Security Vulnerabilities
- CVEs Published In March 2025
A Denial of Service (DoS) vulnerability exists in the file upload feature of imartinez/privategpt version v0.6.2. The vulnerability is due to improper handling of form-data with a large filename in the file upload request. An attacker can exploit this by sending a payload with an excessively large filename, causing the server to become overwhelmed and unavailable to legitimate users.
A vulnerability in haotian-liu/llava version 1.2.0 (LLaVA-1.6) allows for Server-Side Request Forgery (SSRF) through the /run/predict endpoint. An attacker can gain unauthorized access to internal networks or the AWS metadata endpoint by sending crafted requests that exploit insufficient validation of the path parameter. This flaw can lead to unauthorized network access, sensitive data exposure, and further exploitation within the network.
A Server-Side Request Forgery (SSRF) vulnerability exists in lm-sys/fastchat version 0.2.36. The vulnerability is present in the `/queue/join?` endpoint, where insufficient validation of the path parameter allows an attacker to send crafted requests. This can lead to unauthorized access to internal networks or the AWS metadata endpoint, potentially exposing sensitive data and compromising internal servers.
A privilege escalation vulnerability exists in langgenius/dify version 0.9.1. This vulnerability allows a normal user to modify Orchestrate instructions for a chatbot created by an admin user. The issue arises because the application does not properly enforce access controls on the endpoint /console/api/apps/{chatbot-id}/model-config, allowing unauthorized users to alter chatbot configurations.
langgenius/dify version 0.9.1 contains a Server-Side Request Forgery (SSRF) vulnerability. The vulnerability exists due to improper handling of the api_endpoint parameter, allowing an attacker to make direct requests to internal network services. This can lead to unauthorized access to internal servers and potentially expose sensitive information, including access to the AWS metadata endpoint.
A stored cross-site scripting (XSS) vulnerability exists in langgenius/dify version latest, specifically in the chat log functionality. The vulnerability arises because certain HTML tags like <input> and <form> are not disallowed, allowing an attacker to inject malicious HTML into the log via prompts. When an admin views the log containing the malicious HTML, the attacker could steal the admin's credentials or sensitive information. This issue is fixed in version 0.12.1.
A stored cross-site scripting (XSS) vulnerability exists in the latest version of langgenius/dify. The vulnerability is due to improper validation and sanitization of user input in SVG markdown support within the chatbot feature. An attacker can exploit this vulnerability by injecting malicious SVG content, which can execute arbitrary JavaScript code when viewed by an admin, potentially leading to credential theft.
An improper access control vulnerability in danny-avila/librechat versions prior to 0.7.6 allows authenticated users to delete other users' prompts via the groupid parameter. This issue occurs because the endpoint does not verify whether the provided prompt ID belongs to the current user.
An unhandled exception in danny-avila/librechat version 3c94ff2 can lead to a server crash. The issue occurs when the fs module throws an exception while handling file uploads. An unauthenticated user can trigger this exception by sending a specially crafted request, causing the server to crash. The vulnerability is fixed in version 0.7.6.
A vulnerability in danny-avila/librechat version git 81f2936 allows for path traversal due to improper sanitization of file paths by the multer middleware. This can lead to arbitrary file write and potentially remote code execution. The issue is fixed in version 0.7.6.