ai_inference_server vulnerabilities
CVEs whose affected-version data names the ai_inference_server package (go, pip). Each record lists the affected and patched versions; check a specific version with the dependency checker or POST /api/sbom.
92 CVEsRSS
CVE-2026-42561High· 7.5python-multipart: python-multipart: Denial of Service via excessive multipart part headers (CVE-2026-42561)
A flaw was found in python-multipart. A remote attacker can exploit this denial of service (DoS) vulnerability by sending a specially crafted request with an excessive number of part headers or a single very large header value during multi…
CVE-2026-44431Medium· 5.3PoCurllib3 is an HTTP client library for Python
urllib3 is an HTTP client library for Python. From 1.23 to before 2.7.0, cross-origin redirects followed from the low-level API via ProxyManager.connection_from_url().urlopen(..., assert_same_host=False) still forward these sensitive hea…
CVE-2026-44222High· 7.5vllm: vLLM: Denial of Service via malformed multimodal input or token injection (CVE-2026-44222)
A flaw was found in vLLM, an inference and serving engine for large language models (LLMs). This vulnerability allows unauthenticated attackers to cause a Denial of Service (DoS) by supplying image or video placeholder sequences without ma…
CVE-2026-31221High· 8.0pytorch-lightning: PyTorch-Lightning: Arbitrary code execution via insecure deserialization of checkpoint files (CVE-2026-31221)
A flaw was found in PyTorch-Lightning. This vulnerability, categorized as insecure deserialization (CWE-502), exists in the checkpoint loading mechanism. A remote attacker can exploit this by providing a maliciously crafted checkpoint file…
CVE-2026-44223Medium· 6.5vLLM is an inference and serving engine for large language models (LLMs)
vLLM is an inference and serving engine for large language models (LLMs). From 0.18.0 to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step,…
CVE-2026-42308Medium· 6.2Pillow: Pillow: Denial of Service via integer overflow in font processing (CVE-2026-42308)
A flaw was found in Pillow, a Python imaging library. If a font advances for each glyph by an exceeding large amount, an integer overflow can occur when Pillow tracks the current position. This could lead to a denial of service (DoS) condi…
CVE-2026-42310Medium· 4.0Pillow: Pillow: Denial of Service via malicious PDF processing (CVE-2026-42310)
A flaw was found in Pillow, a Python imaging library. A remote attacker could supply a specially crafted malicious PDF file, causing the application to hang indefinitely and consume 100% CPU. This vulnerability leads to a Denial of Service…
CVE-2026-42311High· 7.8Pillow: python-pillow: Pillow: Arbitrary code execution via malicious PSD file processing (CVE-2026-42311)
A flaw was found in Pillow, a Python imaging library. An attacker could exploit this vulnerability by tricking a user into processing a specially crafted malicious PSD file. This could lead to memory corruption, potentially causing the app…
CVE-2026-41506Medium· 6.5golang: github.com/go-git/go-git: go-git: Information disclosure of HTTP authentication credentials via redirects (CVE-2026-41506)
A flaw was found in go-git, an extensible Git implementation library for Go. This vulnerability allows an attacker to potentially obtain sensitive HTTP authentication credentials. This can occur when go-git follows redirects during smart-H…
CVE-2026-42284High· 7.5GitPython: GitPython: Arbitrary code execution via improper validation of clone options (CVE-2026-42284)
A flaw was found in GitPython, a Python library for interacting with Git repositories. A remote attacker could exploit an input validation vulnerability in the `_clone()` function. By crafting a malicious string in the `multi_options` para…
CVE-2026-44244High· 7.3GitPython: GitPython: Arbitrary code execution via injected newlines in Git configuration (CVE-2026-44244)
A flaw was found in GitPython, a Python library used to interact with Git repositories. The `GitConfigParser.set_value()` function does not properly validate input for newlines. This vulnerability allows an attacker to inject malicious con…
CVE-2026-42215High· 7.5GitPython: GitPython: Arbitrary command execution due to bypass of dangerous Git option checks (CVE-2026-42215)
A flaw was found in GitPython, a Python library used to interact with Git repositories. This vulnerability allows an attacker to achieve arbitrary command execution by providing specially crafted arguments (kwargs) to functions such as Rep…
CVE-2026-6357Medium· 5.8pip: pip: Arbitrary code execution or information disclosure via malicious wheel package installation (CVE-2026-6357)
A flaw was found in pip. Prior to version 26.1, pip's self-update check functionality would execute after installing wheel packages. This process involved importing newly installed Python modules. A malicious actor could craft a specially …
CVE-2026-28684High· 7.1python-dotenv: python-dotenv: Arbitrary file overwrite via symbolic link following (CVE-2026-28684)
A flaw was found in python-dotenv. A local attacker can exploit this by crafting a symbolic link, which the `set_key()` and `unset_key()` functions in python-dotenv follow when rewriting `.env` files. This can lead to the overwriting of ar…
CVE-2026-3219Medium· 5.0pip: pip: Incorrect file installation due to improper archive handling (CVE-2026-3219)
A flaw was found in pip. This vulnerability occurs because pip incorrectly processes concatenated tar and ZIP files as ZIP files, regardless of their true format. This improper handling can lead to confusing installation behavior, potentia…
CVE-2026-40164High· 7.5jq is a command-line JSON processor
jq is a command-line JSON processor. Before commit 0c7d133c3c7e37c00b6d46b658a02244fdd3c784, jq used MurmurHash3 with a hardcoded, publicly visible seed (0x432A9843) for all JSON object hash table operations, which allowed an attacker to…
CVE-2026-1839Medium· 6.7transformers: HuggingFace Transformers: Arbitrary code execution via malicious checkpoint file (CVE-2026-1839)
A flaw was found in HuggingFace Transformers. A remote attacker can exploit this vulnerability by supplying a specially crafted checkpoint file (e.g., `rng_state.pth`). The `_load_rng_state()` method in the `Trainer` class loads this file …
CVE-2026-34588High· 8.8OpenEXR: OpenEXR: Arbitrary code execution and information disclosure via crafted EXR file (CVE-2026-34588)
A flaw was found in OpenEXR, an image storage format for the motion picture industry. A remote attacker could exploit an integer overflow vulnerability in the `internal_exr_undo_piz()` function by providing a specially crafted EXR file. Th…
CVE-2026-22815Medium· 5.3aiohttp: AIOHTTP: Denial of Service via insufficient header/trailer handling (CVE-2026-22815)
A flaw was found in AIOHTTP, an asynchronous HTTP client/server framework for Python. Insufficient restrictions in header and trailer handling could allow a remote attacker to cause uncapped memory usage. This can lead to a Denial of Servi…
CVE-2026-25645Medium· 4.7PoCrequests: Requests: Security bypass due to predictable temporary file creation (CVE-2026-25645)
A flaw was found in the `requests` HTTP library, specifically in the `requests.utils.extract_zipped_paths()` function, which is used to load Certificate Authority (CA) bundles. A local attacker can exploit this vulnerability by pre-creatin…
CVE-2026-26209Medium· 5.5⚖ disputedcbor2: cbor2: Denial of Service due to uncontrolled recursion via crafted CBOR payloads (CVE-2026-26209)
A flaw was found in cbor2, a library for encoding and decoding Concise Binary Object Representation (CBOR) data. A remote attacker can exploit this vulnerability by sending a specially crafted CBOR payload containing deeply nested structur…
CVE-2026-30922High· 7.5PoCpyasn1: pyasn1 Vulnerable to Denial of Service via Unbounded Recursion (CVE-2026-30922)
An unbounded recursion flaw has been discovered in the pypi pyasn1 library. This uncontrolled recursion occurs when decoding ASN.1 data with deeply nested structures. An attacker can supply a crafted payload containing nested SEQUENCE (0x3…
CVE-2026-4111High· 7.5A flaw was identified in the RAR5 archive decompression logic of the libarchive library, specifically within the archive_read_data() processing path
A flaw was identified in the RAR5 archive decompression logic of the libarchive library, specifically within the archive_read_data() processing path. When a specially crafted RAR5 archive is processed, the decompression routine may enter…
CVE-2025-14831Medium· 5.3A flaw was found in GnuTLS
A flaw was found in GnuTLS. This vulnerability allows a denial of service (DoS) by excessive CPU (Central Processing Unit) and memory consumption via specially crafted malicious certificates containing a large number of name constraints …
CVE-2026-22701Medium· 5.3filelock: filelock Time-of-Check-Time-of-Use (TOCTOU) in SoftFileLock (CVE-2026-22701)
A Time-of-Check-Time-of-Use (TOCTOU) flaw has been discovered in the pypi filelock package. The TOCTOU race condition vulnerability exists in the SoftFileLock implementation of the filelock package. An attacker with local filesystem access…
CVE-2025-66448High· 7.5vllm: vLLM: Remote Code Execution via malicious model configuration (CVE-2025-66448)
A remote code execution vulnerability has been identified in vLLM. An attacker can exploit a weakness in the model loading process to silently fetch and run unauthorized, malicious Python code on the host system. This happens because the e…
CVE-2025-62426Medium· 6.5vllm: vLLM vulnerable to DoS via large Chat Completion or Tokenization requests with specially crafted `chat_template_kwargs` (CVE-2025-624…
A vulnerability in vLLM allows an authenticated user to trigger unintended tokenization during chat template processing by supplying crafted chat_template_kwargs to the /v1/chat/completions or /tokenize endpoints. By forcing the server to …
CVE-2025-59425High· 7.5vllm: Timing Attack in vLLM API Token Verification Leading to Authentication Bypass (CVE-2025-59425)
A flaw was found in vLLM’s API token authentication logic, where token comparisons were not performed in constant time. This weakness could allow an attacker to exploit timing differences to guess valid tokens and bypass authentication.
CVE-2025-48956High· 7.5vllm: HTTP header size limit not enforced allows Denial of Service from Unauthenticated requests (CVE-2025-48956)
A flaw was found in vLLM. A denial of service (DoS) vulnerability can be triggered by sending a single HTTP GET request with an extremely large X-Forwarded-For header to an HTTP endpoint. This results in server memory exhaustion, potential…
CVE-2025-57809High· 7.5xgrammar: XGrammar affected by Denial of Service by infinite recursion grammars (CVE-2025-57809)
A flaw was found in xgrammar. Recursive grammar definitions could trigger infinite recursion during parsing in GrammarMatcherBase::ExpandEquivalentStackElements, leading to unbounded stack growth and a segmentation fault. This vulnerabilit…