The Autobahn Python library's permessage-brotli extension inflates entire frames before enforcing the maxMessagePayloadSize limit, allowing a small compressed frame to cause excessive memory allocation. This can lead to a denial-of-service (DoS) condition by exhausting server memory. The issue affects versions from 26.7.1 up to but not including 26.10.1 and was fixed in version 26.10.1.
A path traversal vulnerability (CWE-22) exists in langchain-ai's langgraph Python SDK and langgraph versions prior to 0.4.6 and 1.2.14 respectively. The issue arises when RemoteGraph v3 streaming uses affected versions of langgraph-sdk, where caller-controlled assistant_id and thread_id values are interpolated into authenticated GET request paths without proper encoding. This allows relative path segments to be normalized into other readable paths on the server while retaining authentication headers. The vulnerability is fixed in langgraph-sdk 0.4.6 and langgraph 1.2.14.
Happy DOM versions prior to 20.14.6 have a vulnerability in their CookieContainer where domain cookie attributes are not properly validated. This flaw allows acceptance of cookies with domains unrelated to the origin, including public suffixes, and incorrect matching of unrelated hostnames sharing suffixes. This can lead to cookie injection or session fixation in server-side rendering or prerendering contexts when attacker-controlled origins are contacted before victim origins. The issue is resolved in version 20.14.6.
CVE-2026-118660 is a server-side request forgery (SSRF) vulnerability in the Happy DOM JavaScript library prior to version 21.0.0. The library loads iframe src and stylesheet href URLs by default when parsing untrusted HTML on the server, which can be exploited by attackers controlling the HTML to make arbitrary internal or cloud endpoint requests. This may lead to exposure of confidential responses or credentials. The vulnerability is fixed in Happy DOM version 21.0.0.
CVE-2026-118659 is a Server-Side Request Forgery (SSRF) vulnerability in the Happy DOM server-renderer package prior to version 21.0.2. The vulnerability allows unauthenticated HTTP/2 requests with attacker-controlled paths to be resolved against the server's target origin without verifying if the resolved origin remains trusted. This can lead to the server fetching arbitrary HTTP or HTTPS endpoints, including internal services, and relaying the response back to the attacker. The issue is fixed in Happy DOM version 21.0.2.
CVE-2026-93522 is a heap-based buffer overflow vulnerability in X.Org X Server and XWayland caused by mismatched size usage in the CopyArea CPU-to-FBO path. It affects X.Org X Server versions before 21.1.25 and XWayland versions before 24.1.14. Authenticated X clients could exploit this flaw to crash the X server or potentially execute arbitrary code within the server context.
MLflow tracking servers running without authentication or with basic authentication are vulnerable to an arbitrary file read vulnerability via crafted model versions tagged as prompts with schemeless local paths. This allows remote attackers to read arbitrary files on the server filesystem with the privileges of the MLflow server process. The vulnerability affects MLflow versions 3.10.0 through 3.16.1 and is fixed in version 3.17.0.
MLflow versions from 3.12.0rc0 up to but not including 3.17.0 are vulnerable to unauthenticated remote code execution via a flaw in third-party scorer deserialization. The vulnerability arises because the server imports and instantiates Python classes from request-controlled module paths under allowed package prefixes, which can be exploited by uploading malicious Python modules into writable artifact directories within these allowed packages. This leads to arbitrary code execution with the privileges of the MLflow server process, which by default runs as root in the official image. The issue is critical and has a CVSS score of 9.8. A fix is available in MLflow 3.17.0.
CVE-2026-116686 is a server-side request forgery (SSRF) vulnerability in the Model Context Protocol Python SDK versions from 1.10.0 up to but not including 1.30.0, and from 2.0.0 up to but not including 2.2.0. The vulnerability arises because certain client calls validate server-declared outputSchema documents using jsonschema's default resolver, which can cause unresolved $ref values to fetch HTTP, HTTPS, or file URLs from the client's network or filesystem. This allows a malicious server to trigger blind requests or local file reads and stall sessions due to synchronous resolution without timeout. The issue is fixed in versions 1.30.0 and 2.2.0.
CVE-2026-116684 is an authorization vulnerability in the Model Context Protocol Python SDK versions 1.10.0 through before 1.30.0 and 2.0.0 through before 2.2.0. The flaw occurs when server-side bearer authentication accepts a valid token issued for a different service due to lack of independent audience (aud) validation by the TokenVerifier. This allows an attacker holding a token for one resource to authenticate to a different MCP server and gain unauthorized access. The issue is fixed in versions 1.30.0 and 2.2.0.
Authlib versions 1.6.0 through before 1.9.0 contain a vulnerability in the RFC 9101 JAR extension where the attacker-controlled request_uri parameter is passed to a fetch hook without validation at a pre-authentication endpoint. This allows an unauthenticated Server-Side Request Forgery (SSRF) primitive because the documented example implementation performs an unrestricted HTTP GET on the request_uri. The vulnerability arises because the existing allowlist controls for request_uris are ignored, and the extension does not implement required RFC 9101 controls such as validating the URI location, media type, or timeouts. A fix is available in version 1.9.0.
Authlib versions prior to 1.9.0 have a vulnerability in the default JWT bearer grant implementation where the audience ('aud') claim in JWT assertions is not properly validated. The default get_audiences() method returns an empty list, causing verify_claims() to only check for the presence of an 'aud' claim rather than verifying it matches the authorization server's token endpoint. This allows an attacker with a valid JWT for any audience to obtain an access token from a different authorization server sharing the same signing key.
Kali-GPT contains an OS command injection vulnerability in the execute_tool() function within kali_gpt/tools/mcp_server.py. This flaw allows MCP clients to execute arbitrary shell commands by injecting shell metacharacters into parameters such as target, ports, domain, wordlist, or url. The vulnerability arises because these parameters are passed unsanitized to subprocess.run with shell=True, running with the MCP server user's privileges.