Default Brotli decompression in RestSharp lets a malicious server response trigger an out-of-bounds write and remote code execution
The vulnerability stems from AutomaticDecompression defaulting to a full mode that includes the Brotli algorithm, so the client automatically advertises br encoding when requesting. An attacker can craft an HTTP response that contains a malicious Brotli stream and abuse that default so the built-in .NET runtime decoder triggers a buffer overflow or out-of-bounds write while processing the data. Without user interaction and over a network-reachable path, this can lead to memory disclosure, process crashes, or remote arbitrary code execution, seriously harming confidentiality, integrity, and availability.
Component
RestSharp is a widely used .NET HTTP client library. Its core design simplifies RESTful API interaction by wrapping low-level network request details behind a unified abstraction, supports multiple serialization formats and automatic redirects, and mainly serves enterprise applications that need efficient communication with third-party services, improving convenience and standardization of network data exchange.
Type
Dependency on Vulnerable Third-Party Component (CWE-1395)
When initializing RestClient, explicitly disable Brotli automatic decompression and enable only safer encodings such as GZip or Deflate, or pull a fixed library version into project dependencies. Strengthen validation of third-party API responses and do not directly trust unauthenticated compressed streams.
CSSA Exclusive Early Warning
10.0 Critical
Missing validation of MCP tool configuration in ogx enables unauthenticated SSRF and internal resource disclosure
An attacker can abuse the lack of destination validation on the server_url parameter of MCP tool definitions under the /v1/responses endpoint. With authentication disabled by default, no credentials are needed to craft malicious requests that make the server open HTTP connections to arbitrary internal IPs or cloud metadata endpoints. Although the codebase contains a validate_url_not_private guard, it is not applied to that field, so the attacker can perform server-side request forgery, probe and reach restricted internal services, and relay credentials by injecting custom Header and Authorization tokens, leading to serious information disclosure and potential loss of system control through lateral movement.
Component
ogx is a reactive API service compatible with the OpenAI protocol. Its architecture supports plug-and-play models for complex interactions and aims to give enterprise applications flexible inference and tool-calling. It includes support for multimodal content, file search, and external model calls, with a highly extensible plugin design so users can integrate backend services to improve automation workflows, while staying lightweight in the default deployment to reduce operational complexity.
Apply non-private address validation uniformly to user-supplied URL parameters so loopback, link-local, and reserved addresses are rejected; restrict protocols to http or https; and enforce a hostname allowlist to strictly control outbound connection targets.
CVE Intelligence
2
CVE-2026-116199
CVSS
8.1 High
2026-10-25
Google SSO authentication bypass in Snipe-IT leads to privilege escalation
Before Snipe-IT 8.6.3, GoogleAuthController::handleGoogleCallback had an authentication-bypass flaw (CWE-285). It did not enforce a check of the user's activated flag, so a deactivated user whose username matched a Google account email could still sign in via Google SSO. A deactivated user with an email-format username and a still-valid Google identity can restore a session at their original privilege level, including administrator rights. All Snipe-IT instances not updated to 8.6.3 are affected. Exploitation needs no complex interaction—only a valid Google identity—to obtain unauthorized system access.
Component
Snipe-IT is open-source IT asset-management software for tracking hardware, software licenses, and other assets, and supports multiple authentication methods including Google SSO.
Risks
From ordinary user to administrator: If an attacker controls a deactivated account that still has a linked Google identity, they can restore administrator privileges directly
Complete system control: An attacker can use this vulnerability to perform arbitrary actions on the victim system and, depending on restored privileges, view/change/delete data or create new fully privileged accounts
No-user-interaction attack: An attacker can trigger the vulnerability with an existing Google identity credential and needs no additional user interaction
Regularly audit Google SSO bindings for deactivated users
Monitor anomalous login behavior, especially attempts from deactivated accounts
CVE-2026-116207
CVSS
9.1 Critical
2026-10-25
SQL injection in YesWiki SearchManager leads to sensitive data disclosure and denial of service
In YesWiki versions from 4.6.0 before 4.7.0 and from 5.0.0-alpha1 before 5.0.0-alpha4, SearchManager::prepareSearchRequest() has an unauthenticated SQL injection flaw (CWE-89). Field names from the public Bazar entry-search API query and searchfields parameters are placed inside backtick-quoted aliases and only escaped with mysqli_real_escape_string. An attacker can inject backticks to run subqueries that read password hashes and other data, or exhaust database threads with SLEEP(). All YesWiki users not updated to a fixed version are affected. No authentication is required for remote exploitation, leading to sensitive information disclosure or service unavailability.
Component
YesWiki is a PHP-based open-source wiki engine for collaborative content creation and management, widely used for internal knowledge bases and community documentation.
Risks
Sensitive data disclosure: An attacker can read password hashes, personally identifiable information, and other sensitive configuration data from the database
Denial of service: By running time-consuming functions such as SLEEP(), an attacker can exhaust database connection threads and interrupt service
Unauthenticated exploitation: The flaw sits on a public API; an attacker needs no credentials, so risk is extremely high
Immediately upgrade YesWiki to 4.7.0 or 5.0.0-alpha4 or later
Enforce strict input validation and filtering on the Bazar search API
Monitor the database for anomalous queries and sudden spikes in connection count
Package Poisoning
10
Package Poisoning
npm2026-10-26
util-foriepft@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
d40f3d0e9a085432770c1e80910a0857
Package Poisoning
PyPI2026-10-26
claudeopuseni@0.1.1
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
a669ef36d26a1724030c50cc1d37d3aa
Package Poisoning
npm2026-10-26
@s4yhiitestpoc/r2-control@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
125689aeaaf9ec5171c3c079e841246c
Package Poisoning
npm2026-10-26
@s4yhiitestpoc/scan-control-a@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
012b4e19101d5f74554d28a378f1dfde
Package Poisoning
npm2026-10-26
@s4yhiitestpoc/scan-dns-c@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
f6b99a737dcefed54930b03e840c7fc5
Package Poisoning
npm2026-10-26
css-flexgap-shim@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
9f855a42c5332031852c8a25d5e170d7
Package Poisoning
npm2026-10-26
css-grid-snap-util@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
5f3ac7a4fa5919949b97e83212dae2da
Package Poisoning
npm2026-10-26
css-hnhxwd-polyfill@>= 0
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
316f55c8159ab14a1dc4b23b697959bb
Package Poisoning
npm2026-10-26
@walknpm/todo-cli@2.0.1
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
MD5
f3c6c2f19a8dca61cd04b4204bbd9ab7
Package Poisoning
npm2026-10-26
dsh-plugin-blender-3d@0.2.1
flagged as malicious
This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.
