{"answers":{"affected":"22 open source packages ship an affected version, including tensorflow, tensorflow-cpu and tensorflow-gpu. It affects 1 product.","exploited":"There is no known exploitation of CVE-2021-37678 in the wild and no public exploit code.","fix":"Upgrade tensorflow to a fixed version for your release line: 2.3.4, 2.4.3 or 2.5.1.","what":"CVE-2021-37678 is a critical unsafe deserialization vulnerability in Google Tensorflow, published on 12 Aug 2021. It can only be exploited from the same machine or local network, with no action from the victim."},"fix_prompt":"Check this codebase for CVE-2021-37678 and fix it if it is affected.\n\nWhat it is: CVE-2021-37678 is a critical unsafe deserialization vulnerability in Google Tensorflow, published on 12 Aug 2021. It can only be exploited from the same machine or local network, with no action from the victim.\nExploitation: There is no known exploitation of CVE-2021-37678 in the wild and no public exploit code.\n\nAffected packages and fixed versions:\n- tensorflow (Bitnami): affected from 2.3.0; fixed in 2.3.4, 2.4.3, 2.5.1\n- tensorflow (PyPI): affected all versions; fixed in 2.3.4\n- tensorflow (PyPI): affected all versions; fixed in 2.3.4\n- tensorflow (PyPI): affected from 2.3.0; fixed in 2.3.4, 2.4.3\n- tensorflow (PyPI): affected from 2.4.0; fixed in 2.4.3\n- tensorflow (PyPI): affected from 2.4.0; fixed in 2.4.3\n- tensorflow (PyPI): affected from 2.5.0; fixed in 2.5.1\n- tensorflow (PyPI): affected from 2.5.0; fixed in 2.5.1\n- tensorflow-cpu (PyPI): affected all versions; fixed in 2.3.4\n- tensorflow-cpu (PyPI): affected all versions; fixed in 2.3.4\n- tensorflow-cpu (PyPI): affected from 2.3.0; fixed in 2.3.4, 2.4.3\n- tensorflow-cpu (PyPI): affected from 2.4.0; fixed in 2.4.3\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-2021-37678/ (Markdown: https://insights.sbomify.com/CVE-2021-37678/index.md)","id":"CVE-2021-37678","markdown":"https://insights.sbomify.com/CVE-2021-37678/index.md","name":"CVE-2021-37678","record":{"aliases":["BIT-tensorflow-2021-37678","GHSA-r6jx-9g48-2r5r","PYSEC-2021-300","PYSEC-2021-591","PYSEC-2021-789"],"analysis":{"component_type":"library","fix_type":"upgrade","impacts":["code_execution"],"network_reachable":"adjacent_or_local","non_default_config":false,"product_category":"TensorFlow machine learning library","requires_auth":"none","summary":"Attackers can execute arbitrary code by tricking TensorFlow into deserializing malicious Keras models in YAML format.","user_interaction":"not_required","vuln_class":"deserialization"},"analysis_generated_by":"qwen3.8-flash-next","assigner":"GitHub_M","cpe":[{"product":"tensorflow","vendor":"google"}],"cvss":{"score":9.3,"severity":"critical","source":"cna","vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H","version":"3.1"},"cwes":[{"id":"CWE-502","name":"Deserialization of Untrusted Data"}],"description":"TensorFlow is an end-to-end open source platform for machine learning. In affected versions TensorFlow and Keras can be tricked to perform arbitrary code execution when deserializing a Keras model from YAML format. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/python/keras/saving/model_config.py#L66-L104) uses `yaml.unsafe_load` which can perform arbitrary code execution on the input. Given that YAML format support requires a significant amount of work, we have removed it for now. We have patched the issue in GitHub commit 23d6383eb6c14084a8fc3bdf164043b974818012. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range.","ecosystems":["Bitnami","PyPI"],"exploits":{"exploitdb":0,"first_seen":"2021-09-09","github_poc":1,"max_poc_stars":2,"metasploit":0,"nuclei":0,"refs":[{"date":"2021-09-09T12:55:55Z","kind":"github_poc","ref":"https://github.com/fran-CICS/ExploitTensorflowCVE-2021-37678","stars":2,"title":"TP Seguridad Informática UTN FRBA 2021"}],"tier":"poc_unverified"},"has_fix":true,"headline":"Arbitrary code execution due to YAML deserialization","id":"CVE-2021-37678","indexable":true,"kev":null,"kind":"cve","links":{"ecosystem":"Bitnami","product":"google:tensorflow","products":["google:tensorflow"],"type":"deserialization","year":"2021"},"n_packages":22,"n_references":2,"next":{"headline":"Heap OOB in nested `tf.map_fn` with `RaggedTensor`s in TensorFlow","id":"CVE-2021-37679"},"packages":[{"ecosystem":"Bitnami","fixed":["2.3.4","2.4.3","2.5.1"],"introduced":["2.3.0","2.4.0","2.5.0"],"last_affected":[],"package":"tensorflow","purl":"pkg:bitnami/tensorflow","source":"BIT-tensorflow-2021-37678"},{"ecosystem":"PyPI","fixed":["2.3.4"],"introduced":["0"],"last_affected":[],"package":"tensorflow","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4"],"introduced":["0"],"last_affected":[],"package":"tensorflow","purl":"pkg:pypi/tensorflow","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4","2.4.3"],"introduced":["2.3.0","2.4.0"],"last_affected":[],"package":"tensorflow","purl":"pkg:pypi/tensorflow","source":"PYSEC-2021-300"},{"ecosystem":"PyPI","fixed":["2.4.3"],"introduced":["2.4.0"],"last_affected":[],"package":"tensorflow","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.4.3"],"introduced":["2.4.0"],"last_affected":[],"package":"tensorflow","purl":"pkg:pypi/tensorflow","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.5.1"],"introduced":["2.5.0"],"last_affected":[],"package":"tensorflow","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.5.1"],"introduced":["2.5.0"],"last_affected":[],"package":"tensorflow","purl":"pkg:pypi/tensorflow","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4"],"introduced":["0"],"last_affected":[],"package":"tensorflow-cpu","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4"],"introduced":["0"],"last_affected":[],"package":"tensorflow-cpu","purl":"pkg:pypi/tensorflow-cpu","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4","2.4.3"],"introduced":["2.3.0","2.4.0"],"last_affected":[],"package":"tensorflow-cpu","purl":"pkg:pypi/tensorflow-cpu","source":"PYSEC-2021-591"},{"ecosystem":"PyPI","fixed":["2.4.3"],"introduced":["2.4.0"],"last_affected":[],"package":"tensorflow-cpu","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.4.3"],"introduced":["2.4.0"],"last_affected":[],"package":"tensorflow-cpu","purl":"pkg:pypi/tensorflow-cpu","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.5.1"],"introduced":["2.5.0"],"last_affected":[],"package":"tensorflow-cpu","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.5.1"],"introduced":["2.5.0"],"last_affected":[],"package":"tensorflow-cpu","purl":"pkg:pypi/tensorflow-cpu","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4"],"introduced":["0"],"last_affected":[],"package":"tensorflow-gpu","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4"],"introduced":["0"],"last_affected":[],"package":"tensorflow-gpu","purl":"pkg:pypi/tensorflow-gpu","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.3.4","2.4.3"],"introduced":["2.3.0","2.4.0"],"last_affected":[],"package":"tensorflow-gpu","purl":"pkg:pypi/tensorflow-gpu","source":"PYSEC-2021-789"},{"ecosystem":"PyPI","fixed":["2.4.3"],"introduced":["2.4.0"],"last_affected":[],"package":"tensorflow-gpu","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.4.3"],"introduced":["2.4.0"],"last_affected":[],"package":"tensorflow-gpu","purl":"pkg:pypi/tensorflow-gpu","source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.5.1"],"introduced":["2.5.0"],"last_affected":[],"package":"tensorflow-gpu","purl":null,"source":"GHSA-r6jx-9g48-2r5r"},{"ecosystem":"PyPI","fixed":["2.5.1"],"introduced":["2.5.0"],"last_affected":[],"package":"tensorflow-gpu","purl":"pkg:pypi/tensorflow-gpu","source":"GHSA-r6jx-9g48-2r5r"}],"prev":{"headline":"Missing validation in shape inference for `Dequantize` in TensorFlow","id":"CVE-2021-37677"},"products":["tensorflow"],"published":"2021-08-12","related":[{"cvss_score":9.8,"headline":"TensorFlow vulnerable to heap out-of-buffer read in the QuantizeAndDequantize operation","id":"CVE-2023-25668","kev_date_added":null,"published":"2023-03-24"},{"cvss_score":9.8,"headline":"Integer Overflow or Wraparound in Google TensorFlow","id":"CVE-2018-7575","kev_date_added":null,"published":"2019-04-24"},{"cvss_score":9.1,"headline":"Path traversal in n/a","id":"CVE-2021-35958","kev_date_added":null,"published":"2021-06-30"},{"cvss_score":9,"headline":"Integer truncation in Shard API usage","id":"CVE-2020-15202","kev_date_added":null,"published":"2020-09-25"},{"cvss_score":9,"headline":"Data leak in Tensorflow","id":"CVE-2020-15205","kev_date_added":null,"published":"2020-09-25"},{"cvss_score":9,"headline":"Denial of Service in Tensorflow","id":"CVE-2020-15206","kev_date_added":null,"published":"2020-09-25"}],"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":null,"exploitation":null,"technical_impact":null},"state":"PUBLISHED","updated":"2024-08-04","vendors":["tensorflow"]},"summary":"Attackers can execute arbitrary code by tricking TensorFlow into deserializing malicious Keras models in YAML format.","url":"https://insights.sbomify.com/CVE-2021-37678/"}