Threat briefs

CVE-2025-12805 describes a flaw in Red Hat OpenShift AI (RHOAI) llama-stack-operator that allows unauthorized access to Llama Stack services in other namespaces via direct network requests due to missing NetworkPolicy restrictions, potentially enabling attackers to view or manipulate sensitive data.

Multiple critical vulnerabilities in n8n, including prototype pollution, code injection, and SQL injection, allow authenticated users to achieve remote code execution, read sensitive files, and perform unauthorized database operations.

Doveadm credentials are verified using direct comparison, making it susceptible to timing oracle attacks, allowing attackers to determine credentials and gain full access.

CVE-2025-59032 describes a vulnerability in ManageSieve's AUTHENTICATE command, where using a literal as a SASL initial response can crash the ManageSieve service, leading to a denial-of-service condition.

CVE-2026-22744 is a code injection vulnerability in Spring AI's RedisFilterExpressionConverter which allows an attacker to inject arbitrary commands into RediSearch TAG blocks via unescaped user-controlled strings, affecting versions 1.0.0 before 1.0.5 and 1.1.0 before 1.1.4.

Spring AI's spring-ai-bedrock-converse library is vulnerable to Server-Side Request Forgery (SSRF) due to insufficient validation of user-supplied media URLs in multimodal messages, allowing attackers to trigger HTTP requests to internal or external destinations.

A SpEL injection vulnerability exists in Spring AI's SimpleVectorStore when a user-supplied value is used as a filter expression key, potentially allowing malicious actors to execute arbitrary code in vulnerable applications.

A SQL injection vulnerability (CVE-2026-4910) exists in Shenzhen Ruiming Technology Streamax Crocus bis 1.3.44 via the /RemoteFormat.do endpoint, allowing remote attackers to execute arbitrary SQL commands by manipulating the State argument.

The cpp-httplib library prior to version 0.39.0 forwards stored authentication credentials to arbitrary hosts via HTTP redirects, potentially exposing sensitive information to malicious actors.

A stack-based buffer overflow vulnerability (CVE-2026-4905) exists in Tenda AC5 firmware version 15.03.06.47 allowing remote attackers to execute arbitrary code by manipulating the 'index' argument in a POST request to the /goform/WifiWpsOOB endpoint.

vLLM versions before 0.18.0 are vulnerable to remote code execution due to hardcoded trust of remote code, even when explicitly disabled by the user, allowing attackers to execute arbitrary code via malicious model repositories.

A vulnerability in Incus versions prior to 6.23.0 allows for arbitrary read and write access as root on the host server by exploiting a missing chroot isolation in the pongo2 template engine.

The c2cciutils package is vulnerable to denial of service due to an uncontrolled recursion vulnerability (CWE-674) in the pyasn dependency, specifically versions before 1.1.65.

A path traversal vulnerability exists in Sharp CMS versions prior to 9.20.0 due to improper sanitization of file extensions, potentially allowing attackers to bypass security restrictions and access sensitive files.

Forge is vulnerable to signature forgery in Ed25519 due to a missing check that S < L, allowing non-canonical signatures and potentially bypassing authentication/authorization logic, affecting versions before 1.4.0.

Node-forge's certificate chain verification fails to enforce RFC 5280 basicConstraints, allowing leaf certificates without basicConstraints and keyUsage extensions to act as Certificate Authorities, leading to potential certificate forgery and man-in-the-middle attacks.

A critical vulnerability in OpenClaw versions prior to 2026.3.22 allows for remote code execution and privilege escalation by permitting an operator.pairing approver to approve device requests with broader operator scopes than authorized, potentially leading to unauthorized administrative access.

A vulnerability in the openclaw npm package before version 2026.3.22 allowed mutating internal ACP chat commands without requiring operator.admin scope enforcement, potentially allowing unauthorized control-plane actions.

CVE-2023-7338 is a remote code execution vulnerability affecting Ruckus Unleashed when gateway mode is enabled, allowing authenticated remote attackers to execute arbitrary code by sending specially crafted requests through the web-based management interface.

Ory Polis versions prior to 26.2.0 are vulnerable to DOM-based XSS due to improper handling of the `callbackUrl` parameter, allowing attackers to execute arbitrary JavaScript in a user's browser.

The openclaw npm package before version 2026.3.22 allows unauthorized pre-authentication computation due to improper handling of inbound Nostr DMs, where crypto and dispatch work are performed before enforcing sender and pairing policies.

Netty incorrectly parses quoted strings in HTTP/1.1 chunked transfer encoding extension values, enabling request smuggling attacks by terminating chunk header parsing at \r\n inside quoted strings instead of rejecting the malformed request.

A denial of service vulnerability exists in Netty's HTTP/2 server implementation where an unauthenticated user can exhaust server CPU resources by sending a flood of CONTINUATION frames with zero-byte payloads, bypassing size-based mitigations and leading to service unavailability with minimal bandwidth usage; affected versions include netty-codec-http2 < 4.1.132.Final and netty-codec-http2 versions >= 4.2.0.Alpha1 and < 4.2.10.Final.

OpenBao versions before 2.5.2 lack user confirmation for OIDC direct callback mode, allowing attackers to perform remote phishing and bypass authentication.

A SQL injection vulnerability exists in the ListCourierMessages Admin API of Ory Kratos versions prior to 26.2.0 due to flaws in its pagination implementation, allowing attackers to craft malicious tokens if the pagination secret is known or the default secret is used.

A SQL injection vulnerability exists in Kysely versions prior to 0.28.14 due to insufficient backslash escaping in the `DefaultQueryCompiler.sanitizeStringLiteral()` function, potentially allowing attackers to inject arbitrary SQL when using the MySQL dialect, specifically affecting `CreateIndexBuilder.where()` and `CreateViewBuilder.as()` methods.

Mattermost versions 11.2.x <= 11.2.2, 10.11.x <= 10.11.10, 11.4.x <= 11.4.0, 11.3.x <= 11.3.1 are vulnerable to terminal injection, allowing attackers to manipulate administrator terminals via crafted messages containing ANSI and OSC escape sequences.

Kysely versions 0.28.12 and 0.28.13 are vulnerable to SQL injection due to insufficient escaping of backslashes in the `sanitizeStringLiteral` method, potentially leading to arbitrary SQL execution on MySQL servers.

EVerest versions prior to 2026.02.0 exhibit a data race vulnerability (CVE-2026-26074) where concurrent network requests and physical events can corrupt the event queue, leading to potential denial of service or other undefined behavior.

This rule detects when secrets are accessed via an unusual user agent, user name, and source IP in a Kubernetes cluster, indicating potential credential access attempts by attackers seeking sensitive information.