{"answers":{"affected":"3 open source packages ship an affected version, including transformers. It affects 1 product.","exploited":"It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-1839 exists.","fix":"Upgrade transformers to a fixed version for your release line: 5.0.0 or 5.0.0rc3.","what":"CVE-2026-1839 is a medium-severity unsafe deserialization vulnerability in huggingface/transformers, published on 7 Apr 2026. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link."},"fix_prompt":"Check this codebase for CVE-2026-1839 and fix it if it is affected.\n\nWhat it is: CVE-2026-1839 is a medium-severity unsafe deserialization vulnerability in huggingface/transformers, published on 7 Apr 2026. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link.\nExploitation: It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-1839 exists.\n\nAffected packages and fixed versions:\n- transformers (PyPI): affected all versions; fixed in 5.0.0\n- transformers (PyPI): affected all versions; fixed in 5.0.0rc3\n- transformers (PyPI): affected all versions; fixed in 5.0.0rc3\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-2026-1839/ (Markdown: https://insights.sbomify.com/CVE-2026-1839/index.md)","id":"CVE-2026-1839","markdown":"https://insights.sbomify.com/CVE-2026-1839/index.md","name":"CVE-2026-1839","record":{"aliases":["CLSA-2026-1779065183","ECHO-9bd3-047a-4b4a","GHSA-69w3-r845-3855","PYSEC-2026-2288"],"analysis":{"component_type":"library","fix_type":"upgrade","impacts":["code_execution"],"network_reachable":"adjacent_or_local","non_default_config":false,"product_category":"HuggingFace Transformers library","requires_auth":"none","summary":"Attackers can execute arbitrary code by supplying malicious checkpoint files to the Trainer class due to unsafe deserialization in torch.load.","user_interaction":"required","vuln_class":"deserialization"},"analysis_generated_by":"qwen3.8-flash-next","assigner":"@huntr_ai","cpe":[{"product":"transformers","vendor":"huggingface"}],"cvss":{"score":6.5,"severity":"medium","source":"cna","vector":"CVSS:3.0/AV:L/AC:H/PR:N/UI:R/S:U/C:H/I:L/A:H","version":"3.0"},"cwes":[{"id":"CWE-502","name":"Deserialization of Untrusted Data"}],"description":"A vulnerability in the HuggingFace Transformers library, specifically in the `Trainer` class, allows for arbitrary code execution. The `_load_rng_state()` method in `src/transformers/trainer.py` at line 3059 calls `torch.load()` without the `weights_only=True` parameter. This issue affects all versions of the library supporting `torch\u003e=2.2` when used with PyTorch versions below 2.6, as the `safe_globals()` context manager provides no protection in these versions. An attacker can exploit this vulnerability by supplying a malicious checkpoint file, such as `rng_state.pth`, which can execute arbitrary code when loaded. The issue is resolved in version v5.0.0rc3.","ecosystems":["PyPI"],"exploits":{"exploitdb":0,"first_seen":null,"github_poc":0,"max_poc_stars":null,"metasploit":0,"nuclei":0,"refs":[],"tier":"weaponised"},"has_fix":true,"headline":"Arbitrary Code Execution via Unsafe torch.load() in Trainer Checkpoint Loading in huggingface/transformers","id":"CVE-2026-1839","indexable":true,"kev":null,"kind":"cve","links":{"ecosystem":"PyPI","product":null,"products":[],"type":"deserialization","year":"2026"},"n_packages":3,"n_references":2,"next":{"headline":"Missing authentication for critical function in Hubbell Aclara Metrum Cellular Web Interface","id":"CVE-2026-1840"},"packages":[{"ecosystem":"PyPI","fixed":["5.0.0"],"introduced":["0"],"last_affected":[],"package":"transformers","purl":"pkg:pypi/transformers","source":"PYSEC-2026-2288"},{"ecosystem":"PyPI","fixed":["5.0.0rc3"],"introduced":["0"],"last_affected":[],"package":"transformers","purl":null,"source":"GHSA-69w3-r845-3855"},{"ecosystem":"PyPI","fixed":["5.0.0rc3"],"introduced":["0"],"last_affected":[],"package":"transformers","purl":"pkg:pypi/transformers","source":"GHSA-69w3-r845-3855"}],"prev":{"headline":"Hostel \u003c= 1.1.6 - Reflected Cross-Site Scripting via 'shortcode_id' Parameter","id":"CVE-2026-1838"},"products":["huggingface/transformers"],"published":"2026-04-07","related":[{"cvss_score":9.6,"headline":"Deserialization of Untrusted Data in huggingface/transformers","id":"CVE-2023-7018","kev_date_added":null,"published":"2023-12-20"},{"cvss_score":9,"headline":"Deserialization of Untrusted Data in huggingface/transformers","id":"CVE-2023-6730","kev_date_added":null,"published":"2023-12-19"},{"cvss_score":8.8,"headline":"Hugging Face Transformers MaskFormer Model Deserialization of Untrusted Data Remote Code Execution Vulnerability","id":"CVE-2024-11393","kev_date_added":null,"published":"2024-11-22"},{"cvss_score":8.8,"headline":"Hugging Face Transformers Trax Model Deserialization of Untrusted Data Remote Code Execution Vulnerability","id":"CVE-2024-11394","kev_date_added":null,"published":"2024-11-22"},{"cvss_score":8,"headline":"Policy Bypass in LightGlue Nested Config Resolution in huggingface/transformers","id":"CVE-2026-5241","kev_date_added":null,"published":"2026-06-03"},{"cvss_score":7.8,"headline":"Arbitrary Remote Code Execution via `_attn_implementation_internal` Config Injection in huggingface/transformers","id":"CVE-2026-4372","kev_date_added":null,"published":"2026-05-24"},{"cvss_score":7.8,"headline":"Hugging Face Transformers Perceiver Model Deserialization of Untrusted Data Remote Code Execution Vulnerability","id":"CVE-2025-14920","kev_date_added":null,"published":"2025-12-23"}],"sources":[{"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)"}],"ssvc":{"automatable":"no","exploitation":"poc","technical_impact":"total"},"state":"PUBLISHED","updated":"2026-04-07","vendors":["huggingface"]},"summary":"Attackers can execute arbitrary code by supplying malicious checkpoint files to the Trainer class due to unsafe deserialization in torch.load.","url":"https://insights.sbomify.com/CVE-2026-1839/"}