial of Service (ReDoS) attack. The regular expression used to process code blocks in docstrings contains nested quantifiers, leading to exponential backtracking when processing input with a large number of newline characters. An attacker can exploit this by providing a specially crafted payload, causing high CPU usage and potential application downtime, effectively resulting in a Denial of Service (DoS) scenario.
Severity ?
5.3 (Medium)
CWE
- CWE-1333 - Inefficient Regular Expression Complexity
Assigner
References
Impacted products
| Vendor | Product | Version | ||
|---|---|---|---|---|
| huggingface | huggingface/transformers |
Affected:
unspecified , < 4.50.0
(custom)
|
{
"containers": {
"adp": [
{
"metrics": [
{
"other": {
"content": {
"id": "CVE-2025-2099",
"options": [
{
"Exploitation": "poc"
},
{
"Automatable": "yes"
},
{
"Technical Impact": "partial"
}
],
"role": "CISA Coordinator",
"timestamp": "2025-05-19T13:38:03.242262Z",
"version": "2.0.3"
},
"type": "ssvc"
}
}
],
"providerMetadata": {
"dateUpdated": "2025-05-19T13:45:13.385Z",
"orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"shortName": "CISA-ADP"
},
"title": "CISA ADP Vulnrichment"
}
],
"cna": {
"affected": [
{
"product": "huggingface/transformers",
"vendor": "huggingface",
"versions": [
{
"lessThan": "4.50.0",
"status": "affected",
"version": "unspecified",
"versionType": "custom"
}
]
}
],
"descriptions": [
{
"lang": "en",
"value": "A vulnerability in the `preprocess_string()` function of the `transformers.testing_utils` module in huggingface/transformers version v4.48.3 allows for a Regular Expression Denial of Service (ReDoS) attack. The regular expression used to process code blocks in docstrings contains nested quantifiers, leading to exponential backtracking when processing input with a large number of newline characters. An attacker can exploit this by providing a specially crafted payload, causing high CPU usage and potential application downtime, effectively resulting in a Denial of Service (DoS) scenario."
}
],
"metrics": [
{
"cvssV3_0": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "LOW",
"baseScore": 5.3,
"baseSeverity": "MEDIUM",
"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:L",
"version": "3.0"
}
}
],
"problemTypes": [
{
"descriptions": [
{
"cweId": "CWE-1333",
"description": "CWE-1333 Inefficient Regular Expression Complexity",
"lang": "en",
"type": "CWE"
}
]
}
],
"providerMetadata": {
"dateUpdated": "2025-05-19T11:22:36.908Z",
"orgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"shortName": "@huntr_ai"
},
"references": [
{
"url": "https://huntr.com/bounties/97b780f3-ffca-424f-ad5d-0e1c57a5bde4"
},
{
"url": "https://github.com/huggingface/transformers/commit/8cb522b4190bd556ce51be04942720650b1a3e57"
}
],
"source": {
"advisory": "97b780f3-ffca-424f-ad5d-0e1c57a5bde4",
"discovery": "EXTERNAL"
},
"title": "Regular Expression Denial of Service (ReDoS) in huggingface/transformers"
}
},
"cveMetadata": {
"assignerOrgId": "c09c270a-b464-47c1-9133-acb35b22c19a",
"assignerShortName": "@huntr_ai",
"cveId": "CVE-2025-2099",
"datePublished": "2025-05-19T11:22:36.908Z",
"dateReserved": "2025-03-07T17:39:16.856Z",
"dateUpdated": "2025-05-19T13:45:13.385Z",
"state": "PUBLISHED"
},
"dataType": "CVE_RECORD",
"dataVersion": "5.1"
}