{"url":"http://public2.vulnerablecode.io/api/packages/49112?format=json","purl":"pkg:pypi/vllm@0.14.1","type":"pypi","namespace":"","name":"vllm","version":"0.14.1","qualifiers":{},"subpath":"","is_vulnerable":true,"next_non_vulnerable_version":"0.20.0","latest_non_vulnerable_version":"0.20.0","affected_by_vulnerabilities":[{"url":"http://public2.vulnerablecode.io/api/vulnerabilities/37260?format=json","vulnerability_id":"VCID-za3a-c9m1-jqgz","summary":"vLLM is an inference and serving engine for large language models (LLMs). From 0.7.0 to before 0.19.0, the VideoMediaIO.load_base64() method at vllm/multimodal/media/video.py splits video/jpeg data URLs by comma to extract individual JPEG frames, but does not enforce a frame count limit. The num_frames parameter (default: 32), which is enforced by the load_bytes() code path, is completely bypassed in the video/jpeg base64 path. An attacker can send a single API request containing thousands of comma-separated base64-encoded JPEG frames, causing the server to decode all frames into memory and crash with OOM. This vulnerability is fixed in 0.19.0.","references":[{"reference_url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-pq5c-rjhq-qp7p","reference_id":"","reference_type":"","scores":[{"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-pq5c-rjhq-qp7p"}],"fixed_packages":[{"url":"http://public2.vulnerablecode.io/api/packages/49120?format=json","purl":"pkg:pypi/vllm@0.19.0","is_vulnerable":true,"affected_by_vulnerabilities":[{"vulnerability":"VCID-jzjy-kj6h-4bas"}],"resource_url":"http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.19.0"}],"aliases":["CVE-2026-34755","GHSA-pq5c-rjhq-qp7p","PYSEC-2026-144"],"risk_score":null,"exploitability":null,"weighted_severity":null,"resource_url":"http://public2.vulnerablecode.io/vulnerabilities/VCID-za3a-c9m1-jqgz"}],"fixing_vulnerabilities":[{"url":"http://public2.vulnerablecode.io/api/vulnerabilities/49873?format=json","vulnerability_id":"VCID-ggsq-9qgx-vyf6","summary":"vLLM vulnerable to Server-Side Request Forgery (SSRF) through MediaConnector\nA Server-Side Request Forgery (SSRF) vulnerability exists in the `MediaConnector` class within the vLLM project's multimodal feature set. The load_from_url and load_from_url_async methods obtain and process media from URLs provided by users, using different Python parsing libraries when restricting the target host. These two parsing libraries have different interpretations of backslashes, which allows the host name restriction to be bypassed. This allows an attacker to coerce the vLLM server into making arbitrary requests to internal network resources.\n\nThis vulnerability is particularly critical in containerized environments like `llm-d`, where a compromised vLLM pod could be used to scan the internal network, interact with other pods, and potentially cause Denial of Service or access sensitive data. For example, an attacker could make the vLLM pod send malicious requests to an internal `llm-d` management endpoint, leading to system instability by falsely reporting metrics like the KV cache state.","references":[{"reference_url":"https://github.com/vllm-project/vllm","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm"},{"reference_url":"https://github.com/vllm-project/vllm/commit/f46d576c54fb8aeec5fc70560e850bed38ef17d7","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/commit/f46d576c54fb8aeec5fc70560e850bed38ef17d7"},{"reference_url":"https://github.com/vllm-project/vllm/pull/32746","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/pull/32746"},{"reference_url":"https://nvd.nist.gov/vuln/detail/CVE-2026-24779","reference_id":"CVE-2026-24779","reference_type":"","scores":[],"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-24779"},{"reference_url":"https://github.com/advisories/GHSA-qh4c-xf7m-gxfc","reference_id":"GHSA-qh4c-xf7m-gxfc","reference_type":"","scores":[],"url":"https://github.com/advisories/GHSA-qh4c-xf7m-gxfc"},{"reference_url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-qh4c-xf7m-gxfc","reference_id":"GHSA-qh4c-xf7m-gxfc","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-qh4c-xf7m-gxfc"}],"fixed_packages":[{"url":"http://public2.vulnerablecode.io/api/packages/49112?format=json","purl":"pkg:pypi/vllm@0.14.1","is_vulnerable":true,"affected_by_vulnerabilities":[{"vulnerability":"VCID-za3a-c9m1-jqgz"}],"resource_url":"http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.14.1"}],"aliases":["CVE-2026-24779","GHSA-qh4c-xf7m-gxfc"],"risk_score":null,"exploitability":null,"weighted_severity":null,"resource_url":"http://public2.vulnerablecode.io/vulnerabilities/VCID-ggsq-9qgx-vyf6"},{"url":"http://public2.vulnerablecode.io/api/vulnerabilities/49908?format=json","vulnerability_id":"VCID-q8jt-32dy-w7cp","summary":"vLLM has RCE In Video Processing\n**A chain of vulnerabilities in vLLM allow Remote Code Execution (RCE):**\n\n1. **Info Leak** - PIL error messages expose memory addresses, bypassing ASLR\n2. **Heap Overflow** - JPEG2000 decoder in OpenCV/FFmpeg has a heap overflow that lets us hijack code execution\n\n**Result:** Send a malicious video URL to vLLM Completions or Invocations **for a video model** -> Execute arbitrary commands on the server\n\nCompletely default vLLM instance directly from pip, or docker, does not have authentication so \"None\" privileges are required, but even with non-default api-key enabled configuration this exploit is feasible through invocations route that allows payload to execute pre-auth.\n\nExample heap target is provided, other heap targets can be exploited as well to achieve rce. Leak allows for simple ASLR bypass. Leak + heap overflow achieves RCE on versions prior to 0.14.1.\n\nDeployments not serving a video model are not affected.\n\n---","references":[{"reference_url":"https://github.com/vllm-project/vllm","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm"},{"reference_url":"https://github.com/vllm-project/vllm/pull/31987","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/pull/31987"},{"reference_url":"https://github.com/vllm-project/vllm/pull/32319","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/pull/32319"},{"reference_url":"https://github.com/vllm-project/vllm/releases/tag/v0.14.1","reference_id":"","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/releases/tag/v0.14.1"},{"reference_url":"https://nvd.nist.gov/vuln/detail/CVE-2026-22778","reference_id":"CVE-2026-22778","reference_type":"","scores":[],"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-22778"},{"reference_url":"https://github.com/advisories/GHSA-4r2x-xpjr-7cvv","reference_id":"GHSA-4r2x-xpjr-7cvv","reference_type":"","scores":[],"url":"https://github.com/advisories/GHSA-4r2x-xpjr-7cvv"},{"reference_url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-4r2x-xpjr-7cvv","reference_id":"GHSA-4r2x-xpjr-7cvv","reference_type":"","scores":[],"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-4r2x-xpjr-7cvv"}],"fixed_packages":[{"url":"http://public2.vulnerablecode.io/api/packages/49112?format=json","purl":"pkg:pypi/vllm@0.14.1","is_vulnerable":true,"affected_by_vulnerabilities":[{"vulnerability":"VCID-za3a-c9m1-jqgz"}],"resource_url":"http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.14.1"}],"aliases":["CVE-2026-22778","GHSA-4r2x-xpjr-7cvv"],"risk_score":null,"exploitability":null,"weighted_severity":null,"resource_url":"http://public2.vulnerablecode.io/vulnerabilities/VCID-q8jt-32dy-w7cp"}],"risk_score":null,"resource_url":"http://public2.vulnerablecode.io/packages/pkg:pypi/vllm@0.14.1"}