CVE Feed

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, backend/open_webui/utils/oauth.py::_process_picture_url calls validate_url(picture_url) on the initial URL only, then invokes aiohttp.ClientSession.get(picture_url, ...) without allow_redirects=False. aiohttp's default is allow_redirects=True, max_redirects=10; the function does not pass the project's AIOHTTP_CLIENT_ALLOW_REDIRECTS env constant either. An attacker with a valid OAuth IdP identity can therefore submit a public URL that 302-redirects to an internal address and read the internal response body via the attacker's own profile_image_url field. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, POST /api/chat/completions accepts an image_url.url value that, when it does NOT start with http://, https://, or data:image/, is interpreted as a file id and resolved against the global file table with no ownership check. an authenticated user can therefore set image_url.url to another user's file id, the server reads that file from disk, base64-encodes it, and injects the data URI into the LLM request. the user then prompts the LLM to describe / OCR the file and reads the content back. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, Open WebUI lets an authenticated user attach arbitrary file_id values to their own chat message without checking whether they own or can read those files. If the attacker then shares that chat and grants themselves read access, has_access_to_file() treats the victim file as accessible through the shared chat, and the file endpoints read or delete the victim file. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6,Open WebUI renders Mermaid blocks from Markdown files in the file preview panel and inserts the generated SVG into the DOM using innerHTML. Because Mermaid is configured with securityLevel: 'loose', attacker-controlled Mermaid content can be rendered unsafely in this flow. A working payload was validated through the Markdown preview path, resulting in JavaScript execution in the victim’s browser under the application origin. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, Open WebUI lets a user who can create, update, or import workspace models store arbitrary meta.knowledge entries on their model without checking whether they own or can read the referenced files. Open WebUI then treats meta.knowledge entries of type file as an authorization source in two places: the built-in view_file tool reads the file's extracted text, and has_access_to_file()'s model branch authorizes the file content and file delete endpoints. A malicious model owner can therefore attach another user's file ID to their model metadata and read or delete that private file. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, Open WebUI patched SVG XSS in user profile images and webhook profile images but forgot to apply the same fix to model profile images. The ModelMeta class has no validate_profile_image_url field validator, and the model image serving endpoint has no MIME allowlist or nosniff header. Any authenticated user with workspace.models permission (enabled by default) can store a data:image/svg+xml;base64,... payload in a model's profile image and achieve full account takeover of anyone who navigates to the image URL. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, a path traversal vulnerability exists in open-webui's cache file serving endpoint that allows any authenticated user to read files from sibling directories outside the intended cache directory, by exploiting an incomplete startswith containment check that lacks a trailing path separator. The root cause is that serve_cache_file() in open_webui/main.py validates the resolved path with file_path.startswith(os.path.abspath(CACHE_DIR)) — without appending os.sep. This allows any path resolving to a sibling directory whose name begins with cache (e.g. cache_sibling, cache_backup, cached_models) to pass validation. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, Open WebUI's prompt version-history endpoints authorize the prompt_id in the URL but then act on caller-supplied history IDs without verifying that the history row belongs to that prompt (history_entry.prompt_id == prompt.id). This affects /api/v1/prompts/id/{prompt_id}/history/diff, /api/v1/prompts/id/{prompt_id}/update/version, and /api/v1/prompts/id/{prompt_id}/history/{history_id}. An authenticated user with access to any prompt they control, plus a victim prompt_history.id, can read or delete another user's private prompt history. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, Open WebUI has a Broken Object Level Authorization (BOLA) vulnerability in the builtin search_knowledge_files tool. When native function calling is enabled and the selected model has no attached knowledge bases, an authenticated user can call search_knowledge_files with an arbitrary knowledge_id. The function then returns file metadata from that knowledge base without checking whether the user has read access. This allows unauthorized enumeration of private or restricted knowledge base files. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, the SafePlaywrightURLLoader implements a validate_url function to prevent SSRF attacks by checking the IP address of the user-provided URL. However, this validation is performed only on the initial URL. Since Playwright automatically follows HTTP redirects (301/302) by default, an attacker can bypass the validation by providing a safe URL that redirects to a restricted internal network address (e.g., localhost, Docker container network, or Cloud Metadata). This allows the application to access internal services despite ENABLE_RAG_LOCAL_WEB_FETCH being set to False This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, Open WebUI added collection-level ACL checks, but the patch can still be bypassed when Milvus multitenancy mode is enabled. The ACL allows unknown non-KB collection names as legacy/ephemeral collections. In Milvus multitenancy mode, that user-controlled collection name becomes a resource_id and is interpolated into a Milvus expression without escaping. This is caused by an incomplete fix for CVE-2026-44560 This vulnerability is fixed in 0.9.6.

GNU libidn before 1.44 is prone to out-of-bounds reads of uninitialized memory in the ToUnicode APIs because of mishandling in idna_to_unicode_internal. The affected code is not present in libidn2.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, several direct, index-addressed Ollama proxy routes accept a caller-supplied url_idx path parameter and use it as a raw index into the admin-configured OLLAMA_BASE_URLS list. Access control on these routes validates only whether the user may use the requested model, never which backend the request is routed to. Any authenticated user can append an arbitrary url_idx to force their request onto an Ollama backend they were never authorized to reach, including internal, higher-privilege, or explicitly admin-disabled backends. This vulnerability is fixed in 0.9.6.

Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.8.11, the ydoc:document:join Socket.IO handler checks note ownership only when the document_id starts with note: (colon). However, the YdocManager storage layer normalizes all document IDs by replacing colons with underscores (document_id.replace(":", "_")). An attacker can join a document room using note_<id> (underscore) instead of note:<id> (colon), bypassing the authorization check entirely while accessing the same underlying Yjs document. The server then returns the full document state, leaking the victim's private note contents. This vulnerability is fixed in 0.8.11.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.10.0, the "Shareable Playground" (or "Public Flows" in code) contains a potential arbitrary file-read vulnerability, depending on the exact flow configuration used. By making a flow public, public execution of the flow is allowed. The execution request can contain a list of files that gets read by Langflow and fed into the LLM. The files path can be any path supported by the storage - it can be either a local file or S3 path if supported by the local configuration This vulnerability is fixed in 1.10.0.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.0, Langflow's /api/v1/monitor router exposes 7 endpoints that perform read, write, and delete operations on user-owned resources — messages, sessions, build artifacts, and LLM transaction logs — without verifying that the authenticated requester owns the targeted resource. Any authenticated user can read, modify, rename, or permanently delete another user's data by supplying the target's resource ID or flow_id. This is a classic IDOR/BOLA vulnerability. Notably, the same source file (monitor.py) contains one correctly-implemented endpoint that uses an ownership check, demonstrating the correct pattern was known but inconsistently applied. This vulnerability is fixed in 1.9.0.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.0, Langflow is vulnerable to Path Traversal in the Knowledge Bases API (POST /api/v1/knowledge_bases). This occurs because user-supplied knowledge base names are used directly to create file paths without proper sanitization or containment checks. An authenticated attacker can exploit this flaw to create directories and write files anywhere on the server's filesystem. This vulnerability is fixed in 1.9.0.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.1, an Insecure Direct Object Reference (IDOR) vulnerability in /api/v1/responses endpoint allows an authenticated attacker to execute any flow belonging to another user by specifying the victim's flow ID in the request. This vulnerability is fixed in 1.9.1.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.7.0, the logout button does not clear the session. The previous user stays logged in unless another user explicitly logs in. This vulnerability is fixed in 1.7.0.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.0.19, an attacker can send a /api/v1/files/upload/ request without any authentication token/cookies and abuse a very long multipart form boundary to make the langflow app unusable for all users for an indefinite amount of time. This vulnerability is fixed in 1.0.19.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.2, the "Shareable Playground" (or "Public Flows" in code) contains a critical RCE vulnerability. Shareable Playground feature works by enabling the execution of workflows by unauthenticated users, by accessing a link. Specifically, it enables the route /api/v1/build_public_tmp to execute any public flow, given a public flow ID. When the route executes the flow, it allows for providing arbitrary custom Python code as the nodes code, inside the JSON payload. The vulnerable field is data.nodes[X].data.node.template.code.value. This vulnerability is fixed in 1.9.2.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.2, by controlling a files that are digested into the RAG, an attacker can direct the node to read any file on the file-system by absolute path. All components based on BaseFileComponent are vulnerable to the vulnerability. This includes Docling (DoclingInlineComponent), Docling Serve, DoclingRemoteComponent), Read File (FileComponent), NVIDIA Retriever Extraction (NvidiaIngestComponent), Video File (VideoFileComponent), and Unstructured API (UnstructuredComponent). This vulnerability is fixed in 1.9.2.

GNU SASL before 2.2.4 lacks sanitization of a short challenge in _gsasl_ntlm_client_step in the NTLM client, which could result in memory disclosure via a crafted server.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to 1.9.1, unauthenticated users can upload any amount of data to the server without any limitations. No need for any prior knowledge, only network access to Langflow. This can lead to space exhaustion on the server. In addition, in the response, the absolute path of the uploaded file is reported to the attacker, which is an information leak that can assist in chaining other primitives. This vulnerability is fixed in 1.9.1.

A missing access control check when linking trackers to campaigns through the campaign-trackers.php script of Revive Adserver 6.0.6 and earlier could allow a low‑privileged user to link their trackers to campaigns owned by other managers on the same instance, resulting in inconsistent ownership relationships. Ownership validation has been added to ensure that campaigns can only be linked to trackers owned by the same advertiser.

Low‑privileged session IDs generated for the web admin console could be reused in the XML‑RPC API, whose authentication is normally restricted to admin users. An attacker could leverage this to gain unauthorised access and exploit API‑level vulnerabilities. The session context (web/API) is now recorded along with other session data, preventing session IDs from being used interchangeably.

Low‑privileged users could use their Full Name as a vector for a stored XSS attack. The name is included in system‑generated emails, whose content is stored in the details field of the userlog table. An admin user viewing the email content through userlog-details.php would have any malicious JavaScript payload executed due to missing output sanitisation. Proper escaping has been added to the userlog details output.

A missing validation of user input when saving delivery limitations in Revive Adserver 6.0.6 and earlier could allow a low‑privileged user to use the logical parameter to inject malicious PHP code into the compiledlimitations field on the database and have it executed during banner delivery. Input sanitisation has been improved to ensure that the parameter is properly validated.

A missing sanitisation of user input in the zone-include.php script of Revive Adserver 6.0.6 and earlier. A low‑privileged user could exploit the clientid parameter to perform blind SQL injection attacks. Input sanitisation has been improved to ensure that all parameters processed by the script are properly validated.

An access control bypass allows an advertiser‑level user to activate or deactivate a banner in Revive Adserver 6.0.6 and earlier, even when such permissions were not granted. The banner-edit.php script allowed the banner status to be overwritten solely based on banner edit permissions. The status field has been removed from the hidden form fields in the banner edit screen.

The XML‑RPC API addUser method has a validation bypass introduced in the fix for CVE‑2025‑55129. As a result, API users could create usernames that enabled impersonation or stored XSS attacks. Proper validation has been added where it was missing.

A stored XSS can be exploited by leveraging the usernames as an attack vector. When an admin user viewed the audit log details for affected entries, any malicious JavaScript payload embedded in the username would be executed due to missing output sanitisation. Proper escaping has been added to the audit log details output.

A missing access control check when invoking various modify methods in the XML‑RPC API of Revive Adserver 6.0.6 and earlier. The API allowed entities to be reassigned to different parent entities, leading to inconsistent ownership relationships. This issue was exploitable only in combination with CVE‑2026‑34917 or with third‑party API extensions that expose API functionality to low‑privileged users. Access control checks have been added to validate access to parent entities in the API modify methods.

A missing access control check when linking banners or campaigns to a zone through the zone-include.php script of Revive Adserver 6.0.6 and earlier, or via its API allows a low‑privileged user could link their zones to banners or campaigns owned by other managers on the same instance, resulting in inconsistent ownership relationships. Ownership validation has been added to ensure that banners and campaigns can only be linked to zones managed by the same account.

A missing validation of user input exists when saving delivery limitations in Revive Adserver 6.0.6 and earlier. A low‑privileged user could add an unexpected component parameter and inject malicious PHP code into the compiledlimitations field, which would then be executed during banner delivery. Input sanitisation has been improved to ensure that unexpected parameters are filtered out.

A missing sanitisation of user input in the zone-include.php script of Revive Adserver 6.0.6 and earlier could allow a low‑privileged user to exploit the clientid parameter to perform blind SQL injection attacks. Input sanitisation has been improved to ensure that all parameters processed by the script are properly validated.

dhcpcd through 10.3.2, fixed in commit 78ea09e, contains a heap use-after-free vulnerability in the control socket handling within src/control.c that allows local unprivileged attackers to trigger memory corruption when privilege separation is disabled. Attackers can connect to the control socket and send a privileged command such as -x, causing control_recvdata() to free the client object while the same READ+HANGUP event subsequently reaches control_hangup() with the stale pointer, resulting in a use-after-free condition exploitable in deployments using --disable-privsep or where privsep initialization has failed with the control socket operating in mode 0666.

yt-dlp is a command-line audio/video downloader. From 2023.09.24 until 2026.06.09, if curl is used as an external downloader for yt-dlp, cookies may be leaked to an unintended host upon HTTP redirect or when the host for download fragments differs from their parent manifest's. At the file download stage, the cookies are passed by yt-dlp to the file downloader via --cookie. However, unless these are loaded from a file, this operation does not activate the cookie engine. As a result, curl will send cookies with requests to domains or paths for which the cookies are not scoped. This vulnerability is fixed in 2026.06.09.

Uncontrolled Search Path Element vulnerability in ABB Control Builder A, ABB 800xA for Advant Master. This issue affects Control Builder A: through 1.4/4; 800xA for Advant Master: through 6.0.3-1, through 6.1.1-1, 6.1.1-3, 6.2.0-1.

dhcpcd through 10.3.2, fixed in commit 708b4a5, contains a memory leak vulnerability in the IPv6 Router Advertisement route information handling that allows an unauthenticated same-link attacker to cause denial of service by sending crafted Router Advertisements. Attackers can repeatedly send Router Advertisements containing Route Information options with a lifetime of zero, triggering unfreed allocations in routeinfo_findalloc() that cause linear memory exhaustion and eventual daemon crash.

yt-dlp is a command-line audio/video downloader. Prior to 2026.06.09, if aria2c is used as an external downloader for a fragmented manifest format (such as an HLS/DASH stream), yt-dlp passes insufficiently sanitized input to aria2c that allows an attacker to perform an arbitrary file write. On Windows platforms, this can lead to immediate arbitrary code execution. On non-Windows platforms, this can lead to arbitrary code execution upon the next invocation of yt-dlp. This vulnerability is fixed in 2026.06.09.

Bootimus through 0.1.70 contains a broken access control vulnerability that allows authenticated low-privileged users to perform administrative actions by exploiting missing role enforcement in the JWTMiddleware function in internal/auth/auth.go, which validates JWT tokens and account status but fails to inspect the is_admin flag. Attackers can send requests to any endpoint under the /api/users path to create new administrator accounts or reset administrator passwords, thereby gaining full control of the server and the ability to modify boot menus and installation scripts served to PXE clients.

yt-dlp is a command-line audio/video downloader. Prior to 2026.06.09, a vulnerability exists in yt-dlp that allows a remote attacker to write arbitrary OS-shortcut files (such as .desktop, .url, .webloc) to the user's filesystem, bypassing the remediation for CVE-2024-38519. The allowlist explicitly included the unsafe extensions .desktop, .url, and .webloc so that the functionality of the --write-link option (and its variants) could be preserved. These allowlist inclusions can be exploited by an attacker to write malicious OS-shortcut files in the context of a media or subtitles download. This vulnerability is fixed in 2026.06.09.

dhcpcd through 10.3.2, fixed in commit 2f00c7b, contains a one-byte stack out-of-bounds write vulnerability in dhcp6_makemessage() in src/dhcp6.c that allows unauthenticated same-link attackers to write beyond a fixed local buffer by serializing an oversized RFC6603 OPTION_PD_EXCLUDE option body. Attackers can send a crafted DHCPv6 ADVERTISE message containing an IA_PD IAPREFIX /0 with a valid OPTION_PD_EXCLUDE using an exclude prefix length of /121 through /128 to trigger the out-of-bounds write and potentially corrupt adjacent stack memory.

dhcpcd through 10.3.2, fixed in commit 5733d3c, contains a heap use-after-free vulnerability that allows unauthenticated same-link attackers to crash the daemon by sending a crafted DHCPv6 RENEW reply with RFC6603 OPTION_PD_EXCLUDE and both preferred and valid lifetimes set to zero. Attackers acting as or impersonating a DHCPv6 server can trigger dhcp6_deprecatedele() to free a delegated child address while an outer TAILQ_FOREACH_SAFE iterator in dhcp6_deprecateaddrs() still holds the freed pointer, causing a use-after-free when TAILQ_REMOVE is reached.

tarfile.extractall() with the 'data' or 'tar' filter could be bypassed by a crafted archive where a hardlink references a symlink stored at a deeper name than the hardlink itself.  The extraction fallback validated the symlink at it's archived location but recreated it at the hardlink's shallower path, letting a relative target the filter judged contained escape the destination directory.  This allowed a malicious tar archive to create a symlink pointing outside the destination, enabling out-of-destination file reads or writes. This was an incomplete fix of CVE-2025-4330.

Missing symlink validation in Language Servers for AWS may allow an arbitrary file write outside of the workspace trust boundary. This may occur when a local user opens a workspace with a maliciously crafted symlink that resolves to a file path outside the workspace trust boundary. To remediate this issue, users should upgrade to version 1.69.0 or higher.

Improper trust boundary enforcement in Language Servers for AWS before version 1.65.0 on all supported platforms may allow a for arbitrary code execution. If a local user opens a maliciously crafted workspace, any commands within the project configuration files may be automatically executed. This issue requires the user to trust the workspace when prompted. To remediate this issue, users should upgrade to Language Servers for AWS version 1.65.0 or higher.

Tenable Identity Exposure contains multiple unauthenticated API endpoints under /w/api/* that expose sensitive application configuration data including cleartext LDAP credentials, SAML configuration, user accounts, and directory settings to unauthenticated remote attackers. Affected responses are served with Cache-Control: public headers and without Vary: Cookie, allowing reverse proxies and CDNs to cache and serve sensitive data to unauthenticated users even after authentication is applied.

n8n is an open source workflow automation platform. Prior to 1.123.43, 2.22.1, and 2.20.7, an attacker with write access to the git repository connected to an n8n Source Control configuration could commit a malicious Data Table JSON file containing a crafted column name. When an administrator performed a Source Control Pull, n8n imported the file and could lead to SQL injection on the internal PostgreSQL instance. Exploitation requires the n8n instance uses PostgreSQL as its database backend, the Source Control feature is enabled and connected to a repository the attacker can write to, and an administrator triggers a Source Control Pull. This vulnerability is fixed in 1.123.43, 2.22.1, and 2.20.7.