vLLM is an inference and serving engine for large language models (LLMs). Prior to version 0.14.1, a Server-Side Request Forgery (SSRF) vulnerability exists in the `MediaConnector` class within the vLLM project's multimodal feature set. The load_from_url and load_from_url_async methods obtain and process media from URLs provided by users, using different Python parsing libraries when restricting the target host. These two parsing libraries have different interpretations of backslashes, which allows the host name restriction to be bypassed. This allows an attacker to coerce the vLLM server into making arbitrary requests to internal network resources. This vulnerability is particularly critical in containerized environments like `llm-d`, where a compromised vLLM pod could be used to scan the internal network, interact with other pods, and potentially cause denial of service or access sensitive data. For example, an attacker could make the vLLM pod send malicious requests to an internal `llm-d` management endpoint, leading to system instability by falsely reporting metrics like the KV cache state. Version 0.14.1 contains a patch for the issue.

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.6.5 and prior to 0.8.5, having vLLM integration with mooncake, are vulnerable to remote code execution due to using pickle based serialization over unsecured ZeroMQ sockets. The vulnerable sockets were set to listen on all network interfaces, increasing the likelihood that an attacker is able to reach the vulnerable ZeroMQ sockets to carry out an attack. vLLM instances that do not make use of the mooncake integration are not vulnerable. This issue has been patched in version 0.8.5.

vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.14.0, vLLM loads Hugging Face `auto_map` dynamic modules during model resolution without gating on `trust_remote_code`, allowing attacker-controlled Python code in a model repo/path to execute at server startup. An attacker who can influence the model repo/path (local directory or remote Hugging Face repo) can achieve arbitrary code execution on the vLLM host during model load. This happens before any request handling and does not require API access. Version 0.14.0 fixes the issue.

vLLM is an inference and serving engine for large language models (LLMs). From 0.8.3 to before 0.14.1, when an invalid image is sent to vLLM's multimodal endpoint, PIL throws an error. vLLM returns this error to the client, leaking a heap address. With this leak, we reduce ASLR from 4 billion guesses to ~8 guesses. This vulnerability can be chained a heap overflow with JPEG2000 decoder in OpenCV/FFmpeg to achieve remote code execution. This vulnerability is fixed in 0.14.1.

vllm-project vllm version 0.6.0 contains a vulnerability in the AsyncEngineRPCServer() RPC server entrypoints. The core functionality run_server_loop() calls the function _make_handler_coro(), which directly uses cloudpickle.loads() on received messages without any sanitization. This can result in remote code execution by deserializing malicious pickle data.

vLLM, an inference and serving engine for large language models (LLMs), has an issue in versions 0.6.5 through 0.8.4 that ONLY impacts environments using the `PyNcclPipe` KV cache transfer integration with the V0 engine. No other configurations are affected. vLLM supports the use of the `PyNcclPipe` class to establish a peer-to-peer communication domain for data transmission between distributed nodes. The GPU-side KV-Cache transmission is implemented through the `PyNcclCommunicator` class, while CPU-side control message passing is handled via the `send_obj` and `recv_obj` methods on the CPU side.​ The intention was that this interface should only be exposed to a private network using the IP address specified by the `--kv-ip` CLI parameter. The vLLM documentation covers how this must be limited to a secured network. The default and intentional behavior from PyTorch is that the `TCPStore` interface listens on ALL interfaces, regardless of what IP address is provided. The IP address given was only used as a client-side address to use. vLLM was fixed to use a workaround to force the `TCPStore` instance to bind its socket to a specified private interface. As of version 0.8.5, vLLM limits the `TCPStore` socket to the private interface as configured.

vLLM is an inference and serving engine for large language models (LLMs). From 0.16.0 to before 0.19.0, a server-side request forgery (SSRF) vulnerability in download_bytes_from_url allows any actor who can control batch input JSON to make the vLLM batch runner issue arbitrary HTTP/HTTPS requests from the server, without any URL validation or domain restrictions. This can be used to target internal services (e.g. cloud metadata endpoints or internal HTTP APIs) reachable from the vLLM host. This vulnerability is fixed in 0.19.0.

Azure Monitor Elevation of Privilege Vulnerability

Webtools in Brocade SANnav before version 2.1.1 allows unauthenticated users to make requests to arbitrary hosts due to a misconfiguration; this is commonly referred to as Server-Side Request Forgery (SSRF).

The Fusion Builder WordPress plugin before 3.6.2, used in the Avada theme, does not validate a parameter in its forms which could be used to initiate arbitrary HTTP requests. The data returned is then reflected back in the application's response. This could be used to interact with hosts on the server's local network bypassing firewalls and access control measures.

Server-side request forgery (ssrf) in Azure Compute Gallery allows an unauthorized attacker to elevate privileges over a network.

uppy is vulnerable to Server-Side Request Forgery (SSRF)

Dataease is an open source data analytics and visualization platform. In Dataease versions up to 2.10.12, the patch introduced to mitigate DB2 JDBC deserialization remote code execution attacks only blacklisted the rmi parameter. The ldap parameter in the DB2 JDBC connection string was not filtered, allowing attackers to exploit the DB2 JDBC connection string to trigger server-side request forgery (SSRF). In higher versions of Java, ldap deserialization (autoDeserialize) is disabled by default, preventing remote code execution, but SSRF remains exploitable. Versions up to 2.10.12 are affected. The issue is fixed in version 2.10.13. Updating to 2.10.13 or later is recommended. No known workarounds are documented aside from upgrading.

Server-Side Request Forgery (SSRF) in Pypi calibreweb prior to 0.6.16.

Stirling-PDF is a locally hosted web application that performs various operations on PDF files. Prior to version 1.1.0, when using the /api/v1/convert/html/pdf endpoint to convert HTML to PDF, the backend calls a third-party tool to process it and includes a sanitizer for security sanitization which can be bypassed and result in SSRF. This issue has been patched in version 1.1.0.

Stirling-PDF is a locally hosted web application that performs various operations on PDF files. Prior to version 1.1.0, when using the /api/v1/convert/markdown/pdf endpoint to convert Markdown to PDF, the backend calls a third-party tool to process it and includes a sanitizer for security sanitization which can be bypassed and result in SSRF. This issue has been patched in version 1.1.0.

An improper access control vulnerability in GroupMe allows an a unauthenticated attacker to elevate privileges over a network.

flatCore-CMS version 2.0.8 calls dangerous functions, causing server-side request forgery vulnerabilities.

An SSRF issue was discovered in SquaredUp for SCOM 5.2.1.6654.

Nextcloud Mail is the mail app for Nextcloud, a self-hosted productivity platform. Starting in version 1.13.0 and prior to version 2.2.8 and 3.3.0, an attacker can use an unprotected endpoint in the Mail app to perform a SSRF attack. Nextcloud Mail app versions 2.2.8 and 3.3.0 contain a patch for this issue. As a workaround, disable the mail app.

A vulnerability classified as critical was found in quequnlong shiyi-blog up to 1.2.1. This vulnerability affects unknown code of the file /app/sys/article/optimize. The manipulation of the argument url leads to server-side request forgery. The attack can be initiated remotely. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

Monsta FTP 2.10.1 or below is prone to a server-side request forgery vulnerability due to insufficient restriction of the web fetch functionality. This allows attackers to read arbitrary local files and interact with arbitrary third-party services.

Bitrix24 through 20.0.975 allows SSRF via an intranet IP address in the services/main/ajax.php?action=attachUrlPreview url parameter, if the destination URL hosts an HTML document containing '<meta name="og:image" content="' followed by an intranet URL.

Server-Side Request Forgery (SSRF) vulnerability in ActiveCampaign.This issue affects ActiveCampaign: from n/a through 8.1.14.

WSO2 API Manager 3.0.0 does not properly restrict outbound network access from a Publisher node, opening up the possibility of SSRF to this node's entire intranet.

Volmarg Personal Management System 1.4.64 is vulnerable to SSRF (Server Side Request Forgery) via uploading a SVG file. The server can make unintended HTTP and DNS requests to a server that the attacker controls.

A Server-Side Request Forgery (SSRF) in weixin.php of ChatGPT-wechat-personal commit a0857f6 allows attackers to force the application to make arbitrary requests.

Appspace 6.2.4 allows SSRF via the api/v1/core/proxy/jsonprequest url parameter.

Server Side Template Injection in Gambio 4.9.2.0 allows attackers to run arbitrary code via crafted smarty email template.

Whoogle Search is a self-hosted metasearch engine. In versions prior to 0.8.4, the `element` method in `app/routes.py` does not validate the user-controlled `src_type` and `element_url` variables and passes them to the `send` method which sends a GET request on lines 339-343 in `request.py`, which leads to a server-side request forgery. This issue allows for crafting GET requests to internal and external resources on behalf of the server. For example, this issue would allow for accessing resources on the internal network that the server has access to, even though these resources may not be accessible on the internet. This issue is fixed in version 0.8.4.

GitLab 8.10 and later through 12.9 is vulnerable to an SSRF in a project import note feature.

Whoogle Search is a self-hosted metasearch engine. In versions 0.8.3 and prior, the `window` endpoint does not sanitize user-supplied input from the `location` variable and passes it to the `send` method which sends a `GET` request on lines 339-343 in `request.py,` which leads to a server-side request forgery. This issue allows for crafting GET requests to internal and external resources on behalf of the server. For example, this issue would allow for accessing resources on the internal network that the server has access to, even though these resources may not be accessible on the internet. This issue is fixed in version 0.8.4.

Kyverno, versions 1.16.0 and later, are vulnerable to SSRF due to unrestricted CEL HTTP functions.

GitLab EE 3.0 through 12.8.1 allows SSRF. An internal investigation revealed that a particular deprecated service was creating a server side request forgery risk.

Server-Side Request Forgery (SSRF) vulnerability in WPWebinarSystem WebinarPress wp-webinarsystem allows Server Side Request Forgery.This issue affects WebinarPress: from n/a through <= 1.33.28.

Server-Side Request Forgery (SSRF) vulnerability in Varun Dubey Wbcom Designs - Activity Link Preview For BuddyPress activity-link-preview-for-buddypress allows Server Side Request Forgery.This issue affects Wbcom Designs - Activity Link Preview For BuddyPress: from n/a through <= 1.4.4.

An issue was discovered in ProxyServlet.java in the /proxy servlet in Zimbra Collaboration Suite 8.8 before 8.8.15 Patch 23 and 9.x before 9.0.0 Patch 16. The value of the X-Host header overwrites the value of the Host header in proxied requests. The value of X-Host header is not checked against the whitelist of hosts Zimbra is allowed to proxy to (the zimbraProxyAllowedDomains setting).

Microsoft Exchange Server Remote Code Execution Vulnerability

A server-side request forgery (SSRF) vulnerability was identified in the GitHub Enterprise Server notebook viewer that allowed an attacker to access internal services by exploiting URL parser confusion between the validation layer and the HTTP request library. The hostname validation used a different URL parser than the request library, enabling a crafted URL to pass validation while directing the request to an unintended host. Exploitation required network access to the GitHub Enterprise Server instance. This vulnerability affected all versions of GitHub Enterprise Server prior to 3.21 and was fixed in versions 3.16.18, 3.17.15, 3.18.9, 3.19.6, and 3.20.2. This vulnerability was reported via the GitHub Bug Bounty program.

zrlog v3.1.5 was discovered to contain a Server-Side Request Forgery (SSRF) via the downloadUrl parameter.

monetr is a budgeting application for recurring expenses. Prior to version 1.12.5, a server-side request forgery (SSRF) vulnerability in monetr's Lunch Flow integration allowed any authenticated user on a self-hosted instance to cause the monetr server to issue HTTP GET requests to arbitrary URLs supplied by the caller, with the response body from non-200 upstream responses reflected back in the API error message. This issue has been patched in version 1.12.5.

"HCL Digital Experience is susceptible to Server Side Request Forgery."

PhpSpreadsheet is a library for reading and writing spreadsheet files. In versions 1.30.2 and earlier, 2.0.0 through 2.1.14, 2.2.0 through 2.4.3, 3.3.0 through 3.10.3, and 4.0.0 through 5.5.0, when the filename argument to IOFactory::load() is user-controlled, an attacker can supply a PHP stream wrapper path (such as phar://, ftp://, or ssh2.sftp://) that passes the is_file() check in File::assertFile(). The phar:// wrapper triggers deserialization of the PHAR metadata, which can lead to remote code execution if a suitable gadget chain is available in the application. The ftp:// and ssh2.sftp:// wrappers can be used for server-side request forgery. This issue has been fixed in versions 1.30.3, 2.1.15, 2.4.4, 3.10.4, and 5.6.0.

Server-side request forgery (ssrf) in Azure Databricks allows an unauthorized attacker to elevate privileges over a network.

PraisonAI is a multi-agent teams system. Prior to version 1.6.32, the URL checking logic in PraisonAI has a logical flaw that could be bypassed by attackers, leading to SSRF attacks. This issue has been patched in version 1.6.32.

Hyland Alfresco Transformation Service allows unauthenticated attackers to achieve remote code execution through the argument injection vulnerability, which exists in the document processing functionality.

Altec DocLink (now maintained by Beyond Limits Inc.) version 4.0.336.0 exposes insecure .NET Remoting endpoints over TCP and HTTP/SOAP via Altec.RDCHostService.exe using the ObjectURI "doclinkServer.soap". The service does not require authentication and is vulnerable to unsafe object unmarshalling, allowing remote attackers to read arbitrary files from the underlying system by specifying local file paths. Additionally, attackers can coerce SMB authentication via UNC paths and write arbitrary files to server locations. Because writable paths may be web-accessible under IIS, this can result in unauthenticated remote code execution or denial of service through file overwrite.

DB4Web server, when configured to use verbose debug messages, allows remote attackers to use DB4Web as a proxy and attempt TCP connections to other systems (port scan) via a request for a URL that specifies the target IP address and port, which produces a connection status in the resulting error message.

CrewAI contains a server-side request forgery vulnerability that enables content acquisition from internal and cloud services, facilitated by the RAG search tools not properly validating URLs provided at runtime.

An issue pertaining to CWE-918: Server-Side Request Forgery was discovered in oslabs-beta ThermaKube master.