CVE-2026-54017
HighCVSS 7.7Exploitation Probability (EPSS)
Low risk28th percentile - higher than 28% of all known CVEs
Summary
In versions prior to 0.9.6 of the Open WebUI platform, there is a vulnerability in the terminal-server reverse proxy that allows authenticated users to exploit `../` sequences to escape the intended path scope, potentially leading to unauthorized access to files and endpoints on the terminal server.
Risk Assessment
This vulnerability poses a risk of unauthorized access to internal services and files, which could lead to serious security breaches within the organization.
Recommendation
It is recommended to upgrade to version 0.9.6 or later to mitigate this vulnerability and implement additional security measures to restrict access to terminals.
Related vulnerabilities
- CVE-2026-86149Critical
A vulnerability has been identified in Tenda CP3 27.5.57.101 in the file Net/NetCheckPing.cpp. Manipulation of the interface_name/host arguments causes OS command injection. The attack can be initiated remotely.
- CVE-2026-86148Critical
A vulnerability has been discovered in Tenda CP3 27.5.57.101 in the SystemAsh function of the file Apis/system.c of the Kylin component. Manipulation of the AlarmVoiceURL argument results in OS command injection. The attack can be launched remotely.
- CVE-2026-86060Critical
RouterOS contains an argument-handling flaw in the SSH login path involving usernames that begin with a prohibited character, allowing the trusted RouterOS policy mask to be changed, leading to privilege escalation. Exploitation requires an unauthenticated SSH session to reach the RouterOS login helper. This issue was fixed in versions: 6.49.21 (Long-term), 7.23.4 (Long-term) and 7.24.2 (Stable).
- CVE-2026-67276Critical
RouterOS does not compare the complete RSA public key when matching an SSH authentication request to an authorized user key, checking the key type and modulus but omitting the exponent. Because signature verification uses the client-supplied key, an attacker knowing an authorized RSA modulus can supply a key with exponent one, forge a valid signature, and open an SSH command channel as the target user without the private key. This issue was fixed in versions: 6.49.21 (Long-term), 7.23.4 (Long-term) and 7.24.2 (Stable).
- CVE-2026-86190Critical
WWBN AVideo contains a broken access control vulnerability in videoViewsInfo endpoints that returns complete user records including password hashes, recovery tokens, and live session identifiers to unauthenticated callers when a hash parameter is provided. Attackers can use the disclosed session identifier to hijack viewer sessions, including administrator accounts, and obtain sensitive personal data for all video viewers.
- CVE-2026-86189Critical
WWBN AVideo contains a path traversal vulnerability in notify.ffmpeg.json.php that allows unauthenticated attackers to write files to arbitrary locations by supplying a caller-chosen path in the avideoRelativePath parameter. Attackers can replay any previously issued ciphertext as a notifyCode token, which is decrypted but never validated, to bypass authentication and write files to the application root and subdirectories.
- CVE-2026-86184Critical
Lara Dashboard before 1.3.0 contains an authentication bypass vulnerability in the screenshot-login route that allows unauthenticated attackers to authenticate as any user by email when APP_ENV is not production. Attackers can request the GET /screenshot-login/{email} endpoint with a registered email address to receive a fully authenticated session, enabling access to user administration, settings, database contents, and arbitrary code execution through the module installer.
- CVE-2026-10196Critical
The Mail Mint – Email Marketing, Newsletter, Email Automation & WooCommerce Emails plugin for WordPress is vulnerable to PHP Object Injection in all versions up to, and including, 1.31.0 via deserialization of untrusted input in the 'handle_form_submission' function. This makes it possible for unauthenticated attackers to inject a PHP Object. The additional presence of a POP chain allows attackers to execute code on the server. The vulnerability was partially patched in version 1.23.1.
- CVE-2026-86124Critical
AutoAgent contains an unauthenticated remote code execution vulnerability in the TCP server that binds to all interfaces and executes attacker-supplied commands as root. Attackers can connect to the exposed communication port and execute arbitrary bash commands within the container, gaining access to bind-mounted host workspace directories.
- CVE-2026-86121Critical
Cua computer-server versions before 0.3.42 skip authentication when the CONTAINER_NAME environment variable is unset and bind to all interfaces by default, allowing unauthenticated attackers to execute arbitrary commands. Attackers can reach TCP port 8000 to run shell commands via the run_command endpoint, read and write arbitrary files through file operation endpoints, and access interactive PTY shells without authentication.
Original NVD description (English source)
Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, the terminal-server reverse proxy in `backend/open_webui/routers/terminals.py` does not fully confine the user-controlled `path` segment before forwarding it to an admin-configured terminal server. An authenticated user who has been granted access to a terminal server can craft `path` values containing encoded `../` traversal sequences that escape the intended path (or policy) scope on that server, reaching unintended endpoints and files on the terminal-server host. Where the terminal server fans requests out to internal services, this also gives SSRF-style reach into those services. This is a separate code path from the `/api/v1/retrieval/process/web` SSRF (GHSA-c6xv-rcvw-v685), with its own input. Two distinct vectors are consolidated here: first, raw path forwarding / single-encoded traversal (original report); and second, a bypass of the subsequently-added `_sanitize_proxy_path` mitigation using double-encoded dots (`%252e%252e`). The attacker-controlled input is the request `path`, supplied by the non-admin user, not anything an administrator configures, so this is not an admin-trust / Rule-9 situation. Version 0.9.6 fixes the issue.

