AI and machine learning: vulnerabilities and fixes

TensorFlow, PyTorch, Jupyter, LangChain and LLM tools. 54 CVEs tracked.

AI and machine learning has 54 tracked vulnerabilities, 10 critical and 1 exploited in the wild (CISA KEV). A fixed release or vendor patch exists for 41. Most recent: 2026-10-23. Highest priority: CVE-2026-59822 (score 73/100).

• CVE-2026-59822 HIGH 8.2 EXPLOITED UPGRADE BerriAI LiteLLM Improper Authentication Vulnerability
• CVE-2026-44181 CRITICAL 10.0 UPGRADE Jupyter Enterprise Gateway launches remote Jupyter Notebook kernels across distributed clusters like Apache Sp
• CVE-2026-44180 CRITICAL 9.8 UPGRADE Jupyter Enterprise Gateway launches remote Jupyter Notebook kernels across distributed clusters like Apache Sp
• CVE-2026-44182 CRITICAL 10.0 UPGRADE Jupyter Enterprise Gateway launches remote Jupyter Notebook kernels across distributed clusters like Apache Sp
• CVE-2026-65700 CRITICAL 9.8 h2oGPT through 0.2.1 contains a path traversal vulnerability in the OpenAI-compatible files API that allows un
• CVE-2026-73487 CRITICAL 9.8 UPGRADE Flowise before 3.1.3 contains a regex-based Python code validator bypass in CSV and Airtable Agent nodes that
• CVE-2026-35030 CRITICAL 9.1 UPGRADE LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format
• CVE-2026-35052 CRITICAL 9.8 UPGRADE D-Tale is the combination of a Flask back-end and a React front-end to view & analyze Pandas data structures
• CVE-2026-35029 HIGH 8.8 UPGRADE LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format
• CVE-2026-69256 HIGH 8.8 UPGRADE Flowise is a drag & drop user interface to build a customized large language model flow
• CVE-2026-67425 HIGH 8.6 Flyto2 Core is an execution kernel for automation and AI-agent workflows
• CVE-2026-67622 CRITICAL 9.9 UPGRADE Flowise through 3.1.4 contains an insecure direct object reference vulnerability in the OpenAI Assistants inte
• CVE-2026-71962 HIGH 7.5 UPGRADE Flowise versions 2.2.4 through 3.1.4 contain a missing authorization vulnerability in the POST /api/v1/openai-
• CVE-2026-43629 HIGH 8.1 UPGRADE llama.cpp builds b4882 through b9058 contain a heap buffer overflow vulnerability in the KV cache state restor
• CVE-2026-43631 HIGH 8.1 UPGRADE llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in the vocab pointer o
• CVE-2026-58659 HIGH 7.8 UPGRADE PyTorch Lightning through 2.6.5, fixed in commit d710d68, contains a remote code execution vulnerability in th
• CVE-2026-70472 HIGH 8.8 UPGRADE Flowise is a drag & drop user interface to build a customized large language model flow
• CVE-2026-63086 HIGH 8.6 text-generation-inference through 3.3.7 contains a server-side request forgery (SSRF) vulnerability in the Ope
• CVE-2026-65918 HIGH 7.1 UPGRADE PyTorch torchvision through 0.28.0, fixed in commit 4e05dc2, contains an out-of-bounds heap read vulnerability
• CVE-2026-107685 CRITICAL 9.1 Improper Certificate Validation vulnerability in the WebSocket streaming client in agentjido req_llm allows an
• CVE-2026-43632 HIGH 8.1 UPGRADE llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in llama-server affect
• CVE-2026-56340 HIGH 8.8 UPGRADE vLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in multimodal embeddings processing
• CVE-2026-73484 HIGH 8.1 UPGRADE Flowise before 3.1.3 contains a sandbox escape vulnerability in pythonCodeValidator.ts that fails to block nat
• CVE-2026-57845 HIGH 8.8 Hugging Face Datasets through 5.1.0 contains an insecure deserialization vulnerability that allows attackers t
• CVE-2026-114135 HIGH 8.8 Jupyter Server is the backend for Jupyter web applications
• CVE-2026-73264 HIGH 7.6 Prowler is a cloud security platform
• CVE-2026-114612 CRITICAL 9.9 LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format
• CVE-2026-43630 MEDIUM 6.5 UPGRADE llama.cpp builds b4882 before b8487 contain an out-of-bounds read vulnerability in the recurrent memory state
• CVE-2026-73079 HIGH 8.5 Sub2API is an AI API gateway platform designed to distribute and manage API quotas from AI product subscriptio
• CVE-2026-114140 HIGH 8.8 The nbconvert tool, jupyter nbconvert, converts Jupyter notebooks to various other formats via Jinja templates
• CVE-2026-17500 MEDIUM 5.3 A vulnerability was detected in ggml-org llama.cpp d006858/e15efe0
• CVE-2026-17501 MEDIUM 5.3 A flaw has been found in ggml-org llama.cpp e15efe0
• CVE-2026-48766 HIGH 7.6 TypeBot is a chatbot builder tool
• CVE-2026-12484 HIGH 7.8 A vulnerability in keras-team/keras version 3.15.0 allows unsafe deserialization of attacker-controlled PyTorc
• CVE-2026-73603 MEDIUM 5.3 UPGRADE Flowise before 3.1.4 fails to validate chatflow visibility in the unauthenticated text-to-speech endpoint, all
• CVE-2026-62186 HIGH 7.6 UPGRADE OpenClaw versions before 2026.6.8 contain an authorization bypass vulnerability in OpenAI-compatible HTTP mode
• CVE-2026-70640 HIGH 7.0 UPGRADE llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Androi
• CVE-2026-43622 HIGH 7.8 UPGRADE llama.cpp builds b1886 through b7445 contain a double free vulnerability in the LLaMA-Android JNI wrapper wher
• CVE-2026-43628 HIGH 7.8 UPGRADE llama.cpp builds b3978 through b9058 contain an integer underflow and out-of-bounds read vulnerability in the
• CVE-2026-70638 HIGH 7.8 UPGRADE llama.cpp builds b1886 through b7445 contain an integer overflow vulnerability in the LLaMA-Android JNI wrappe
• CVE-2026-75060 HIGH 8.4 UPGRADE In JetBrains PyCharm before 2026.2.1 code execution was possible via unauthenticated Jupyter MCP tools
• CVE-2026-15628 MEDIUM 6.3 A security flaw has been discovered in zhayujie chatgpt-on-wechat CowAgent up to 2.1.1
• CVE-2026-43627 HIGH 7.8 UPGRADE llama.cpp builds b1283 through b9058 contain an integer overflow vulnerability in the llama_batch_init() funct
• CVE-2026-48121 MEDIUM 6.7 @langchain/langgraph-checkpoint-mongodb provides a LangGraph.js CheckpointSaver implementation that uses Mongo
• CVE-2026-114510 HIGH 7.1 Flexible GraphRAG is a Python framework for building graph and retrieval-augmented generation applications acr
• CVE-2026-66005 MEDIUM 6.3 Jan through 0.8.4, fixed in commit 3e1c1e7, contains a CORS misconfiguration vulnerability in its local API se
• CVE-2026-15531 MEDIUM 5.3 A vulnerability has been found in yashbhalgat HashNeRF-pytorch up to 82885e698295982504eb6a26d060a6b2473e3706
• CVE-2026-48762 MEDIUM 5.4 TypeBot is a chatbot builder tool
• CVE-2026-64650 MEDIUM 6.3 The `@ai-sdk/harness-opencode` tool is an HarnessV1 adapter backed by @openai/codex-sdk, which drives the code
• CVE-2026-24271 MEDIUM 6.2 NVIDIA TensorRT-LLM contains a vulnerability in the OpenAI-compatible inference API, where an attacker could c
• CVE-2026-47475 MEDIUM 6.2 NVIDIA TensorRT-LLM contains a vulnerability in the OpenAI-compatible inference API where an attacker could tr
• CVE-2026-70639 MEDIUM 5.5 UPGRADE llama.cpp builds b1886 through b7445 contain a null pointer dereference vulnerability in the LLaMA-Android JNI
• CVE-2026-109296 MEDIUM 4.3 Open WebUI before 0.11.4 contains a missing authorization vulnerability that allows authenticated users to rea
• CVE-2026-18581 LOW 3.3 A vulnerability was determined in ggml-org llama.cpp e15efe0

FAQ

How do I fix AI and machine learning vulnerabilities?

Open the CVE for your version, check the affected range, then follow the listed action: upgrade to the fixed release, or downgrade or mitigate if no fix exists yet. Commands for each OS are shown on every CVE page.

Is AI and machine learning being exploited right now?

Yes: 1 AI and machine learning CVE is on the CISA Known Exploited Vulnerabilities list, so patch it first.

Which AI and machine learning CVE should I patch first?

CVE-2026-59822 currently ranks highest (73/100): BerriAI LiteLLM Improper Authentication Vulnerability.

Software in this group

People search for: ai, llm, tensorflow, pytorch, chatgpt, langchain.

← Back to latest CVEs