ParisNeo/lollms-webui

LoLLMs WEBUI provides the Web user interface for Lord of Large Language and Multi modal Systems. A critical Server-Side Request Forgery (SSRF) vulnerability has been identified in all known existing versions of `lollms-webui`. The `@router.post("/api/proxy")` endpoint allows unauthenticated attackers to force the server into making arbitrary GET requests. This can be exploited to access internal services, scan local networks, or exfiltrate sensitive cloud metadata (e.g., AWS/GCP IAM tokens). As of time of publication, no known patched versions are available.

A Local File Inclusion (LFI) vulnerability exists in the '/reinstall_extension' endpoint of the parisneo/lollms-webui application, specifically within the `name` parameter of the `@router.post("/reinstall_extension")` route. This vulnerability allows attackers to inject a malicious `name` parameter, leading to the server loading and executing arbitrary Python files from the upload directory for discussions. This issue arises due to the concatenation of `data.name` directly with `lollmsElfServer.lollms_paths.extensions_zoo_path` and its use as an argument for `ExtensionBuilder().build_extension()`. The server's handling of the `__init__.py` file in arbitrary locations, facilitated by `importlib.machinery.SourceFileLoader`, enables the execution of arbitrary code, such as command execution or creating a reverse-shell connection. This vulnerability affects the latest version of parisneo/lollms-webui and can lead to Remote Code Execution (RCE) when the application is exposed to an external endpoint or the UI, especially when bound to `0.0.0.0` or in `headless mode`. No user interaction is required for exploitation.

parisneo/lollms-webui version 9.6 is vulnerable to Cross-Site Scripting (XSS) and Open Redirect due to inadequate input validation and processing of SVG files during the upload process. The XSS vulnerability allows attackers to embed malicious JavaScript code within SVG files, which is executed upon rendering, leading to potential credential theft and unauthorized data access. The Open Redirect vulnerability arises from insufficient URL validation within SVG files, enabling attackers to redirect users to malicious websites, thereby exposing them to phishing attacks, malware distribution, and reputation damage. These vulnerabilities are present in the application's functionality to send files to the AI module.

A CORS misconfiguration in parisneo/lollms-webui prior to version 10 allows attackers to steal sensitive information such as logs, browser sessions, and settings containing private API keys from other services. This vulnerability can also enable attackers to perform actions on behalf of a user, such as deleting a project or sending a message. The issue impacts the confidentiality and integrity of the information.

A Cross-Site Request Forgery (CSRF) vulnerability exists in the `install_comfyui` endpoint of the `lollms_comfyui.py` file in the parisneo/lollms-webui repository, versions v9.9 to the latest. The endpoint uses the GET method without requiring a client ID, allowing an attacker to trick a victim into installing ComfyUI. If the victim's device does not have sufficient capacity, this can result in a crash.

A path traversal vulnerability exists in the 'cyber_security/codeguard' native personality of the parisneo/lollms-webui, affecting versions up to 9.5. The vulnerability arises from the improper limitation of a pathname to a restricted directory in the 'process_folder' function within 'lollms-webui/zoos/personalities_zoo/cyber_security/codeguard/scripts/processor.py'. Specifically, the function fails to properly sanitize user-supplied input for the 'code_folder_path', allowing an attacker to specify arbitrary paths using '../' or absolute paths. This flaw leads to arbitrary file read and overwrite capabilities in specified directories without limitations, posing a significant risk of sensitive information disclosure and unauthorized file manipulation.

A path traversal vulnerability exists in the parisneo/lollms-webui application, specifically within the `lollms_core/lollms/server/endpoints/lollms_binding_files_server.py` and `lollms_core/lollms/security.py` files. Due to inadequate validation of file paths between Windows and Linux environments using `Path(path).is_absolute()`, attackers can exploit this flaw to read any file on the system. This issue affects the latest version of LoLLMs running on the Windows platform. The vulnerability is triggered when an attacker sends a specially crafted request to the `/user_infos/{path:path}` endpoint, allowing the reading of arbitrary files, as demonstrated with the `win.ini` file. The issue has been addressed in version 9.5 of the software.

A path traversal and arbitrary file upload vulnerability exists in the parisneo/lollms-webui application, specifically within the `@router.get("/switch_personal_path")` endpoint in `./lollms-webui/lollms_core/lollms/server/endpoints/lollms_user.py`. The vulnerability arises due to insufficient sanitization of user-supplied input for the `path` parameter, allowing an attacker to specify arbitrary file system paths. This flaw enables direct arbitrary file uploads, leakage of `personal_data`, and overwriting of configurations in `lollms-webui`->`configs` by exploiting the same named directory in `personal_data`. The issue affects the latest version of the application and is fixed in version 9.4. Successful exploitation could lead to sensitive information disclosure, unauthorized file uploads, and potentially remote code execution by overwriting critical configuration files.

A Cross-Site Request Forgery (CSRF) vulnerability exists in the profile picture upload functionality of the Lollms application, specifically in the parisneo/lollms-webui repository, affecting versions up to 7.3.0. This vulnerability allows attackers to change a victim's profile picture without their consent, potentially leading to a denial of service by overloading the filesystem with files. Additionally, this flaw can be exploited to perform a stored cross-site scripting (XSS) attack, enabling attackers to execute arbitrary JavaScript in the context of the victim's browser session. The issue is resolved in version 9.3.

A vulnerability in parisneo/lollms-webui versions up to 9.3 allows remote attackers to execute arbitrary code. The vulnerability stems from insufficient protection of the `/apply_settings` and `/execute_code` endpoints. Attackers can bypass protections by setting the host to localhost, enabling code execution, and disabling code validation through the `/apply_settings` endpoint. Subsequently, arbitrary commands can be executed remotely via the `/execute_code` endpoint, exploiting the delay in settings enforcement. This issue was addressed in version 9.5.

A path traversal vulnerability exists in the 'save_settings' endpoint of the parisneo/lollms-webui application, affecting versions up to the latest release before 9.5. The vulnerability arises due to insufficient sanitization of the 'config' parameter in the 'apply_settings' function, allowing an attacker to manipulate the application's configuration by sending specially crafted JSON payloads. This could lead to remote code execution (RCE) by bypassing existing patches designed to mitigate such vulnerabilities.

A command injection vulnerability exists in the 'run_xtts_api_server' function of the parisneo/lollms-webui application, specifically within the 'lollms_xtts.py' script. The vulnerability arises due to the improper neutralization of special elements used in an OS command. The affected function utilizes 'subprocess.Popen' to execute a command constructed with a Python f-string, without adequately sanitizing the 'xtts_base_url' input. This flaw allows attackers to execute arbitrary commands remotely by manipulating the 'xtts_base_url' parameter. The vulnerability affects versions up to and including the latest version before 9.5. Successful exploitation could lead to arbitrary remote code execution (RCE) on the system where the application is deployed.

parisneo/lollms-webui is vulnerable to authentication bypass due to insufficient protection over sensitive endpoints. The application checks if the host parameter is not '0.0.0.0' to restrict access, which is inadequate when the application is bound to a specific interface, allowing unauthorized access to endpoints such as '/restart_program', '/update_software', '/check_update', '/start_recording', and '/stop_recording'. This vulnerability can lead to denial of service, unauthorized disabling or overriding of recordings, and potentially other impacts if certain features are enabled in the configuration.

An SQL injection vulnerability exists in the `delete_discussion()` function of the parisneo/lollms-webui application, allowing an attacker to delete all discussions and message data. The vulnerability is exploitable via a crafted HTTP POST request to the `/delete_discussion` endpoint, which internally calls the vulnerable `delete_discussion()` function. By sending a specially crafted payload in the 'id' parameter, an attacker can manipulate SQL queries to delete all records from the 'discussion' and 'message' tables. This issue is due to improper neutralization of special elements used in an SQL command.

parisneo/lollms-webui is vulnerable to a denial of service (DoS) attack due to uncontrolled resource consumption. Attackers can exploit the `/open_code_in_vs_code` and similar endpoints without authentication by sending repeated HTTP POST requests, leading to the opening of Visual Studio Code or the default folder opener (e.g., File Explorer, xdg-open) multiple times. This can render the host machine unusable by exhausting system resources. The vulnerability is present in the latest version of the software.

A Local File Inclusion (LFI) vulnerability exists in the parisneo/lollms-webui application, specifically within the `/personalities` route. An attacker can exploit this vulnerability by crafting a URL that includes directory traversal sequences (`../../`) followed by the desired system file path, URL encoded. Successful exploitation allows the attacker to read any file on the filesystem accessible by the web server. This issue arises due to improper control of filename for include/require statement in the application.

An OS Command Injection vulnerability exists in the '/open_code_folder' endpoint of the parisneo/lollms-webui application, due to improper validation of user-supplied input in the 'discussion_id' parameter. Attackers can exploit this vulnerability by injecting malicious OS commands, leading to unauthorized command execution on the underlying operating system. This could result in unauthorized access, data leakage, or complete system compromise.

A Cross-Site Request Forgery (CSRF) vulnerability in the parisneo/lollms-webui project allows remote attackers to execute arbitrary code on a victim's system. The vulnerability stems from the `/execute_code` API endpoint, which does not properly validate requests, enabling an attacker to craft a malicious webpage that, when visited by a victim, submits a form to the victim's local lollms-webui instance to execute arbitrary OS commands. This issue allows attackers to take full control of the victim's system without requiring direct network access to the vulnerable application.