| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| The ToASCII and ToUnicode functions incorrectly accept Punycode-encoded labels that decode to an ASCII-only label. For example, ToUnicode("xn--example-.com") incorrectly returns the name "example.com" rather than an error. This behavior can lead to privilege escalation in programs using the idna package. For example, a program which performs privilege checks on the ASCII hostname may reject "example.com" but permit "xn--example-.com". If that program subsequently converts the ASCII hostname to Unicode, it will inadvertently permits access to the Unicode name "example.com". |
| Well-crafted inputs reaching ParseAddress, ParseAddressList, and ParseDate were able to trigger excessive CPU exhaustion and memory allocations. |
| Jupyter Server is the backend for Jupyter web applications. In versions 2.17.0 and earlier, a path traversal vulnerability in the REST API allows an authenticated user to escape the configured root_dir and access sibling directories whose names begin with the same prefix as the root_dir. For example, with a root_dir named "test", the API permits access to a sibling directory named "testtest" through a crafted request to the /api/contents endpoint using encoded path components. An attacker can read, write, and delete files in affected sibling directories. Multi-tenant deployments using predictable naming schemes are particularly at risk, as a user with a directory named "user1" could access directories for user10 through user19 and beyond. A user who can choose a single-character folder name could gain access to a significant number of sibling directories.
Version 2.18.0 contains a fix. As a workaround, ensure folder names do not share a common prefix with any sibling directory. |
| AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to version 3.14.0, using ``CookieJar.load()`` with untrusted input may allow arbitrary code execution. Most applications using this function will be doing so with the user's own data, so this is unlikely to affect many applications. Version 3.14.0 patches the issue. If an application does allow attacker controlled files to be loaded, a workaround on older releases would be to sanitize the files before loading. |
| When processing HTTP/2 SETTINGS frames, transport will enter an infinite loop of writing CONTINUATION frames if it receives a SETTINGS_MAX_FRAME_SIZE with a value of 0. |
| When using LookupCNAME with the cgo DNS resolver, a very long CNAME response can trigger a double-free of C memory and a crash. |
| In versions 3.0.0a1 through 3.2.0 of Mistune, there is a ReDoS (Regular Expression Denial of Service) vulnerability in `LINK_TITLE_RE` that allows an attacker who can supply Markdown for parsing to cause denial of service. The regular expression used for parsing link titles contains overlapping alternatives that can trigger catastrophic backtracking. In both the double-quoted and single-quoted branches, a backslash followed by punctuation can be matched either as an escaped punctuation sequence or as two ordinary characters, creating an ambiguous pattern inside a repeated group. If an attacker supplies Markdown containing repeated ! sequences with no closing quote, the regex engine explores an exponential number of backtracking paths. This is reachable through normal Markdown parsing of inline links and block link reference definitions. A small crafted input can therefore cause significant CPU consumption and make applications using Mistune unresponsive. |
| A vulnerability in the `_create_model_version()` handler of `mlflow/server/handlers.py` in mlflow/mlflow versions 3.9.0 and earlier allows an unauthenticated remote attacker to read arbitrary files from the server's filesystem. The issue arises when a `CreateModelVersion` request includes the tag `mlflow.prompt.is_prompt`, which bypasses source path validation. This enables an attacker to store an arbitrary local filesystem path as the model version source. The `get_model_version_artifact_handler()` function later uses this source to serve files without verifying the model version's prompt status, leading to a complete confidentiality compromise. This issue is fixed in version 3.10.0. |
| fast-uri versions 2.3.1 through 3.1.2 and 4.0.0 fail to canonicalize Unicode (IDN) hostnames for HTTP-family URLs. The IDN conversion path calls a helper that does not exist on the global URL constructor, silently leaving the host in its original Unicode form while normalize() and equal() still return values that differ from a WHATWG-compatible URL parser. Applications that use fast-uri to enforce host-based policy (denylists, loopback filtering, redirect validation, outbound proxy routing) before passing the same URL to Node's URL or fetch can be bypassed when the two implementations resolve the same input to different hosts. Patches: upgrade to fast-uri 3.1.3 for the 3.x line or 4.0.1 for the 4.x line. Workarounds: enforce host policy using the same URL parser used for the actual request, or reject non-ASCII hosts before policy checks. |
| Impact:
The undici WebSocket client enforces maxPayloadSize on the cumulative byte count of fragments in a message but does not enforce a limit on the number of fragments. A malicious WebSocket server can stream many small or empty continuation frames that each pass per-frame and cumulative-size validation, collectively causing unbounded memory growth in the client process. The result is memory exhaustion and a denial of service.
Affected applications are those using the undici WebSocket client (new WebSocket(...)) or the WebSocketStream API that can be induced to connect to an attacker-controlled or compromised WebSocket endpoint.
All releases starting at undici 6.17.0 are affected.
Patches: Upgrade to undici >= 6.26.0, >= 7.28.0, or >= 8.5.0. Workarounds:
No workaround is available. The fix must be applied through an upgrade. |
| form-data is a library for creating readable multipart/form-data streams. In versions through 4.0.5, the `field` argument to `FormData#append` and the `filename` option are concatenated verbatim into the `Content-Disposition` header without escaping carriage return (CR), line feed (LF), or double-quote (") characters. An application that passes attacker-controlled data as a field name or filename (for example, an API gateway that turns JSON object keys into multipart field names) allows the attacker to terminate the header line and inject additional headers, or to smuggle entire additional multipart parts, into the request the application forwards to a backend. This can let the attacker add or override form fields (e.g. set `is_admin=true`) seen by the downstream parser. This is an instance of CWE-93 (CRLF injection). The fix escapes CR, LF, and `"` as `%0D`, `%0A`, and `%22` in field names and filenames, matching the serialization browsers use per the WHATWG HTML multipart/form-data encoding algorithm. Exploitation requires the consuming application to use untrusted input as a field name or filename; applications that use only fixed/trusted field names are not affected. Fixed in 2.5.6, 3.0.5, and 4.0.6. |
| Nokogiri before 1.13.2 (CRuby, when using packaged libraries) ships vendored libxml2 2.9.12 and libxslt 1.1.34, which are affected by two upstream CVEs. Via CVE-2021-30560 in libxslt, an application transforming XML with untrusted XSL stylesheets is vulnerable to a denial-of-service attack. Via CVE-2022-23308 in libxml2, an application parsing an untrusted document with parse option DTDVALID set to true and NOENT set to false may be vulnerable to denial of service, memory disclosure, or code execution. Nokogiri 1.13.2 upgrades vendored libxml2 to 2.9.13 and libxslt to 1.1.35. |
| In Bouncy Castle for Java before 1.85, MLS wire decoder allocates attacker-declared opaque length before bounds check. |
| In versions of Spring Security's OAuth2 Authorization Server module 7.0.0 through 7.0.4, when Dynamic Client Registration is explicitly enabled, the registration endpoint performs insufficient validation of certain client metadata fields supplied by the registering client. An attacker who possesses a valid Initial Access Token can register a malicious client with crafted metadata, which, depending on server configuration and how the metadata is later rendered or used, may result in Stored Cross-Site Scripting (XSS), Privilege Escalation, or Server-Side Request Forgery (SSRF). |
| An Improper Authorization vulnerability affecting 3DPassport in 3DSwymer from Release 3DEXPERIENCE R2023x through Release 3DEXPERIENCE R2026x could allow an attacker to gain access to some user accounts. |
| The proxy middleware in mcp-use's inspector forwards requests to a destination the caller names. mountMcpProxy in libraries/typescript/packages/inspector/src/server/proxy/mcp-proxy.ts read the target from the X-Target-URL header or the __mcp_target parameter and proxied to it without inspecting the host, so loopback, link-local and private addresses were all accepted, as were names that resolve to them, and the validation was not reapplied to a redirect the destination returned. A caller could therefore make the server issue requests to addresses reachable only from the host it runs on and read the responses. The current code calls isSafeProxyTarget, which checks the resolved address against private, loopback and link-local ranges before proxying and bounds the number of redirects followed. |
| mcp-go accepted requests on its HTTP transports without checking the Host header. StreamableHTTPServer.ServeHTTP in server/streamable_http.go and SSEServer.ServeHTTP in server/sse.go served any request arriving over a loopback connection regardless of the host it named, and the SSE transport's cross-origin default allowed any origin. A page in a browser could therefore point a name it controlled at the loopback address and reach a server listening there, invoking tools and reading resources that the server exposed on the assumption that only local software could connect. No release before 0.56.0 validated the header on either transport; 0.56.0 adds server/http_localhost.go, which rejects a loopback-bound request carrying a host that is not a loopback name, and wires it into both transports. |
| The mcp-router CLI served its MCP aggregator on every interface and enforced authentication only when the operator asked for it. The serve command in apps/cli/src/commands/serve.ts defaulted its host to the all-interfaces address on a fixed port, and required a token only when the corresponding flag was supplied, so a default invocation exposed the aggregator, and every MCP server it fronted, to anyone able to reach the port. Release 0.6.3 defaults the host to the loopback address and refuses to start without a token whenever the host it is given is not a loopback address; no earlier release carries either check. |
| ToolUniverse ran caller-supplied Python inside a sandbox that could be escaped, on a server that required no authentication. The executor behind the python_code_executor tool, in python_executor_tool.py, inspected the submitted source for a denied list of attribute names and calls but left the attribute-lookup builtins available and did not stop a dunder attribute reached through a string lookup or through a module already permitted, so a caller could walk from a literal's class to its base and enumerate subclasses to obtain a reference to the process and subprocess modules. A per-call argument also let the caller widen the import allow-list before the inspection ran. The HTTP and MCP servers in http_api_server.py and smcp_server.py bound to every interface with debugging enabled and no authentication, so any caller able to reach the port executed code as the server process. Version 1.3.0 adds bearer-token authentication, defaults the bind address to loopback, and hardens the attribute checks. |
| The execute_ruby tool is documented as a read-only Ruby sandbox and is enforced by a pattern denylist together with replacements for the process-spawning methods on Kernel. The pseudo-terminal library's spawn entry points are neither in the denylist nor replaced, so a normal tool call could reach them and start a shell, executing commands as the account running the server and outside the guarded methods. The denylist was introduced with the tool in 1.4.0 and never covered those entry points through 1.6.0. Version 1.6.1 restricts the requires the sandbox permits to a data-only list and blocks dynamic dispatch to execution entry points; 2.0.0 removes the tool. |
