CVEs of 2025.
Every CVE published in 2025 that we hold, ranked by exploit odds.
-
CVE-2025-0851
A path traversal issue in ZipUtils.unzip and TarUtils.untar in Deep Java Library (DJL) on all platforms allows a bad actor to write files to arbitrary locations.
-
CVE-2025-2828
A Server-Side Request Forgery (SSRF) vulnerability exists in the RequestsToolkit component of the langchain-community package (specifically, langchain_community.agent_toolkits.openapi.toolkit.RequestsToolkit) in langchain-ai/langchain version 0.0.27. This vulnerability occurs because the toolkit does not enforce restrictions on requests to remote internet addresses, allowing it to also access local addresses. As a result, an attacker could exploit this flaw to perform port scans, access local services, retrieve instance metadata from cloud environments (e.g., Azure, AWS), and interact with servers on the local network. This issue has been fixed in version 0.0.28.
-
CVE-2025-12490
Netgate pfSense CE Suricata Path Traversal Remote Code Execution Vulnerability. This vulnerability allows remote attackers to create arbitrary files on affected installations of Netgate pfSense. Authentication is required to exploit this vulnerability. The specific flaw exists within the Suricata package. The issue results from the lack of proper validation of a user-supplied path prior to using it in file operations. An attacker can leverage this vulnerability to create files in the context of root. Was ZDI-CAN-28085.
-
CVE-2025-54381
BentoML is a Python library for building online serving systems optimized for AI apps and model inference. In versions 1.4.0 until 1.4.19, the file upload processing system contains an SSRF vulnerability that allows unauthenticated remote attackers to force the server to make arbitrary HTTP requests. The vulnerability stems from the multipart form data and JSON request handlers, which automatically download files from user-provided URLs without validating whether those URLs point to internal network addresses, cloud metadata endpoints, or other restricted resources. The documentation explicitly promotes this URL-based file upload feature, making it an intended design that exposes all deployed services to SSRF attacks by default. Version 1.4.19 contains a patch for the issue.
-
CVE-2024-12905
An Improper Link Resolution Before File Access ("Link Following") and Improper Limitation of a Pathname to a Restricted Directory ("Path Traversal"). This vulnerability occurs when extracting a maliciously crafted tar file, which can result in unauthorized file writes or overwrites outside the intended extraction directory. The issue is associated with index.js in the tar-fs package. This issue affects tar-fs: from 0.0.0 before 1.16.4, from 2.0.0 before 2.1.2, from 3.0.0 before 3.0.8.
-
CVE-2025-30005
Xorcom CompletePBX is vulnerable to a path traversal via the Diagnostics reporting module, which will allow reading of arbitrary files and additionally delete any retrieved file in place of the expected report. This issue affects CompletePBX: all versions up to and prior to 5.2.35
-
CVE-2025-47273
setuptools is a package that allows users to download, build, install, upgrade, and uninstall Python packages. A path traversal vulnerability in `PackageIndex` is present in setuptools prior to version 78.1.1. An attacker would be allowed to write files to arbitrary locations on the filesystem with the permissions of the process running the Python code, which could escalate to remote code execution depending on the context. Version 78.1.1 fixes the issue.
-
CVE-2025-4138
Allows the extraction filter to be ignored, allowing symlink targets to point outside the destination directory, and the modification of some file metadata. You are affected by this vulnerability if using the tarfile module to extract untrusted tar archives using TarFile.extractall() or TarFile.extract() using the filter= parameter with a value of "data" or "tar". See the tarfile extraction filters documentation https://docs.python.org/3/library/tarfile.html#tarfile-extraction-filter for more information. Note that for Python 3.14 or later the default value of filter= changed from "no filtering" to `"data", so if you are relying on this new default behavior then your usage is also affected. Note that none of these vulnerabilities significantly affect the installation of source distributions which are tar archives as source distributions already allow arbitrary code execution during the build process. However when evaluating source distributions it's important to avoid installing source distributions with suspicious links.
-
CVE-2025-4517
Allows arbitrary filesystem writes outside the extraction directory during extraction with filter="data". You are affected by this vulnerability if using the tarfile module to extract untrusted tar archives using TarFile.extractall() or TarFile.extract() using the filter= parameter with a value of "data" or "tar". See the tarfile extraction filters documentation https://docs.python.org/3/library/tarfile.html#tarfile-extraction-filter for more information. Note that for Python 3.14 or later the default value of filter= changed from "no filtering" to `"data", so if you are relying on this new default behavior then your usage is also affected. Note that none of these vulnerabilities significantly affect the installation of source distributions which are tar archives as source distributions already allow arbitrary code execution during the build process. However when evaluating source distributions it's important to avoid installing source distributions with suspicious links.
-
CVE-2025-32799
Conda-build contains commands and tools to build conda packages. Prior to version 25.4.0, the conda-build processing logic is vulnerable to path traversal (Tarslip) attacks due to improper sanitization of tar entry paths. Attackers can craft tar archives containing entries with directory traversal sequences to write files outside the intended extraction directory. This could lead to arbitrary file overwrites, privilege escalation, or code execution if sensitive locations are targeted. This issue has been patched in version 25.4.0.
-
CVE-2025-54802
pyLoad is the free and open-source Download Manager written in pure Python. In versions 0.5.0b3.dev89 and below, there is an opportunity for path traversal in pyLoad-ng CNL Blueprint via package parameter, allowing Arbitrary File Write which leads to Remote Code Execution (RCE). The addcrypted endpoint in pyload-ng suffers from an unsafe path construction vulnerability, allowing unauthenticated attackers to write arbitrary files outside the designated storage directory. This can be abused to overwrite critical system files, including cron jobs and systemd services, leading to privilege escalation and remote code execution as root. This issue is fixed in version 0.5.0b3.dev90.
-
CVE-2025-26646
External control of file name or path in .NET, Visual Studio, and Build Tools for Visual Studio allows an authorized attacker to perform spoofing over a network.
-
CVE-2025-23304
NVIDIA NeMo library for all platforms contains a vulnerability in the model loading component, where an attacker could cause code injection by loading .nemo files with maliciously crafted metadata. A successful exploit of this vulnerability may lead to remote code execution and data tampering.
-
CVE-2025-6445
ServiceStack FindType Directory Traversal Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of ServiceStack. Interaction with this library is required to exploit this vulnerability but attack vectors may vary depending on the implementation. The specific flaw exists within the implementation of the FindType method. The issue results from the lack of proper validation of a user-supplied path prior to using it in file operations. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-25837.
-
CVE-2025-6087
A Server-Side Request Forgery (SSRF) vulnerability was identified in the @opennextjs/cloudflare package. The vulnerability stems from an unimplemented feature in the Cloudflare adapter for Open Next, which allowed unauthenticated users to proxy arbitrary remote content via the /_next/image endpoint. This issue allowed attackers to load remote resources from arbitrary hosts under the victim site’s domain for any site deployed using the Cloudflare adapter for Open Next. For example: https://victim-site.com/_next/image?url=https://attacker.com In this example, attacker-controlled content from attacker.com is served through the victim site’s domain (victim-site.com), violating the same-origin policy and potentially misleading users or other services. Impact: * SSRF via unrestricted remote URL loading * Arbitrary remote content loading * Potential internal service exposure or phishing risks through domain abuse Mitigation: The following mitigations have been put in place: * Server side updates to Cloudflare’s platform to restrict the content loaded via the /_next/image endpoint to images. The update automatically mitigates the issue for all existing and any future sites deployed to Cloudflare using the affected version of the Cloudflare adapter for Open Next * Root cause fix https://github.com/opennextjs/opennextjs-cloudflare/pull/727 to the Cloudflare adapter for Open Next. The patched version of the adapter is found here @opennextjs/[email protected] https://www.npmjs.com/package/@opennextjs/cloudflare/v/1.3.0 * Package dependency update https://github.com/cloudflare/workers-sdk/pull/9608 to create-cloudflare (c3) to use the fixed version of the Cloudflare adapter for Open Next. The patched version of create-cloudflare is found here: [email protected] https://www.npmjs.com/package/create-cloudflare/v/2.49.3 In addition to the automatic mitigation deployed on Cloudflare’s platform, we encourage affected users to upgrade to @opennext/cloudflare v1.3.0 and use the remotePatterns https://nextjs.org/docs/pages/api-reference/components/image#remotepatterns filter in Next config https://nextjs.org/docs/pages/api-reference/components/image#remotepatterns if they need to allow-list external urls with images assets.
-
CVE-2025-4330
Allows the extraction filter to be ignored, allowing symlink targets to point outside the destination directory, and the modification of some file metadata. You are affected by this vulnerability if using the tarfile module to extract untrusted tar archives using TarFile.extractall() or TarFile.extract() using the filter= parameter with a value of "data" or "tar". See the tarfile extraction filters documentation https://docs.python.org/3/library/tarfile.html#tarfile-extraction-filter for more information. Note that for Python 3.14 or later the default value of filter= changed from "no filtering" to `"data", so if you are relying on this new default behavior then your usage is also affected. Note that none of these vulnerabilities significantly affect the installation of source distributions which are tar archives as source distributions already allow arbitrary code execution during the build process. However when evaluating source distributions it's important to avoid installing source distributions with suspicious links.
-
CVE-2025-61923
PrestaShop Checkout is the PrestaShop official payment module in partnership with PayPal. In versions prior to 4.4.1 and 5.0.5, the backoffice is missing validation on input resulting in a directory traversal and arbitrary file disclosure. The vulnerability is fixed in versions 4.4.1 and 5.0.5. No known workarounds exist.
-
CVE-2025-48370
auth-js is an isomorphic Javascript library for Supabase Auth. Prior to version 2.70.0, the library functions getUserById, deleteUser, updateUserById, listFactors and deleteFactor did not require the user supplied values to be valid UUIDs. This could lead to a URL path traversal, resulting in the wrong API function being called. Implementations that follow security best practice and validate user controlled inputs, such as the userId are not affected by this. This issue has been patched in version 2.70.0.
-
CVE-2024-12718
Allows modifying some file metadata (e.g. last modified) with filter="data" or file permissions (chmod) with filter="tar" of files outside the extraction directory. You are affected by this vulnerability if using the tarfile module to extract untrusted tar archives using TarFile.extractall() or TarFile.extract() using the filter= parameter with a value of "data" or "tar". See the tarfile extraction filters documentation https://docs.python.org/3/library/tarfile.html#tarfile-extraction-filter for more information. Only Python versions 3.12 or later are affected by these vulnerabilities, earlier versions don't include the extraction filter feature. Note that for Python 3.14 or later the default value of filter= changed from "no filtering" to `"data", so if you are relying on this new default behavior then your usage is also affected. Note that none of these vulnerabilities significantly affect the installation of source distributions which are tar archives as source distributions already allow arbitrary code execution during the build process. However when evaluating source distributions it's important to avoid installing source distributions with suspicious links.
-
CVE-2025-13435
A security vulnerability has been detected in Dreampie Resty up to 1.3.1.SNAPSHOT. This affects the function Request of the file /resty-httpclient/src/main/java/cn/dreampie/client/HttpClient.java of the component HttpClient Module. Such manipulation of the argument filename leads to path traversal. The attack may be performed from remote. Attacks of this nature are highly complex. The exploitability is reported as difficult. The exploit has been disclosed publicly and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
-
CVE-2025-30224
MyDumper is a MySQL Logical Backup Tool. The MySQL C client library (libmysqlclient) allows authenticated remote actors to read arbitrary files from client systems via a crafted server response to LOAD LOCAL INFILE query, leading to sensitive information disclosure when clients connect to untrusted MySQL servers without explicitly disabling the local infile capability. Mydumper has the local infile option enabled by default and does not have an option to disable it. This can lead to an unexpected arbitrary file read if the Mydumper tool connects to an untrusted server. This vulnerability is fixed in 0.18.2-8.
-
CVE-2025-25295
Label Studio is an open source data labeling tool. A path traversal vulnerability in Label Studio SDK versions prior to 1.0.10 allows unauthorized file access outside the intended directory structure. The flaw exists in the VOC, COCO and YOLO export functionalities. These functions invoke a `download` function on the `label-studio-sdk` python package, which fails to validate file paths when processing image references during task exports. By creating tasks with path traversal sequences in the image field, an attacker can force the application to read files from arbitrary server filesystem locations when exporting projects in any of the mentioned formats. This is authentication-required vulnerability allowing arbitrary file reads from the server filesystem. It may lead to potential exposure of sensitive information like configuration files, credentials, and confidential data. Label Studio versions before 1.16.0 specified SDK versions prior to 1.0.10 as dependencies, and the issue was confirmed in Label Studio version 1.13.2.dev0; therefore, Label Studio users should upgrade to 1.16.0 or newer to mitigate it.
-
CVE-2025-8021
All versions of the package files-bucket-server are vulnerable to Directory Traversal where an attacker can traverse the file system and access files outside of the intended directory.
-
CVE-2025-10284
BBOT's unarchive module could be abused by supplying malicious archives files and when extracted can then perform an arbitrary file write, resulting in remote code execution.
-
CVE-2025-54590
webfinger.js is a TypeScript-based WebFinger client that runs in both browsers and Node.js environments. In versions 2.8.0 and below, the lookup function accepts user addresses for account checking. However, the ActivityPub specification requires preventing access to localhost services in production. This library does not prevent localhost access, only checking for hosts that start with "localhost" and end with a port. Users can exploit this by creating servers that send GET requests with controlled host, path, and port parameters to query services on the instance's host or local network, enabling blind SSRF attacks. This is fixed in version 2.8.1.
-
CVE-2025-12638
Keras version 3.11.3 is affected by a path traversal vulnerability in the keras.utils.get_file() function when extracting tar archives. The vulnerability arises because the function uses Python's tarfile.extractall() method without the security-critical filter='data' parameter. Although Keras attempts to filter unsafe paths using filter_safe_paths(), this filtering occurs before extraction, and a PATH_MAX symlink resolution bug triggers during extraction. This bug causes symlink resolution to fail due to path length limits, resulting in a security bypass that allows files to be written outside the intended extraction directory. This can lead to arbitrary file writes outside the cache directory, enabling potential system compromise or malicious code execution. The vulnerability affects Keras installations that process tar archives with get_file() and does not affect versions where this extraction method is secured with the appropriate filter parameter.
-
CVE-2025-9118
A path traversal vulnerability in the NPM package installation process of Google Cloud Dataform allows a remote attacker to read and write files in other customers' repositories via a maliciously crafted package.json file.
-
CVE-2024-11343
In Progress® Telerik® Document Processing Libraries, versions prior to 2025 Q1 (2025.1.205), unzipping an archive can lead to arbitrary file system access.
-
CVE-2025-32013
LNbits is a Lightning wallet and accounts system. A Server-Side Request Forgery (SSRF) vulnerability has been discovered in LNbits' LNURL authentication handling functionality. When processing LNURL authentication requests, the application accepts a callback URL parameter and makes an HTTP request to that URL using the httpx library with redirect following enabled. The application doesn't properly validate the callback URL, allowing attackers to specify internal network addresses and access internal resources.
-
CVE-2025-8959
HashiCorp's go-getter library subdirectory download feature is vulnerable to symlink attacks leading to unauthorized read access beyond the designated directory boundaries. This vulnerability, identified as CVE-2025-8959, is fixed in go-getter 1.7.9.
-
CVE-2025-25185
GPT Academic provides interactive interfaces for large language models. In 3.91 and earlier, GPT Academic does not properly account for soft links. An attacker can create a malicious file as a soft link pointing to a target file, then package this soft link file into a tar.gz file and upload it. Subsequently, when accessing the decompressed file from the server, the soft link will point to the target file on the victim server. The vulnerability allows attackers to read all files on the server.
-
CVE-2025-52913
A vulnerability in the NuPoint Unified Messaging (NPM) component of Mitel MiCollab through 9.8 SP2 (9.8.2.12) could allow an unauthenticated attacker to conduct a path traversal attack due to insufficient input validation. A successful exploit could allow unauthorized access, enabling the attacker to view, corrupt, or delete users' data and system configurations.
-
CVE-2025-31492
mod_auth_openidc is an OpenID Certified authentication and authorization module for the Apache 2.x HTTP server that implements the OpenID Connect Relying Party functionality. Prior to 2.4.16.11, a bug in a mod_auth_openidc results in disclosure of protected content to unauthenticated users. The conditions for disclosure are an OIDCProviderAuthRequestMethod POST, a valid account, and there mustn't be any application-level gateway (or load balancer etc) protecting the server. When you request a protected resource, the response includes the HTTP status, the HTTP headers, the intended response (the self-submitting form), and the protected resource (with no headers). This is an example of a request for a protected resource, including all the data returned. In the case where mod_auth_openidc returns a form, it has to return OK from check_userid so as not to go down the error path in httpd. This means httpd will try to issue the protected resource. oidc_content_handler is called early, which has the opportunity to prevent the normal output being issued by httpd. oidc_content_handler has a number of checks for when it intervenes, but it doesn't check for this case, so the handler returns DECLINED. Consequently, httpd appends the protected content to the response. The issue has been patched in mod_auth_openidc versions >= 2.4.16.11.
-
CVE-2025-29787
`zip` is a zip library for rust which supports reading and writing of simple ZIP files. In the archive extraction routine of affected versions of the `zip` crate starting with version 1.3.0 and prior to version 2.3.0, symbolic links earlier in the archive are allowed to be used for later files in the archive without validation of the final canonicalized path, allowing maliciously crafted archives to overwrite arbitrary files in the file system when extracted. Users who extract untrusted archive files using the following high-level API method may be affected and critical files on the system may be overwritten with arbitrary file permissions, which can potentially lead to code execution. Version 2.3.0 fixes the issue.
-
CVE-2025-0454
A Server-Side Request Forgery (SSRF) vulnerability was identified in the Requests utility of significant-gravitas/autogpt versions prior to v0.4.0. The vulnerability arises due to a hostname confusion between the `urlparse` function from the `urllib.parse` library and the `requests` library. A malicious user can exploit this by submitting a specially crafted URL, such as `http://localhost:\@google.com/../`, to bypass the SSRF check and perform an SSRF attack.
-
CVE-2025-25065
SSRF vulnerability in the RSS feed parser in Zimbra Collaboration 9.0.0 before Patch 43, 10.0.x before 10.0.12, and 10.1.x before 10.1.4 allows unauthorized redirection to internal network endpoints.
-
CVE-2025-65025
esm.sh is a nobuild content delivery network(CDN) for modern web development. Prior to version 136, the esm.sh CDN service is vulnerable to path traversal during NPM package tarball extraction. An attacker can craft a malicious NPM package containing specially crafted file paths (e.g., package/../../tmp/evil.js). When esm.sh downloads and extracts this package, files may be written to arbitrary locations on the server, escaping the intended extraction directory. This issue has been patched in version 136.
-
CVE-2025-61594
URI is a module providing classes to handle Uniform Resource Identifiers. In versions 0.12.4 and earlier (bundled in Ruby 3.2 series) 0.13.2 and earlier (bundled in Ruby 3.3 series), 1.0.3 and earlier (bundled in Ruby 3.4 series), when using the + operator to combine URIs, sensitive information like passwords from the original URI can be leaked, violating RFC3986 and making applications vulnerable to credential exposure. This is a a bypass for the fix to CVE-2025-27221 that can expose user credentials. This issue has been fixed in versions 0.12.5, 0.13.3 and 1.0.4.
-
CVE-2023-35816
DevExpress before 23.1.3 allows arbitrary TypeConverter conversion.
-
CVE-2025-1211
Versions of the package hackney before 1.21.0 are vulnerable to Server-side Request Forgery (SSRF) due to improper parsing of URLs by URI built-in module and hackey. Given the URL http://[email protected]/, the URI function will parse and see the host as 127.0.0.1 (which is correct), and hackney will refer the host as 127.2.2.2/. This vulnerability can be exploited when users rely on the URL function for host checking.
-
CVE-2025-8267
Versions of the package ssrfcheck before 1.2.0 are vulnerable to Server-Side Request Forgery (SSRF) due to an incomplete denylist of IP address ranges. Specifically, the package fails to classify the reserved IP address space 224.0.0.0/4 (Multicast) as invalid. This oversight allows attackers to craft requests targeting these multicast addresses.
-
CVE-2025-27136
LocalS3 is an Amazon S3 mock service for testing and local development. Prior to version 1.21, the LocalS3 service's bucket creation endpoint is vulnerable to XML External Entity (XXE) injection. When processing the CreateBucketConfiguration XML document during bucket creation, the service's XML parser is configured to resolve external entities. This allows an attacker to declare an external entity that references an internal URL, which the server will then attempt to fetch when parsing the XML. The vulnerability specifically occurs in the location constraint processing, where the XML parser resolves external entities without proper validation or restrictions. When the external entity is resolved, the server makes an HTTP request to the specified URL and includes the response content in the parsed XML document. This vulnerability can be exploited to perform server-side request forgery (SSRF) attacks, allowing an attacker to make requests to internal services or resources that should not be accessible from external networks. The server will include the responses from these internal requests in the resulting bucket configuration, effectively leaking sensitive information. The attacker only needs to be able to send HTTP requests to the LocalS3 service to exploit this vulnerability.
-
CVE-2025-6985
The HTMLSectionSplitter class in langchain-text-splitters version 0.3.8 is vulnerable to XML External Entity (XXE) attacks due to unsafe XSLT parsing. This vulnerability arises because the class allows the use of arbitrary XSLT stylesheets, which are parsed using lxml.etree.parse() and lxml.etree.XSLT() without any hardening measures. In lxml versions up to 4.9.x, external entities are resolved by default, allowing attackers to read arbitrary local files or perform outbound HTTP(S) fetches. In lxml versions 5.0 and above, while entity expansion is disabled, the XSLT document() function can still read any URI unless XSLTAccessControl is applied. This vulnerability allows remote attackers to gain read-only access to any file the LangChain process can reach, including sensitive files such as SSH keys, environment files, source code, or cloud metadata. No authentication, special privileges, or user interaction are required, and the issue is exploitable in default deployments that enable custom XSLT.
-
CVE-2025-31490
AutoGPT is a platform that allows users to create, deploy, and manage continuous artificial intelligence agents that automate complex workflows. Prior to 0.6.1, AutoGPT allows SSRF due to DNS Rebinding in requests wrapper. AutoGPT is built with a wrapper around Python's requests library, hardening the application against SSRF. The code for this wrapper can be found in autogpt_platform/backend/backend/util/request.py. The requested hostname of a URL which is being requested is validated, ensuring that it does not resolve to any local ipv4 or ipv6 addresses. However, this check is not sufficient, as a DNS server may initially respond with a non-blocked address, with a TTL of 0. This means that the initial resolution would appear as a non-blocked address. In this case, validate_url() will return the url as successful. After validate_url() has successfully returned the url, the url is then passed to the real request() function. When the real request() function is called with the validated url, request() will once again resolve the address of the hostname, because the record will not have been cached (due to TTL 0). This resolution may be in the "invalid range". This type of attack is called a "DNS Rebinding Attack". This vulnerability is fixed in 0.6.1.
-
CVE-2025-0184
A Server-Side Request Forgery (SSRF) vulnerability was identified in langgenius/dify version 0.10.2. The vulnerability occurs in the 'Create Knowledge' section when uploading DOCX files. If an external relationship exists in the DOCX file, the reltype value is requested as a URL using the 'requests' module instead of the 'ssrf_proxy', leading to an SSRF vulnerability. This issue was fixed in version 0.11.0.
-
CVE-2025-3445
A Path Traversal "Zip Slip" vulnerability has been identified in mholt/archiver in Go. This vulnerability allows using a crafted ZIP file containing path traversal symlinks to create or overwrite files with the user's privileges or application utilizing the library. When using the archiver.Unarchive functionality with ZIP files, like this: archiver.Unarchive(zipFile, outputDir), A crafted ZIP file can be extracted in such a way that it writes files to the affected system with the same privileges as the application executing this vulnerable functionality. Consequently, sensitive files may be overwritten, potentially leading to privilege escalation, code execution, and other severe outcomes in some cases. It's worth noting that a similar vulnerability was found in TAR files (CVE-2024-0406). Although a fix was implemented, it hasn't been officially released, and the affected project has since been deprecated. The successor to mholt/archiver is a new project called mholt/archives, and its initial release (v0.1.0) removes the Unarchive() functionality.
-
CVE-2025-14198
A vulnerability was detected in Verysync 微力同步 2.21.3. This affects an unknown function of the file /safebrowsing/clientreport/download?key=dummytoken of the component Web Administration Module. Performing manipulation results in information disclosure. The attack is possible to be carried out remotely. The exploit is now public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
-
CVE-2025-31491
AutoGPT is a platform that allows users to create, deploy, and manage continuous artificial intelligence agents that automate complex workflows. Prior to 0.6.1, AutoGPT allows of leakage of cross-domain cookies and protected headers in requests redirect. AutoGPT uses a wrapper around the requests python library, located in autogpt_platform/backend/backend/util/request.py. In this wrapper, redirects are specifically NOT followed for the first request. If the wrapper is used with allow_redirects set to True (which is the default), any redirect is not followed by the initial request, but rather re-requested by the wrapper using the new location. However, there is a fundamental flaw in manually re-requesting the new location: it does not account for security-sensitive headers which should not be sent cross-origin, such as the Authorization and Proxy-Authorization header, and cookies. For example in autogpt_platform/backend/backend/blocks/github/_api.py, an Authorization header is set when retrieving data from the GitHub API. However, if GitHub suffers from an open redirect vulnerability (such as the made-up example of https://api.github.com/repos/{owner}/{repo}/issues/comments/{comment_id}/../../../../../redirect/?url=https://joshua.hu/), and the script can be coerced into visiting it with the Authorization header, the GitHub credentials in the Authorization header will be leaked. This allows leaking auth headers and private cookies. This vulnerability is fixed in 0.6.1.
-
CVE-2025-57814
request-filtering-agent is an http(s).Agent implementation that blocks requests to Private/Reserved IP addresses. Versions 1.x.x and earlier contain a vulnerability where HTTPS requests to 127.0.0.1 bypass IP address filtering, while HTTP requests are correctly blocked. This allows attackers to potentially access internal HTTPS services running on localhost, bypassing the library's SSRF protection. The vulnerability is particularly dangerous when the application accepts user-controlled URLs and internal services are only protected by network-level restrictions. This vulnerability has been fixed in request-filtering-agent version 2.0.0. Users should upgrade to version 2.0.0 or later.
-
CVE-2023-35817
DevExpress before 23.1.3 allows AsyncDownloader SSRF.
CVE records published in 2025 (ids starting CVE-2025-), sorted by how likely they are to be exploited.