| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| The mamba language model framework thru 2.2.6 is vulnerable to insecure deserialization (CWE-502) when loading pre-trained models from HuggingFace Hub. The MambaLMHeadModel.from_pretrained() method uses torch.load() to load the pytorch_model.bin weight file without enabling the security-restrictive weights_only=True parameter. This allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by publishing a malicious model repository on HuggingFace Hub. When a victim loads a model from this repository, arbitrary code is executed on the victim's system in the context of the mamba process. |
| The CosyVoice project thru commit 6e01309e01bc93bbeb83bdd996b1182a81aaf11e (2025-30-21) contains an insecure deserialization vulnerability (CWE-502) in its model loading process. When loading model files (.pt) from a user-specified directory (via the --model_dir argument), the code uses torch.load() without the security-restrictive weights_only=True parameter. This allows the deserialization of arbitrary Python objects via the Pickle module. An attacker can exploit this by providing a maliciously crafted model directory containing .pt files with embedded pickle payloads. When a victim loads this directory using CosyVoice's web interface, the malicious payload is executed, leading to remote code execution on the victim's system. |
| Horovod thru 0.28.1 contains an insecure deserialization vulnerability (CWE-502) in its KVStore HTTP server component. The KVStore server, used for distributed task coordination, lacks authentication and authorization controls, allowing any remote attacker to write arbitrary data via HTTP PUT requests. When a Horovod worker reads data from the KVStore (via HTTP GET), it deserializes the data using cloudpickle.loads() without verifying its source or integrity. An attacker can exploit this by sending a malicious pickle payload to the server before the legitimate data is written, causing the victim worker to deserialize and execute arbitrary code, leading to remote code execution. |
| The imgaug library thru 0.4.0 contains an insecure deserialization vulnerability in its BackgroundAugmenter class within the multicore.py module. The class uses Python's pickle module to deserialize data received via a multiprocessing queue in the _augment_images_worker() method without any safety checks. An attacker who can influence the data placed into this queue (e.g., through social engineering, malicious input scripts, or a compromised shared queue) can provide a malicious pickle payload. When deserialized, this payload can execute arbitrary code in the context of the worker process, leading to remote or local code execution depending on the deployment scenario. |
| The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization (CWE-502) through its predict() method. When a user provides a dataset file path to the predict() method, the framework automatically determines the file format. If the file is a pickle (.pkl) file, it is loaded using pandas.read_pickle() without any validation or security restrictions. This allows the deserialization of arbitrary Python objects via the unsafe pickle module. A remote attacker can exploit this by providing a maliciously crafted pickle file, leading to arbitrary code execution on the system running the Ludwig prediction. |
| HCL AION is affected by a vulnerability where basic authorization tokens are used for authentication. Use of basic authorization mechanisms may expose credentials to potential interception or misuse, especially if not combined with secure transmission practices. |
| JunoClaw is an agentic AI platform built on Juno Network. Prior to 0.x.y-security-1, the upload_wasm MCP tool accepted a filesystem path from the agent and uploaded whatever bytes the path resolved to, with no validation of location, symlink target, file size, or file format. This vulnerability is fixed in 0.x.y-security-1. |
| HCL AION is affected by a vulnerability where sensitive information may be included in URL parameters. Passing sensitive data in URLs may expose it through browser history, logs, or intermediary systems, potentially leading to unintended information disclosure under certain conditions. |
| n8n-MCP is an MCP server that provides AI assistants access to n8n node documentation, properties, and operations. Prior to version 2.47.13, when n8n-mcp runs in HTTP transport mode, authenticated MCP tools/call requests had their full arguments and JSON-RPC params written to server logs by the request dispatcher and several sibling code paths before any redaction. When a tool call carries credential material — most notably n8n_manage_credentials.data — the raw values can be persisted in logs. In deployments where logs are collected, forwarded to external systems, or viewable outside the request trust boundary (shared log storage, SIEM pipelines, support/ops access), this can result in disclosure of: bearer tokens and OAuth credentials sent through n8n_manage_credentials, per-tenant API keys and webhook auth headers embedded in tool arguments, arbitrary secret-bearing payloads passed to any MCP tool. The issue requires authentication (AUTH_TOKEN accepted by the server), so unauthenticated callers cannot trigger it; the runtime exposure is also reduced by an existing console-silencing layer in HTTP mode, but that layer is fragile and the values are still constructed and passed into the logger. This issue has been patched in version 2.47.13. |
| n8n-MCP is an MCP server that provides AI assistants access to n8n node documentation, properties, and operations. Prior to version 2.47.11, when n8n-mcp runs in HTTP transport mode, incoming requests to the POST /mcp endpoint had their request metadata written to server logs regardless of the authentication outcome. In deployments where logs are collected, forwarded to external systems, or viewable outside the request trust boundary (shared log storage, SIEM pipelines, support/ops access), this can result in disclosure of: bearer tokens from the Authorization header, per-tenant API keys from the, x-n8n-key header in multi-tenant setups, JSON-RPC request payloads sent to the MCP endpoint. Access control itself was not bypassed — unauthenticated requests were correctly rejected with 401 Unauthorized — but sensitive values from those rejected requests could still be persisted in logs. This issue has been patched in version 2.47.11. |
| A vulnerability in SQL Expressions allows an authenticated attacker to read arbitrary files from the Grafana server's filesystem. Only instances with the sqlExpressions feature toggle enabled are vulnerable. |
| On 2026-05-11, between approximately 19:20 and 19:26 UTC, 84 malicious versions across 42 @tanstack/* packages were published to the npm registry. The publishes were authenticated via the legitimate GitHub Actions OIDC trusted-publisher binding for TanStack/router, but the publish workflow itself was not modified. The attacker chained three known vulnerability classes — a pull_request_target "Pwn Request" misconfiguration, GitHub Actions cache poisoning across the fork↔base trust boundary, and runtime memory extraction of the OIDC token from the Actions runner process — to publish credential-stealing malware under a trusted identity. Each affected package received exactly two malicious versions, published a few minutes apart. |
| Vvveb before 1.0.8.3 contains a directory listing information disclosure vulnerability that allows unauthenticated attackers to enumerate files and directories by accessing multiple paths lacking proper index directives in .htaccess files. Attackers can access directories such as admin asset paths, plugins, themes, and media folders to view filenames, file sizes, modification timestamps, and unrendered admin templates containing sensitive route maps. |
| OpenTelemetry eBPF Instrumentation provides eBPF instrumentation based on the OpenTelemetry standard. From 0.4.0 to before 0.8.0, a flaw in the Java agent injection path allows a local attacker controlling a Java workload to overwrite arbitrary host files when Java injection is enabled and OBI is running with elevated privileges. The injector trusted TMPDIR from the target process and used unsafe file creation semantics, enabling both filesystem boundary escape and symlink-based file clobbering. This vulnerability is fixed in 0.8.0. |
| vm2 is an open source vm/sandbox for Node.js. In 3.10.5, NodeVM's require.root path restriction can be bypassed using filesystem symlinks, allowing sandboxed code to load modules from outside the allowed root directory in host context. Because path validation uses path.resolve() (which does not dereference symlinks) but module loading uses Node's native require() (which does), an attacker can load arbitrary host-realm modules and achieve remote code execution. This vulnerability is fixed in 3.11.0. |
| gitoxide is an implementation of git written in Rust. Prior to 0.21.1, a malicious tree can be constructed that will, when checked out with gitoxide, permit writing an attacker-controlled symlink into any existing directory the user has write access to. During checkout, all symlink index entries are deferred and created after regular files using a single shared gix_worktree::Stack. Internally, this uses a gix_fs::Stack. gix_fs::Stack::make_relative_path_current() caches validated path prefixes: when the previously-processed leaf component exactly matches the leading component(s) of the next path, the leaf-to-directory transition at gix-fs/src/stack.rs invokes only delegate.push_directory(), never delegate.push(). In gix_worktree::stack::delegate::StackDelegate, when the state member is State::CreateDirectoryAndAttributesStack, Attributes::push_directory() only loads attributes (from the ODB, in the clone case), and does not perform any other checks. The on-disk symlink_metadata() check and unlink-on-collision live in StackDelegate::push()'s invocation of create_leading_directory(), which is therefore bypassed for the cached prefix. The final symlink is created with plain std::os::unix::fs::symlink, which follows symlinks in parent directories. Therefore, it's possible to provide a tree with duplicate symlink and directory entries that exploits this. This vulnerability is fixed in 0.21.1. |
| A logging issue was addressed with improved data redaction. This issue is fixed in iOS 18.7.9 and iPadOS 18.7.9, iOS 26.5 and iPadOS 26.5, macOS Sequoia 15.7.7, macOS Sonoma 14.8.7, macOS Tahoe 26.5, tvOS 26.5, watchOS 26.5. An app may be able to determine kernel memory layout. |
| This issue was addressed with improved checks. This issue is fixed in macOS Tahoe 26.5. An attacker with physical access to a locked device may be able to view sensitive user information. |
| curl might erroneously pass on credentials for a first proxy to a second
proxy.
This can happen when the following conditions are true:
1. curl is setup to use specific different proxies for different URL schemes
2. the first proxy needs credentials
3. the second proxy uses no credentials
4. while using the first proxy (using say `http://`), curl is asked to follow
a redirect to a URL using another scheme (say `https://`), accessed using a
second, different, proxy |
| Deserialization of untrusted data in Microsoft Office SharePoint allows an authorized attacker to execute code over a network. |
