Security Vulnerabilities
- CVEs Published In March 2026
The '/logs' and '/logs-stream' endpoints in the log router allow any authenticated user to read the full application log buffer. These endpoints only require basic authentication ('get_current_active_user') without any privilege checks (e.g., 'is_superuser').
The '/api/v1/files/images/{flow_id}/{file_name}' endpoint serves SVG files with the 'image/svg+xml' content type without sanitizing their content.
Since SVG files can contain embedded JavaScript, an attacker can upload a malicious SVG that executes arbitrary JavaScript when viewed by other users, leading to stored cross-site scripting (XSS). This allows stealing authentication tokens stored in cookies, including JWT access and refresh tokens.
The 'POST /api/v2/files' endpoint does not sanitize the 'filename' parameter from the multipart form data, allowing an attacker to write files to arbitrary locations on the filesystem using path traversal sequences ('../').
A flaw has been found in OpenBMB XAgent 1.0.0. The impacted element is the function FunctionHandler.handle_tool_call of the file XAgent/function_handler.py of the component API Key Handler. This manipulation of the argument api_key causes sensitive information in log files. The attack may be initiated remotely. The exploit has been published and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
A local file disclosure vulnerability in the XInclude processing component of Inkscape 1.1 before 1.3 allows a remote attacker to read local files via a crafted SVG file containing malicious xi:include tags.
The Twilio integration webhook handler accepts any POST request without validating Twilio's 'X-Twilio-Signature'.
When processing media messages, it fetches user-controlled URLs ('MediaUrlN' parameters) using HTTP requests that include the integration's Twilio credentials in the 'Authorization' header.
An attacker can forge a webhook payload pointing to their own server and receive the victim's 'accountSID' and 'authToken' in plaintext (base64-encoded Basic Auth), leading to full compromise of the Twilio account.
WWBN AVideo is an open source video platform. In versions up to and including 26.0, the `objects/playlistsVideos.json.php` endpoint returns the full video contents of any playlist by ID without any authentication or authorization check. Private playlists (including `watch_later` and `favorite` types) are correctly hidden from listing endpoints via `playlistsFromUser.json.php`, but their contents are directly accessible through this endpoint by providing the sequential integer `playlists_id` parameter. Commit bb716fbece656c9fe39784f11e4e822b5867f1ca has a patch for the issue.
WWBN AVideo is an open source video platform. In versions up to and including 26.0, three `list.json.php` endpoints in the Scheduler plugin lack any authentication check, while every other endpoint in the same plugin directories (`add.json.php`, `delete.json.php`, `index.php`) requires `User::isAdmin()`. An unauthenticated attacker can retrieve all scheduled tasks (including internal callback URLs and parameters), admin-composed email messages, and user-to-email targeting mappings by sending simple GET requests. Commit 83390ab1fa8dca2de3f8fa76116a126428405431 contains a patch.
WWBN AVideo is an open source video platform. In versions up to and including 26.0, the `get_api_video_password_is_correct` API endpoint allows any unauthenticated user to verify whether a given password is correct for any password-protected video. The endpoint returns a boolean `passwordIsCorrect` field with no rate limiting, CAPTCHA, or authentication requirement, enabling efficient offline-speed brute-force attacks against video passwords. Commit 01a0614fedcdaee47832c0d913a0fb86d8c28135 contains a patch.
WWBN AVideo is an open source video platform. In versions up to and including 26.0, the AI plugin's `save.json.php` endpoint loads AI response objects using an attacker-controlled `$_REQUEST['id']` parameter without validating that the AI response belongs to the specified video. An authenticated user with AI permissions can reference any AI response ID — including those generated for other users' private videos — and apply the stolen AI-generated content (titles, descriptions, keywords, summaries, or full transcriptions) to their own video, effectively exfiltrating the information. Commit aa2c46a806960a0006105df47765913394eec142 contains a patch.