Высокий · CVSS 7.1EPSS 0,33%Публичный эксплойтЕсть исправление
Denial of Service in Keras via Excessive Memory Allocation in HDF5 Metadata
Allocation of Resources Without Limits or Throttling in the HDF5 weight loading component in Google Keras 3.0.0 through 3.13.0 on all platforms allows a remote attacker to cause a Denial of Service (DoS) through memory exhaustion and a crash of the Python interpreter via a crafted .keras archive containing a valid model.weights.h5 file whose dataset declares an extremely large shape.
BlackSheep ClientSession is vulnerable to CRLF injection
BlackSheep is an asynchronous web framework to build event based web applications with Python. Prior to 2.4.6, the HTTP Client implementation in BlackSheep is vulnerable to CRLF injection. Missing headers validation makes it possible for an attacker to modify the HTTP requests (e.g. insert a new header) or even create a new HTTP request. Exploitation requires developers to pass unsanitized user input directly into headers.The server part is not affected because BlackSheep delegates to an underlying ASGI server handling of response headers. This vulnerability is fixed in 2.4.6.
Weblate is a web based localization tool. Prior to 5.15.2, the screenshot images were served directly by the HTTP server without proper access control. This could allow an unauthenticated user to access screenshots after guessing their filename. This vulnerability is fixed in 5.15.2.
Chainlit versions prior to 2.8.5 contain an authorization bypass through user-controlled key vulnerability. If this vulnerability is exploited, threads may be viewed or thread ownership may be obtained by an attacker who can log in to the product.
GuardDog Path Traversal Vulnerability Leads to Arbitrary File Overwrite and RCE
GuardDog is a CLI tool to identify malicious PyPI packages. Prior to 2.7.1, there is a path traversal vulnerability exists in GuardDog's safe_extract() function that allows malicious PyPI packages to write arbitrary files outside the intended extraction directory, leading to Arbitrary File Overwrite and Remote Code Execution on systems running GuardDog. This vulnerability is fixed in 2.7.1.
GuardDog Zip Bomb Vulnerability in safe_extract() Allows DoS
GuardDog is a CLI tool to identify malicious PyPI packages. Prior to 2.7.1, GuardDog's safe_extract() function does not validate decompressed file sizes when extracting ZIP archives (wheels, eggs), allowing attackers to cause denial of service through zip bombs. A malicious package can consume gigabytes of disk space from a few megabytes of compressed data. This vulnerability is fixed in 2.7.1.
LangChain versions up to and including 0.3.1 contain a regular expression denial-of-service (ReDoS) vulnerability in the MRKLOutputParser.parse() method (libs/langchain/langchain/agents/mrkl/output_parser.py). The parser applies a backtracking-prone regular expression when extracting tool actions from model output. An attacker who can supply or influence the parsed text (for example via prompt injection in downstream applications that pass LLM output directly into MRKLOutputParser.parse()) can trigger excessive CPU consumption by providing a crafted payload, causing significant parsing delays and a denial-of-service condition.
LlamaIndex (run-llama/llama_index) versions up to and including 0.11.6 contain an unsafe deserialization vulnerability in BGEM3Index.load_from_disk() in llama_index/indices/managed/bge_m3/base.py. The function uses pickle.load() to deserialize multi_embed_store.pkl from a user-supplied persist_dir without validation. An attacker who can provide a crafted persist directory containing a malicious pickle file can trigger arbitrary code execution when the victim loads the index from disk.
LlamaIndex (run-llama/llama_index) versions up to and including 0.12.2 contain an uncontrolled resource consumption vulnerability in the VannaPack VannaQueryEngine implementation. The custom_query() logic generates SQL statements from a user-supplied prompt and executes them via vn.run_sql() without enforcing query execution limits In downstream deployments where untrusted users can supply prompts, an attacker can trigger expensive or unbounded SQL operations that exhaust CPU or memory resources, resulting in a denial-of-service condition. The vulnerable execution path occurs in llama_index/packs/vanna/base.py within custom_query().
hermes's raw options logging may disclose secrets passed in via subcommand options argument
hermes is an implementation of the HERMES workflow to automatize software publication with rich metadata. From 0.8.1 to before 0.9.1, hermes subcommands take arbitrary options under the -O argument. These have been logged in raw form. If users provide sensitive data such as API tokens (e.g., via hermes deposit -O invenio_rdm.auth_token SECRET), these are written to the log file in plain text, making them available to whoever can access the log file. This vulnerability is fixed in 0.9.1.
wlc may leak API keys due to an insecure API key configuration
wlc is a Weblate command-line client using Weblate's REST API. Prior to 1.17.0, wlc supported providing unscoped API keys in the setting. This practice was discouraged for years, but the code was never removed. This might cause the API key to be leaked to different servers.
wlc is a Weblate command-line client using Weblate's REST API. Prior to 1.17.0, the SSL verification would be skipped for some crafted URLs. This vulnerability is fixed in 1.17.0.
Высокий · CVSS 8.6EPSS 0,28%Нет опубликованного исправления
Label Studio vulnerable to full account takeover by chaining Stored XSS + IDOR in User Profile via custom_hotkeys field
Label Studio is a multi-type data labeling and annotation tool. In 1.22.0 and earlier, a persistent stored cross-site scripting (XSS) vulnerability exists in the custom_hotkeys functionality of the application. An authenticated attacker (or one who can trick a user/administrator into updating their custom_hotkeys) can inject JavaScript code that executes in other users’ browsers when those users load any page using the templates/base.html template. Because the application exposes an API token endpoint (/api/current-user/token) to the browser and lacks robust CSRF protection on some API endpoints, the injected script may fetch the victim’s API
MindsDB has improper sanitation of filepath that leads to information disclosure and DOS
MindsDB is a platform for building artificial intelligence from enterprise data. Prior to version 25.11.1, an unauthenticated path traversal in the file upload API lets any caller read arbitrary files from the server filesystem and move them into MindsDB’s storage, exposing sensitive data. The PUT handler in file.py directly joins user-controlled data into a filesystem path when the request body is JSON and source_type is not "url". Only multipart uploads and URL-sourced uploads receive sanitization; JSON uploads lack any call to clear_filename or equivalent checks. This vulnerability is fixed in 25.11.1.
MLFlow versions up to and including 3.4.0 are vulnerable to DNS rebinding attacks due to a lack of Origin header validation in the MLFlow REST server. This vulnerability allows malicious websites to bypass Same-Origin Policy protections and execute unauthorized calls against REST endpoints. An attacker can query, update, and delete experiments via the affected endpoints, leading to potential data exfiltration, destruction, or manipulation. The issue is resolved in version 3.5.0.
A vulnerability was found in AcademySoftwareFoundation OpenColorIO up to 2.5.0. This issue affects the function ConvertToRegularExpression of the file src/OpenColorIO/FileRules.cpp. Performing a manipulation results in out-of-bounds read. The attack needs to be approached locally. The exploit has been made public and could be used. The patch is named ebdbb75123c9d5f4643e041314e2bc988a13f20d. To fix this issue, it is recommended to deploy a patch. The fix was added to the 2.5.1 milestone.
A security flaw has been discovered in lief-project LIEF up to 0.17.1. Affected by this issue is the function Parser::parse_binary of the file src/ELF/Parser.tcc of the component ELF Binary Parser. The manipulation results in null pointer dereference. The attack must be initiated from a local position. The exploit has been released to the public and may be used for attacks. Upgrading to version 0.17.2 can resolve this issue. The patch is identified as 81bd5d7ea0c390563f1c4c017c9019d154802978. It is recommended to upgrade the affected component.
Высокий · CVSS 7.5EPSS 0,34%Публичный эксплойтЕсть исправление
ComfyUI-Manager is Vulnerable to CRLF Injection in Configuration Handler
ComfyUI-Manager is an extension designed to enhance the usability of ComfyUI. Prior to versions 3.39.2 and 4.0.5, an attacker can inject special characters into HTTP query parameters to add arbitrary configuration values to the config.ini file. This can lead to security setting tampering or modification of application behavior. This issue has been patched in versions 3.39.2 and 4.0.5.
vLLM is vulnerable to DoS in Idefics3 vision models via image payload with ambiguous dimensions
vLLM is an inference and serving engine for large language models (LLMs). In versions from 0.6.4 to before 0.12.0, users can crash the vLLM engine serving multimodal models that use the Idefics3 vision model implementation by sending a specially crafted 1x1 pixel image. This causes a tensor dimension mismatch that results in an unhandled runtime error, leading to complete server termination. This issue has been patched in version 0.12.0.