| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| vLLM is an inference and serving engine for large language models. In versions from 0.22.1 through 0.28.0, the operator-supplied model revision pin (--revision / --code-revision) is not propagated to several Hugging Face artifact loads for the FunAudioChat and Tarsier2 architectures: the WhisperFeatureExtractor and speech_tokenizer PreTrainedTokenizerFast loads in vllm/model_executor/models/funaudiochat.py and the Qwen2VLConfig.from_pretrained call used by Tarsier2ProcessingInfo in vllm/model_executor/models/qwen2_vl.py. As a result, deployments pinned to a reviewed revision still resolve these behavior-affecting processor, tokenizer, and config artifacts from the repository's default revision, so a later change to the upstream default branch can alter audio preprocessing, speech tokenizer behavior, or Tarsier2 configuration without any change to the operator's configured pin. This is a supply-chain integrity and reproducibility failure for pinned deployments; it is residual to the earlier fix tracked as GHSA-3ww4-5jv9-j5gm / CVE-2026-47155 and does not constitute remote code execution or a trust_remote_code=False bypass. The issue is fixed in version 0.28.0. |
| Contributor PHP Object Injection in Nested Pages <= 3.3.2 versions. |
| Unauthenticated Cross Site Scripting (XSS) in Post and Page Builder by BoldGrid <= 1.27.14 versions. |
| pgPointcloud through 1.2.5 contains a heap out-of-bounds read vulnerability in dimensional patch WKB deserialization that allows authenticated database users to read adjacent heap memory. Attackers can supply crafted pcpatch values with attacker-controlled size fields to copy heap memory into stored patches for exfiltration or crash the PostgreSQL backend. |
| TDuck survey form through 6.0 fails to validate write passwords on submission endpoints, enforcing the check only on the front end. Remote unauthenticated attackers can submit form entries directly to public submission APIs without providing the password by using the form key from share links. |
| Contrast (Edgeless Systems) before 1.8.1 logs the workload secret to stderr, and thus to Kubernetes logs, when the Contrast initializer is configured with CONTRAST_LOG_LEVEL set to info or debug. Because info is the default, all installations that do not customize the initializer log level are affected. This exposes workload secrets — normally accessible only to the Contrast Coordinator, the initializer, the seedshare owner, and the workload owner — to Kubernetes users with get or list permission on pods/logs and to anyone with read access to the Kubernetes log storage, such as the cloud provider. Deployments that do not use workload secrets are unaffected. |
| Heap-based buffer overflow in Microsoft Standard XPS allows an authorized attacker to elevate privileges over a network. |
| ** REJECT ** DO NOT USE THIS CANDIDATE NUMBER. ConsultIDs: none. Reason: This candidate was withdrawn by its CNA. Further investigation showed that it was not a security issue. Notes: The sole source documents PE-format conformance defects in innosetup-5.5.9.exe with no exploit, attack path, or untrusted search path condition (CWE-426/427), and the author states Windows loads these files normally; the record's remote/exploited claims are unsupported, as is the product maintainer's contention. |
| Integer underflow (wrap or wraparound) in Microsoft Standard XPS allows an authorized attacker to elevate privileges locally. |
| Heap-based buffer overflow in Microsoft Standard XPS allows an authorized attacker to elevate privileges locally. |
| Heap-based buffer overflow in Microsoft Standard XPS allows an authorized attacker to elevate privileges locally. |
| Out-of-bounds read in Microsoft Standard XPS allows an authorized attacker to disclose information locally. |
| Heap-based buffer overflow in Microsoft Standard XPS allows an authorized attacker to elevate privileges locally. |
| Heap-based buffer overflow in Microsoft Standard XPS allows an authorized attacker to elevate privileges over a network. |
| All versions of Zammad including the latest alpha enable the local zammad user to escalate privileges to root. |
| A vulnerability was found in ST Engineering iDirect Evolution and Velocity WebServer Evolution up to 20260717. This affects an unknown part of the component Location Header Handler. Performing a manipulation of the argument Host results in open redirect. It is possible to initiate the attack remotely. The exploit has been made public and could be used. The vendor was contacted early about this disclosure but did not respond in any way. |
| In MONAI 1.6.0, PersistentDataset (monai/data/dataset.py) explicitly rejects the combination track_meta=True with weights_only=True, forcing users who cache MetaTensors (the default tensor type in MONAI >= 1.0) to run torch.load(hashfile, weights_only=False). Related cache helpers in monai/data/utils.py also call pickle.loads on cached content and derive cache keys with hashlib.md5. As a result, a local user with write access to a shared or world-writable cache_dir (e.g. /tmp/monai_cache, HPC scratch, ~/.cache/monai) can place a malicious pickle file that is deserialized the next time another user's MONAI pipeline reads the cache, resulting in arbitrary code execution in that user's context. All released versions of the monai pip package are affected; no patched version is available as of the advisory. |
| MONAI through 1.6.0 contains an eval injection vulnerability in _get_fake_spatial_shape() in monai/bundle/scripts.py. The function validates shape expressions with a helper that walks the AST and only collects ast.Name nodes, rejecting any name other than 'p' or 'n', before passing the string to eval(). Expressions built solely from constants and attribute, subscript, or call nodes (for example "(1).__class__.__bases__[0].__subclasses__()" or "int.__class__.__init__.__globals__") contain no ast.Name nodes and therefore bypass the allowlist. Because the shape value originates from bundle metadata consumed by _get_real_input_data and verify_net_in_out (reachable through the bundle 'verify_net_in_out' CLI flow), an attacker who can influence a bundle's metadata can escape the eval sandbox via object introspection chains and achieve code execution in this non-default flow. |
| bbs-go through 4.4.6 contains a permission bypass vulnerability in the AdminMiddleware authorization logic where the read-only dashboard.user.view permission rule matches the /api/admin/user/synccount endpoint before the intended dashboard.user.update rule. Authenticated users with only view permissions can call the synccount endpoint to trigger expensive full-table user recounts and cache invalidations, causing denial of service through repeated concurrent database operations. |
| In JetBrains YouTrack before 2026.2.18991 sSRF via stored XHTML injection was possible during PDF export |
