{"id":"GHSA-ggpf-24jw-3fcw","aliases":[],"status":"published","title":"CVE-2025-24357 Malicious model remote code execution fix bypass with PyTorch \u003c 2.6.0","description":"## Description\n\nhttps://github.com/vllm-project/vllm/security/advisories/GHSA-rh4j-5rhw-hr54 reported a vulnerability where loading a malicious model could result in code execution on the vllm host. The fix applied to specify `weights_only=True` to calls to `torch.load()` did not solve the problem prior to PyTorch 2.6.0.\n\nPyTorch has issued a new CVE about this problem: https://github.com/advisories/GHSA-53q9-r3pm-6pq6\n\nThis means that versions of vLLM using PyTorch before 2.6.0 are vulnerable to this problem.\n## Background Knowledge\nWhen users install VLLM according to the official manual\n![image](https://github.com/user-attachments/assets/d17e0bdb-26f2-46d6-adf6-0b17e5ddf5c7)\n\nBut the version of PyTorch is specified in the requirements. txt file\n![image](https://github.com/user-attachments/assets/94aad622-ad6d-4741-b772-c342727c58c7)\n\nSo by default when the user install VLLM, it will install the PyTorch with version 2.5.1\n![image](https://github.com/user-attachments/assets/04ff31b0-aad1-490a-963d-00fda91da47b)\n\nIn CVE-2025-24357, weights_only=True was used for patching, but we know this is not secure.\nBecause we found that using Weights_only=True in pyTorch before 2.5.1 was unsafe\n\nHere, we use this interface to prove that it is not safe.\n![image](https://github.com/user-attachments/assets/0d86efcd-2aad-42a2-8ac6-cc96b054c925)\n\n\n## Fix\nupdate PyTorch version to 2.6.0\n\n## Credit\nThis vulnerability was found By Ji'an Zhou and Li'shuo Song","severity":{"version":"3.1","score":9.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","level":"critical","from":"ghsa"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-1395"],"published":"2025-04-23T02:26:06Z","modified":"2026-08-07T08:12:20Z","fixes":[{"ecosystem":"PyPI","package":"vllm","fixed_in":["0.8.0"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"PyPI","package":"vllm","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"0.8.0"}]}]},{"source":"osv","ecosystem":"PyPI","package":"vllm","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"0.8.0"}]}],"versions":["0.0.1","0.1.0","0.1.1","0.1.2","0.1.3","0.1.4","0.1.5","0.1.6","0.1.7","0.2.0","0.2.1","0.2.1.post1","0.2.2","0.2.3","0.2.4","0.2.5","0.2.6","0.2.7","0.3.0","0.3.1","0.3.2","0.3.3","0.4.0","0.4.0.post1","0.4.1","0.4.2","0.4.3","0.5.0","0.5.0.post1","0.5.1","0.5.2","0.5.3","0.5.3.post1","0.5.4","0.5.5","0.6.0","0.6.1","0.6.1.post1","0.6.1.post2","0.6.2","0.6.3","0.6.3.post1","0.6.4","0.6.4.post1","0.6.5","0.6.6","0.6.6.post1","0.7.0","0.7.1","0.7.2","0.7.3"]}],"references":[{"url":"https://github.com/pytorch/pytorch/security/advisories/GHSA-53q9-r3pm-6pq6","tags":["web"]},{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-ggpf-24jw-3fcw","tags":["web"]},{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-rh4j-5rhw-hr54","tags":["web"]},{"url":"https://github.com/advisories/GHSA-ggpf-24jw-3fcw"},{"url":"https://github.com/vllm-project/vllm","tags":["package"]}],"sources":{"ghsa":[{"id":"GHSA-ggpf-24jw-3fcw","title":"CVE-2025-24357 Malicious model remote code execution fix bypass with PyTorch \u003c 2.6.0","description":"## Description\n\nhttps://github.com/vllm-project/vllm/security/advisories/GHSA-rh4j-5rhw-hr54 reported a vulnerability where loading a malicious model could result in code execution on the vllm host. The fix applied to specify `weights_only=True` to calls to `torch.load()` did not solve the problem prior to PyTorch 2.6.0.\n\nPyTorch has issued a new CVE about this problem: https://github.com/advisories/GHSA-53q9-r3pm-6pq6\n\nThis means that versions of vLLM using PyTorch before 2.6.0 are vulnerable to this problem.\n## Background Knowledge\nWhen users install VLLM according to the official manual\n![image](https://github.com/user-attachments/assets/d17e0bdb-26f2-46d6-adf6-0b17e5ddf5c7)\n\nBut the version of PyTorch is specified in the requirements. txt file\n![image](https://github.com/user-attachments/assets/94aad622-ad6d-4741-b772-c342727c58c7)\n\nSo by default when the user install VLLM, it will install the PyTorch with version 2.5.1\n![image](https://github.com/user-attachments/assets/04ff31b0-aad1-490a-963d-00fda91da47b)\n\nIn CVE-2025-24357, weights_only=True was used for patching, but we know this is not secure.\nBecause we found that using Weights_only=True in pyTorch before 2.5.1 was unsafe\n\nHere, we use this interface to prove that it is not safe.\n![image](https://github.com/user-attachments/assets/0d86efcd-2aad-42a2-8ac6-cc96b054c925)\n\n\n## Fix\nupdate PyTorch version to 2.6.0\n\n## Credit\nThis vulnerability was found By Ji'an Zhou and Li'shuo Song","status":"published","published":"2025-04-23T02:26:06Z","modified":"2025-04-23T02:26:07Z","cvss":[{"version":"3.1","score":9.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","level":"critical"}],"cwes":["CWE-1395"],"updated_at":"2026-11-02T14:38:40.628275Z"}],"osv":[{"id":"GHSA-ggpf-24jw-3fcw","title":"CVE-2025-24357 Malicious model remote code execution fix bypass with PyTorch \u003c 2.6.0","description":"## Description\n\nhttps://github.com/vllm-project/vllm/security/advisories/GHSA-rh4j-5rhw-hr54 reported a vulnerability where loading a malicious model could result in code execution on the vllm host. The fix applied to specify `weights_only=True` to calls to `torch.load()` did not solve the problem prior to PyTorch 2.6.0.\n\nPyTorch has issued a new CVE about this problem: https://github.com/advisories/GHSA-53q9-r3pm-6pq6\n\nThis means that versions of vLLM using PyTorch before 2.6.0 are vulnerable to this problem.\n## Background Knowledge\nWhen users install VLLM according to the official manual\n![image](https://github.com/user-attachments/assets/d17e0bdb-26f2-46d6-adf6-0b17e5ddf5c7)\n\nBut the version of PyTorch is specified in the requirements. txt file\n![image](https://github.com/user-attachments/assets/94aad622-ad6d-4741-b772-c342727c58c7)\n\nSo by default when the user install VLLM, it will install the PyTorch with version 2.5.1\n![image](https://github.com/user-attachments/assets/04ff31b0-aad1-490a-963d-00fda91da47b)\n\nIn CVE-2025-24357, weights_only=True was used for patching, but we know this is not secure.\nBecause we found that using Weights_only=True in pyTorch before 2.5.1 was unsafe\n\nHere, we use this interface to prove that it is not safe.\n![image](https://github.com/user-attachments/assets/0d86efcd-2aad-42a2-8ac6-cc96b054c925)\n\n\n## Fix\nupdate PyTorch version to 2.6.0\n\n## Credit\nThis vulnerability was found By Ji'an Zhou and Li'shuo Song","status":"published","published":"2025-04-23T02:26:06Z","modified":"2026-08-07T08:12:20.513960856Z","cvss":[{"version":"3.1","score":9.8,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","level":"critical"}],"cwes":["CWE-1395"],"updated_at":"2026-11-02T15:46:47.822182Z"}]}}