Django REST framework

Package Instance

Lookup for vulnerable packages by Package URL.

Purlpkg:pypi/vllm@0.6.6.post1
Typepypi
Namespace
Namevllm
Version0.6.6.post1
Qualifiers
Subpath
Is_vulnerabletrue
Next_non_vulnerable_version0.20.0
Latest_non_vulnerable_version0.20.0
Affected_by_vulnerabilities
0
url VCID-737m-tpkz-qffm
vulnerability_id VCID-737m-tpkz-qffm
summary vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Maliciously constructed statements can lead to hash collisions, resulting in cache reuse, which can interfere with subsequent responses and cause unintended behavior. Prefix caching makes use of Python's built-in hash() function. As of Python 3.12, the behavior of hash(None) has changed to be a predictable constant value. This makes it more feasible that someone could try exploit hash collisions. The impact of a collision would be using cache that was generated using different content. Given knowledge of prompts in use and predictable hashing behavior, someone could intentionally populate the cache using a prompt known to collide with another prompt in use. This issue has been addressed in version 0.7.2 and all users are advised to upgrade. There are no known workarounds for this vulnerability.
references
0
reference_url https://github.com/python/cpython/commit/432117cd1f59c76d97da2eaff55a7d758301dbc7
reference_id
reference_type
scores
url https://github.com/python/cpython/commit/432117cd1f59c76d97da2eaff55a7d758301dbc7
1
reference_url https://github.com/vllm-project/vllm/pull/12621
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/pull/12621
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-rm76-4mrf-v9r8
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/security/advisories/GHSA-rm76-4mrf-v9r8
fixed_packages
0
url pkg:pypi/vllm@0.7.2
purl pkg:pypi/vllm@0.7.2
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-e8w2-9rwg-u7ba
1
vulnerability VCID-fxgs-s1vm-8bez
2
vulnerability VCID-k1qz-xe9c-2bg3
3
vulnerability VCID-nctw-rz8h-f3af
4
vulnerability VCID-svzy-7pke-2bdr
5
vulnerability VCID-u659-sd9h-tkf3
6
vulnerability VCID-ugds-eqgw-fbbz
7
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.7.2
aliases CVE-2025-25183, GHSA-rm76-4mrf-v9r8, PYSEC-2025-62
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-737m-tpkz-qffm
1
url VCID-e8w2-9rwg-u7ba
vulnerability_id VCID-e8w2-9rwg-u7ba
summary vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.9.0, when a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT (Time to First Token). These timing differences caused by matching chunks are significant enough to be recognized and exploited. This issue has been patched in version 0.9.0.
references
0
reference_url https://github.com/vllm-project/vllm/commit/77073c77bc2006eb80ea6d5128f076f5e6c6f54f
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/commit/77073c77bc2006eb80ea6d5128f076f5e6c6f54f
1
reference_url https://github.com/vllm-project/vllm/pull/17045
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/pull/17045
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-4qjh-9fv9-r85r
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/security/advisories/GHSA-4qjh-9fv9-r85r
fixed_packages
0
url pkg:pypi/vllm@0.9.0
purl pkg:pypi/vllm@0.9.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-nctw-rz8h-f3af
1
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.9.0
aliases CVE-2025-46570, GHSA-4qjh-9fv9-r85r, PYSEC-2025-53
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-e8w2-9rwg-u7ba
2
url VCID-fxgs-s1vm-8bez
vulnerability_id VCID-fxgs-s1vm-8bez
summary vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.6.5 and prior to 0.8.5, having vLLM integration with mooncake, are vulnerable to remote code execution due to using pickle based serialization over unsecured ZeroMQ sockets. The vulnerable sockets were set to listen on all network interfaces, increasing the likelihood that an attacker is able to reach the vulnerable ZeroMQ sockets to carry out an attack. vLLM instances that do not make use of the mooncake integration are not vulnerable. This issue has been patched in version 0.8.5.
references
0
reference_url https://github.com/vllm-project/vllm/blob/32b14baf8a1f7195ca09484de3008063569b43c5/vllm/distributed/kv_transfer/kv_pipe/mooncake_pipe.py#L179
reference_id
reference_type
scores
0
value 9.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/blob/32b14baf8a1f7195ca09484de3008063569b43c5/vllm/distributed/kv_transfer/kv_pipe/mooncake_pipe.py#L179
1
reference_url https://github.com/vllm-project/vllm/commit/a5450f11c95847cf51a17207af9a3ca5ab569b2c
reference_id
reference_type
scores
0
value 9.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/commit/a5450f11c95847cf51a17207af9a3ca5ab569b2c
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-hj4w-hm2g-p6w5
reference_id
reference_type
scores
0
value 9.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/security/advisories/GHSA-hj4w-hm2g-p6w5
3
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-x3m8-f7g5-qhm7
reference_id
reference_type
scores
0
value 9.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/security/advisories/GHSA-x3m8-f7g5-qhm7
fixed_packages
0
url pkg:pypi/vllm@0.8.5
purl pkg:pypi/vllm@0.8.5
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-5ec1-1h6d-tuaq
1
vulnerability VCID-e8w2-9rwg-u7ba
2
vulnerability VCID-nctw-rz8h-f3af
3
vulnerability VCID-qake-z4ec-wkdu
4
vulnerability VCID-svzy-7pke-2bdr
5
vulnerability VCID-ugds-eqgw-fbbz
6
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.8.5
aliases CVE-2025-32444, PYSEC-2025-42
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-fxgs-s1vm-8bez
3
url VCID-k1qz-xe9c-2bg3
vulnerability_id VCID-k1qz-xe9c-2bg3
summary vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. The outlines library is one of the backends used by vLLM to support structured output (a.k.a. guided decoding). Outlines provides an optional cache for its compiled grammars on the local filesystem. This cache has been on by default in vLLM. Outlines is also available by default through the OpenAI compatible API server. The affected code in vLLM is vllm/model_executor/guided_decoding/outlines_logits_processors.py, which unconditionally uses the cache from outlines. A malicious user can send a stream of very short decoding requests with unique schemas, resulting in an addition to the cache for each request. This can result in a Denial of Service if the filesystem runs out of space. Note that even if vLLM was configured to use a different backend by default, it is still possible to choose outlines on a per-request basis using the guided_decoding_backend key of the extra_body field of the request. This issue applies only to the V0 engine and is fixed in 0.8.0.
references
0
reference_url https://github.com/vllm-project/vllm/blob/53be4a863486d02bd96a59c674bbec23eec508f6/vllm/model_executor/guided_decoding/outlines_logits_processors.py
reference_id
reference_type
scores
0
value 6.5
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
url https://github.com/vllm-project/vllm/blob/53be4a863486d02bd96a59c674bbec23eec508f6/vllm/model_executor/guided_decoding/outlines_logits_processors.py
1
reference_url https://github.com/vllm-project/vllm/pull/14837
reference_id
reference_type
scores
0
value 6.5
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
url https://github.com/vllm-project/vllm/pull/14837
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-mgrm-fgjv-mhv8
reference_id
reference_type
scores
0
value 6.5
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
url https://github.com/vllm-project/vllm/security/advisories/GHSA-mgrm-fgjv-mhv8
fixed_packages
0
url pkg:pypi/vllm@0.8.0
purl pkg:pypi/vllm@0.8.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-5ec1-1h6d-tuaq
1
vulnerability VCID-e8w2-9rwg-u7ba
2
vulnerability VCID-fxgs-s1vm-8bez
3
vulnerability VCID-nctw-rz8h-f3af
4
vulnerability VCID-qake-z4ec-wkdu
5
vulnerability VCID-svzy-7pke-2bdr
6
vulnerability VCID-ugds-eqgw-fbbz
7
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.8.0
aliases CVE-2025-29770, GHSA-mgrm-fgjv-mhv8, PYSEC-2025-223
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-k1qz-xe9c-2bg3
4
url VCID-nctw-rz8h-f3af
vulnerability_id VCID-nctw-rz8h-f3af
summary vLLM is an inference and serving engine for large language models (LLMs). In versions from 0.6.4 to before 0.12.0, users can crash the vLLM engine serving multimodal models that use the Idefics3 vision model implementation by sending a specially crafted 1x1 pixel image. This causes a tensor dimension mismatch that results in an unhandled runtime error, leading to complete server termination. This issue has been patched in version 0.12.0.
references
0
reference_url https://github.com/vllm-project/vllm
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm
1
reference_url https://github.com/vllm-project/vllm/commit/0ec84221718d920c3f46da879cc354f94b8fb59e
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/commit/0ec84221718d920c3f46da879cc354f94b8fb59e
2
reference_url https://github.com/vllm-project/vllm/pull/29881
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/pull/29881
3
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-grg2-63fw-f2qr
reference_id
reference_type
scores
0
value 7.5
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
url https://github.com/vllm-project/vllm/security/advisories/GHSA-grg2-63fw-f2qr
4
reference_url https://nvd.nist.gov/vuln/detail/CVE-2026-22773
reference_id CVE-2026-22773
reference_type
scores
url https://nvd.nist.gov/vuln/detail/CVE-2026-22773
5
reference_url https://github.com/advisories/GHSA-grg2-63fw-f2qr
reference_id GHSA-grg2-63fw-f2qr
reference_type
scores
url https://github.com/advisories/GHSA-grg2-63fw-f2qr
fixed_packages
0
url pkg:pypi/vllm@0.12.0
purl pkg:pypi/vllm@0.12.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.12.0
aliases CVE-2026-22773, GHSA-grg2-63fw-f2qr, PYSEC-2026-143
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-nctw-rz8h-f3af
5
url VCID-svzy-7pke-2bdr
vulnerability_id VCID-svzy-7pke-2bdr
summary vLLM is an inference and serving engine for large language models (LLMs). In versions starting from 0.7.0 to before 0.9.0, in the file vllm/multimodal/hasher.py, the MultiModalHasher class has a security and data integrity issue in its image hashing method. Currently, it serializes PIL.Image.Image objects using only obj.tobytes(), which returns only the raw pixel data, without including metadata such as the image’s shape (width, height, mode). As a result, two images of different sizes (e.g., 30x100 and 100x30) with the same pixel byte sequence could generate the same hash value. This may lead to hash collisions, incorrect cache hits, and even data leakage or security risks. This issue has been patched in version 0.9.0.
references
0
reference_url https://github.com/vllm-project/vllm/commit/99404f53c72965b41558aceb1bc2380875f5d848
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/commit/99404f53c72965b41558aceb1bc2380875f5d848
1
reference_url https://github.com/vllm-project/vllm/pull/17378
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/pull/17378
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-c65p-x677-fgj6
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/security/advisories/GHSA-c65p-x677-fgj6
fixed_packages
0
url pkg:pypi/vllm@0.9.0
purl pkg:pypi/vllm@0.9.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-nctw-rz8h-f3af
1
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.9.0
aliases CVE-2025-46722, GHSA-c65p-x677-fgj6, PYSEC-2025-43
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-svzy-7pke-2bdr
6
url VCID-u659-sd9h-tkf3
vulnerability_id VCID-u659-sd9h-tkf3
summary vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. When vLLM is configured to use Mooncake, unsafe deserialization exposed directly over ZMQ/TCP on all network interfaces will allow attackers to execute remote code on distributed hosts. This is a remote code execution vulnerability impacting any deployments using Mooncake to distribute KV across distributed hosts. This vulnerability is fixed in 0.8.0.
references
0
reference_url https://github.com/vllm-project/vllm/commit/288ca110f68d23909728627d3100e5a8db820aa2
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/commit/288ca110f68d23909728627d3100e5a8db820aa2
1
reference_url https://github.com/vllm-project/vllm/pull/14228
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/pull/14228
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-x3m8-f7g5-qhm7
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/security/advisories/GHSA-x3m8-f7g5-qhm7
fixed_packages
0
url pkg:pypi/vllm@0.8.0
purl pkg:pypi/vllm@0.8.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-5ec1-1h6d-tuaq
1
vulnerability VCID-e8w2-9rwg-u7ba
2
vulnerability VCID-fxgs-s1vm-8bez
3
vulnerability VCID-nctw-rz8h-f3af
4
vulnerability VCID-qake-z4ec-wkdu
5
vulnerability VCID-svzy-7pke-2bdr
6
vulnerability VCID-ugds-eqgw-fbbz
7
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.8.0
aliases CVE-2025-29783, GHSA-x3m8-f7g5-qhm7, PYSEC-2025-63
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-u659-sd9h-tkf3
7
url VCID-ugds-eqgw-fbbz
vulnerability_id VCID-ugds-eqgw-fbbz
summary vLLM, an inference and serving engine for large language models (LLMs), has a Regular Expression Denial of Service (ReDoS) vulnerability in the file `vllm/entrypoints/openai/tool_parsers/pythonic_tool_parser.py` of versions 0.6.4 up to but excluding 0.9.0. The root cause is the use of a highly complex and nested regular expression for tool call detection, which can be exploited by an attacker to cause severe performance degradation or make the service unavailable. The pattern contains multiple nested quantifiers, optional groups, and inner repetitions which make it vulnerable to catastrophic backtracking. Version 0.9.0 contains a patch for the issue.
references
0
reference_url https://github.com/vllm-project/vllm/commit/4fc1bf813ad80172c1db31264beaef7d93fe0601
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/commit/4fc1bf813ad80172c1db31264beaef7d93fe0601
1
reference_url https://github.com/vllm-project/vllm/pull/18454
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/pull/18454
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-w6q7-j642-7c25
reference_id
reference_type
scores
url https://github.com/vllm-project/vllm/security/advisories/GHSA-w6q7-j642-7c25
fixed_packages
0
url pkg:pypi/vllm@0.9.0
purl pkg:pypi/vllm@0.9.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-nctw-rz8h-f3af
1
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.9.0
aliases CVE-2025-48887, PYSEC-2025-50
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-ugds-eqgw-fbbz
8
url VCID-w9kt-yaqy-47fb
vulnerability_id VCID-w9kt-yaqy-47fb
summary vLLM is a library for LLM inference and serving. vllm/model_executor/weight_utils.py implements hf_model_weights_iterator to load the model checkpoint, which is downloaded from huggingface. It uses the torch.load function and the weights_only parameter defaults to False. When torch.load loads malicious pickle data, it will execute arbitrary code during unpickling. This vulnerability is fixed in v0.7.0.
references
0
reference_url https://github.com/vllm-project/vllm/commit/d3d6bb13fb62da3234addf6574922a4ec0513d04
reference_id
reference_type
scores
0
value 8.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/commit/d3d6bb13fb62da3234addf6574922a4ec0513d04
1
reference_url https://github.com/vllm-project/vllm/pull/12366
reference_id
reference_type
scores
0
value 8.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/pull/12366
2
reference_url https://github.com/vllm-project/vllm/security/advisories/GHSA-rh4j-5rhw-hr54
reference_id
reference_type
scores
0
value 8.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
url https://github.com/vllm-project/vllm/security/advisories/GHSA-rh4j-5rhw-hr54
3
reference_url https://pytorch.org/docs/stable/generated/torch.load.html
reference_id
reference_type
scores
0
value 8.8
scoring_system cvssv3.1
scoring_elements CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
url https://pytorch.org/docs/stable/generated/torch.load.html
fixed_packages
0
url pkg:pypi/vllm@0.7.0
purl pkg:pypi/vllm@0.7.0
is_vulnerable true
affected_by_vulnerabilities
0
vulnerability VCID-737m-tpkz-qffm
1
vulnerability VCID-e8w2-9rwg-u7ba
2
vulnerability VCID-fxgs-s1vm-8bez
3
vulnerability VCID-k1qz-xe9c-2bg3
4
vulnerability VCID-nctw-rz8h-f3af
5
vulnerability VCID-svzy-7pke-2bdr
6
vulnerability VCID-u659-sd9h-tkf3
7
vulnerability VCID-ugds-eqgw-fbbz
8
vulnerability VCID-za3a-c9m1-jqgz
resource_url http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.7.0
aliases CVE-2025-24357, GHSA-rh4j-5rhw-hr54, PYSEC-2025-58
risk_score null
exploitability null
weighted_severity null
resource_url http://public2.vulnerablecode.io/vulnerabilities/VCID-w9kt-yaqy-47fb
Fixing_vulnerabilities
Risk_scorenull
Resource_urlhttp://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.6.6.post1