{"answers":{"affected":"4 open source packages ship an affected version, including pytorch and torch. It affects 1 product.","exploited":"There is no known exploitation of CVE-2025-32434 in the wild and no public exploit code.","fix":"Upgrade pytorch to 2.6.0 or later.","what":"CVE-2025-32434 is a critical unsafe deserialization vulnerability in pytorch, published on 18 Apr 2025. An attacker can exploit it over the internet without logging in, with no action from the victim."},"fix_prompt":"Check this codebase for CVE-2025-32434 and fix it if it is affected.\n\nWhat it is: CVE-2025-32434 is a critical unsafe deserialization vulnerability in pytorch, published on 18 Apr 2025. An attacker can exploit it over the internet without logging in, with no action from the victim.\nExploitation: There is no known exploitation of CVE-2025-32434 in the wild and no public exploit code.\n\nAffected packages and fixed versions:\n- pytorch (Bitnami): affected all versions; fixed in 2.6.0\n- torch (PyPI): affected all versions; fixed in 2.6.0\n- torch (PyPI): affected all versions; fixed in 2.6.0\n- torch (PyPI): affected all versions; fixed in 2.6.0\n\nTask:\n1. Search the dependency manifests and lockfiles (for example package.json, package-lock.json, requirements.txt, poetry.lock, uv.lock, pom.xml, build.gradle, go.mod, Cargo.lock, Gemfile.lock) and any Dockerfiles for the affected packages, including transitive dependencies.\n2. If an affected version is in use, upgrade it to the fixed version for its release line with the project's package manager, so the lockfile is regenerated. Prefer the smallest upgrade that fixes it.\n3. If there is no fix, or the upgrade would break the build, do not guess: explain the risk and the safest workaround instead.\n4. Run the tests and summarise what changed and why.\n5. If nothing in the project is affected, say so and change nothing.\n\nReference: https://insights.sbomify.com/CVE-2025-32434/ (Markdown: https://insights.sbomify.com/CVE-2025-32434/index.md)","id":"CVE-2025-32434","markdown":"https://insights.sbomify.com/CVE-2025-32434/index.md","name":"CVE-2025-32434","record":{"aliases":["AZL-60877","AZL-60880","BIT-pytorch-2025-32434","CLSA-2025-1760660047","DEBIAN-CVE-2025-32434","DLA-4389-1","GHSA-53q9-r3pm-6pq6","PYSEC-2025-41"],"analysis":{"component_type":"library","fix_type":"upgrade","impacts":["code_execution"],"network_reachable":"yes","non_default_config":false,"product_category":"PyTorch machine learning library","requires_auth":"none","summary":"Attackers can achieve remote code execution by tricking users into loading malicious models with weights_only enabled in PyTorch.","user_interaction":"not_required","vuln_class":"deserialization"},"analysis_generated_by":"qwen3.8-flash-next","assigner":"GitHub_M","cpe":[{"product":"pytorch","vendor":"linuxfoundation"}],"cvss":{"score":9.3,"severity":"critical","source":"cna","vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N","version":"4.0"},"cwes":[{"id":"CWE-502","name":"Deserialization of Untrusted Data"}],"description":"PyTorch is a Python package that provides tensor computation with strong GPU acceleration and deep neural networks built on a tape-based autograd system. In version 2.5.1 and prior, a Remote Command Execution (RCE) vulnerability exists in PyTorch when loading a model using torch.load with weights_only=True. This issue has been patched in version 2.6.0.","ecosystems":["Bitnami","PyPI"],"exploits":{"exploitdb":0,"first_seen":"2025-11-13","github_poc":4,"max_poc_stars":2,"metasploit":0,"nuclei":0,"refs":[{"date":"2026-06-30T20:00:55Z","kind":"github_poc","ref":"https://github.com/AlexanderGumeniuk/CVE-2025-32434","stars":2,"title":null},{"date":"2025-12-15T19:59:13Z","kind":"github_poc","ref":"https://github.com/B1tBit/CVE-2025-32434-exploit","stars":1,"title":"A script for exploiting a vulnerability in PyTorch with subsequent RCE in library versions \u003c 2.6.0"},{"date":"2025-11-13T04:27:05Z","kind":"github_poc","ref":"https://github.com/cyhe50/cve-2025-32434-poc","stars":0,"title":null},{"date":"2026-03-08T11:15:50Z","kind":"github_poc","ref":"https://github.com/Soildworks/Agentic-CLIP-Benchmark","stars":0,"title":"An automated, high-precision zero-shot evaluation pipeline for OpenAI's CLIP model on CIFAR-10. Features 88.80% accuracy, Safetensors security mitigation (CVE-2025-32434), and AI Native (Trae) workflow."}],"tier":"poc_unverified"},"has_fix":true,"headline":"PyTorch: `torch.load` with `weights_only=True` leads to remote code execution","id":"CVE-2025-32434","indexable":true,"kev":null,"kind":"cve","links":{"ecosystem":"Bitnami","product":null,"products":[],"type":"deserialization","year":"2025"},"n_packages":4,"n_references":1,"next":{"headline":"Hydra no restricted eval after nix-eval-jobs migration","id":"CVE-2025-32435"},"packages":[{"ecosystem":"Bitnami","fixed":["2.6.0"],"introduced":["0"],"last_affected":[],"package":"pytorch","purl":"pkg:bitnami/pytorch","source":"BIT-pytorch-2025-32434"},{"ecosystem":"PyPI","fixed":["2.6.0"],"introduced":["0"],"last_affected":[],"package":"torch","purl":null,"source":"GHSA-53q9-r3pm-6pq6"},{"ecosystem":"PyPI","fixed":["2.6.0"],"introduced":["0"],"last_affected":[],"package":"torch","purl":"pkg:pypi/torch","source":"GHSA-53q9-r3pm-6pq6"},{"ecosystem":"PyPI","fixed":["2.6.0"],"introduced":["0"],"last_affected":[],"package":"torch","purl":"pkg:pypi/torch","source":"PYSEC-2025-41"}],"prev":{"headline":"Erlang Erlang/OTP SSH Server Missing Authentication for Critical Function","id":"CVE-2025-32433"},"products":["pytorch"],"published":"2025-04-18","related":[{"cvss_score":9.8,"headline":"Deserialization in n/a","id":"CVE-2024-48063","kev_date_added":null,"published":"2024-10-29"},{"cvss_score":9.8,"headline":"PyTorch vulnerable to arbitrary code execution","id":"CVE-2022-45907","kev_date_added":null,"published":"2022-11-26"},{"cvss_score":8.8,"headline":"PyTorch Vulnerable to Remote Code Execution via Untrusted Checkpoint Files","id":"CVE-2026-24747","kev_date_added":null,"published":"2026-01-27"},{"cvss_score":7.8,"headline":"Pytorch use-after-free vulnerability","id":"CVE-2024-31583","kev_date_added":null,"published":"2024-04-17"},{"cvss_score":7.5,"headline":"Denial of service in n/a","id":"CVE-2025-55551","kev_date_added":null,"published":"2025-09-25"},{"cvss_score":7.5,"headline":"Denial of service in n/a","id":"CVE-2025-55553","kev_date_added":null,"published":"2025-09-25"}],"sources":[{"class":"attribution","licence":"Apache-2.0","name":"Bitnami","prefix":"BIT"},{"class":"attribution","licence":"CVE Terms of Use (MITRE)","name":"CVE List","prefix":"CVE"},{"class":"attribution","licence":"CC BY 4.0","name":"GitHub Advisory Database","prefix":"GHSA"},{"class":"attribution","licence":"CC BY 4.0","name":"PyPA Advisory Database","prefix":"PYSEC"},{"class":"attribution","licence":"US public information, NVD API terms","name":"NVD"},{"class":"attribution","licence":"CWE Terms of Use","name":"CWE (MITRE)"},{"class":"attribution","licence":"FIRST CVSS licence","name":"CVSS (FIRST)"},{"class":"open","licence":"CC0","name":"PoC-in-GitHub"}],"ssvc":{"automatable":"yes","exploitation":"none","technical_impact":"total"},"state":"PUBLISHED","updated":"2025-12-01","vendors":["pytorch"]},"summary":"Attackers can achieve remote code execution by tricking users into loading malicious models with weights_only enabled in PyTorch.","url":"https://insights.sbomify.com/CVE-2025-32434/"}