CVE Vulnerability Catalog

Vouch Proxy prior to 0.48.0 parses the total part count from an attacker-controlled multipart cookie name and passes it to make([]string, numParts) without checking that the value is positive or reasonably bounded. A cookie name such as VouchCookie_1of10000000000 causes an attempted allocation of roughly 160 GB and a fatal Go out-of-memory condition.

MCP Memory Keeper is an MCP server for persistent context management in AI coding assistants. Prior to 0.13.0, context_import in src/index.ts passes the caller-controlled filePath directly to fs.readFileSync without restricting the path to an export directory. An MCP client, including an LLM agent induced to call the tool, can use ../ traversal or an absolute path to target any file readable by the server process. A valid JSON file is parsed and imported into the caller's session, allowing its full contents to be retrieved through context_get or context_export, while JSON.parse errors for non-JSON files can return leading file bytes in a SyntaxError message. The two disclosure modes can expose other exported sessions, JSON credentials or service-account files, environment files, and portions of SSH keys or other local files. This issue is fixed in version 0.13.0.

plone.app.textfield provides a zope.schema-style field type called RichText for storing a value with a related MIME type. Prior to 2.0.2, 3.0.2, and 4.0.1, depending on the release line, RichTextValue.output returns an unsanitized stored RichText value when mimeType equals outputMimeType, including values that claim the text/x-html-safe output type. This equality shortcut bypasses the safe_html transform even though the transform itself correctly removes event-handler attributes and unsafe URI schemes. Equal types can result from a RichText field configured with the same mimeType and outputMimeType or from REST API input that supplies text/x-html-safe as its content type. The raw stored value is then emitted through tal:content=structure without escaping, allowing a user who can set a RichText field to store JavaScript that executes in a viewer's browser. This issue is fixed in versions 2.0.2, 3.0.2, and 4.0.1.

MCP Context Forge is an AI gateway, registry, and proxy for MCP, A2A, REST, and gRPC APIs. Prior to 1.0.2, the python_sandbox_server in mcp-servers/python/python_sandbox_server/src/python_sandbox_server/server_fastmcp.py exposes raw getattr through safe_builtins, omits a required _getattr_ guard, and relies on validate_code checks for literal dangerous dunder strings. An attacker can construct dunder names at runtime, traverse the Python class hierarchy, reach subprocess.Popen, and execute OS commands with the server process privileges through the execute_code MCP tool. The HTTP/SSE transport can expose this tool without authentication, while stdio-only deployments have reduced network reachability. The issue affects the python_sandbox_server subproject and does not directly affect the core Context Forge gateway or proxy components. This issue is fixed in version 1.0.2.

Fabric CA is a Certificate Authority for Hyperledger Fabric. Prior to 1.5.21, when fabric-ca is configured with an LDAP backend, Client.GetUser in lib/server/ldap/client.go inserts the username from HTTP Basic authentication into the LDAP uid search UserFilter without escaping LDAP metacharacters. An unauthenticated attacker with network access to the CA enrollment endpoint can alter the LDAP search before password validation and potentially steer authentication attempts toward a victim account. Deployments that do not use an LDAP backend are unaffected. This issue is fixed in version 1.5.21.

motionEye (mEye) is an online interface for a piece of software called "motion," which is a video surveillance program with motion detection. Prior to 0.44.0, motionEye accepts the client-controlled meye_username and meye_password_hash cookies as authentication material without server-side session validation. An unauthenticated attacker who knows a target username and corresponding hash can set the cookies manually or cause them to be loaded by submitting blank credentials through the switch-user authentication flow, after which the server authenticates the attacker as that user. The administrator username and password-hash value are stored in /etc/motioneye/motion.conf, which is globally readable by default, allowing a local shell user to obtain reusable administrator credential material. Successful impersonation can enable account lockout, password changes and persistence, data enumeration, data destruction, and data exfiltration. This issue is fixed in version 0.44.0.

Inspektor Gadget from 0.28.0 until 0.53.1 has a flaw in the USDT note parser (pkg/uprobetracer/usdt.go). When a custom gadget with a SEC("usdt/...") eBPF section attaches to a crafted ELF binary, an unprivileged container can trigger an out-of-bounds panic or consume gigabytes of memory, crashing or exhausting the privileged Inspektor Gadget process.

A High severity Denial of Service (DoS) vulnerability in Confluence Data Center versions 8.9.0, 9.0.1, 9.1.0, 9.2.0, 9.3.1, 9.4.0, 9.5.1, 10.0.2, 10.1.0, and 10.2.0. An authenticated attacker can cause a resource to be unavailable for its intended users by temporarily or indefinitely disrupting services of a host connected to a network.

A High severity Improper Authorization vulnerability in Jira Service Management Data Center version 11.3.0. An authenticated attacker can gain unintended access, potentially leading to exposure of resources or functionality, and possibly enabling sensitive information disclosure or arbitrary code execution.

A High severity Improper Authorization vulnerability in Confluence Data Center versions 7.4.0, 7.13.0, 8.5.0, 8.9.0, 9.0.1, 9.1.0, 9.2.0, 9.3.1, 9.4.0, 9.5.1, 10.0.2, 10.1.0, and 10.2.0. An authenticated attacker can gain unintended access, potentially leading to exposure of resources or functionality, and possibly enabling sensitive information disclosure or arbitrary code execution.

A Remote Code Execution vulnerability in Koha results from the lack of proper validation of a user-supplied string before passing it to the eval function in the web service listening on TCP port 8081 by default. An authenticated attacker can exploit this vulnerability to execute code in the context of the service account.

Concrete CMS 9 before 9.5.3 and Concrete 8.5.21 contain a stored XSS vulnerability in the Feature, Feature Link, Hero Image, and Image blocks. The external link URL was insufficiently validated and rendered without output escaping, allowing a user with page-editing permissions to inject arbitrary JavaScript.

Concrete CMS 9 (9.0.0 through 9.5.2) does not perform an authorization check on the user selector autocomplete endpoint (/ccm/system/user/autocomplete). The endpoint only validates a CSRF-style token that is not bound to the caller's identity or permissions, and the token is issued to anonymous visitors.

Yonyou U8 CRM contains an unauthenticated SQL injection vulnerability in the fillbacksettingedit.php configuration endpoint where the DontCheckLogin=1 parameter bypasses authentication and the id parameter is incorporated into SQL queries without sanitization. Attackers can exploit this flaw to execute arbitrary SQL commands and, on Microsoft SQL Server deployments with xp_cmdshell enabled, write backdoor files and execute arbitrary operating system commands. Exploitation evidence was first observed by the Shadowserver Foundation on 2025-02-13.

Yonyou U8 Cloud contains an unauthenticated Java deserialization vulnerability in the nc.impl.pub.filesystem.FileManageServlet component that allows remote unauthenticated attackers to execute arbitrary OS commands by sending a serialized payload via POST request. Attackers can exploit the doAction method, which passes raw HTTP request body data directly to ObjectInputStream.readObject() without filtering, to achieve remote code execution. Exploitation evidence was first observed by the Shadowserver Foundation on 2025-02-13.

Tornado before 6.5.7 contains a credential leak vulnerability in CurlAsyncHTTPClient where pycurl handles are reused across requests without proper state clearing. Attackers can obtain sensitive credentials by issuing requests through the same client instance, allowing TLS certificates or proxy authentication to persist across unintended requests.

Tornado before 6.5.8 contains an incomplete fix for cookie attribute injection that allows attackers to inject arbitrary cookie attributes by passing capitalized or legacy keyword arguments to set_cookie. Attackers can embed semicolon-delimited data in capitalized parameters like Domain, Path, or SameSite to bypass validation and modify cookie security attributes.

Tornado before 6.5.8 contains a memory amplification vulnerability in parse_multipart_form_data that splits multipart data before validating the max_parts limit. Attackers can send crafted multipart requests with many parts to create large transient lists, exhausting server memory and causing denial of service.

atomic-agents-stack before 1.1.0 contains a path traversal vulnerability in the dashboard HTTP server that allows remote attackers to read arbitrary files by supplying directory traversal sequences in request paths. Attackers can bypass path containment checks by including '../' segments in requests to the DashboardHandler.do_GET endpoint to access files outside the intended agents_root directory.

atomic-agents-stack before 1.1.0 accepts cleartext HTTP schemes in the HTTP MCP server-registry backend factory, allowing network man-in-the-middle attackers to rewrite catalog responses. Attackers can inject arbitrary command and argument values that are spawned as local subprocesses by MCPClientPool to achieve code execution on the agent host.