VAITP

Gradio is an open-source Python package designed for quick prototyping. Prior to version 6.6.0, a Server-Side Request Forgery (SSRF) vulnerability in Gradio allows an attacker to make arbitrary HTTP requests from a victim's server by hosting a malicious Gradio Space. When a victim application uses `gr.load()` to load an attacker-controlled Space, the malicious `proxy_url` from the config is trusted and added to the allowlist, enabling the attacker to access internal services, cloud metadata endpoints, and private networks through the victim's infrastructure. Version 6.6.0 fixes the issue.

Gradio is an open-source Python package designed for quick prototyping. Prior to version 6.6.0, the _redirect_to_target() function in Gradio's OAuth flow accepts an unvalidated _target_url query parameter, allowing redirection to arbitrary external URLs. This affects the /logout and /login/callback endpoints on Gradio apps with OAuth enabled (i.e. apps running on Hugging Face Spaces with gr.LoginButton). Starting in version 6.6.0, the _target_url parameter is sanitized to only use the path, query, and fragment, stripping any scheme or host.

Gradio is an open-source Python package designed for quick prototyping. Prior to version 6.7, Gradio apps running on Window with Python 3.13+ are vulnerable to an absolute path traversal issue that enables unauthenticated attackers to read arbitrary files from the file system. Python 3.13+ changed the definition of `os.path.isabs` so that root-relative paths like `/windows/win.ini` on Windows are no longer considered absolute paths, resulting in a vulnerability in Gradio's logic for joining paths safely. This can be exploited by unauthenticated attackers to read arbitrary files from the Gradio server, even when Gradio is set up with authentication. Version 6.7 fixes the issue.

Gradio is an open-source Python package designed for quick prototyping. Starting in version 4.16.0 and prior to version 6.6.0, Gradio applications running outside of Hugging Face Spaces automatically enable "mocked" OAuth routes when OAuth components (e.g. `gr.LoginButton`) are used. When a user visits `/login/huggingface`, the server retrieves its own Hugging Face access token via `huggingface_hub.get_token()` and stores it in the visitor's session cookie. If the application is network-accessible, any remote attacker can trigger this flow to steal the server owner's HF token. The session cookie is signed with a hardcoded secret derived from the string `"-v4"`, making the payload trivially decodable. Version 6.6.0 fixes the issue.

The NVDA Dev & Test Toolbox is an NVDA add-on for gathering tools to help NVDA development and testing. A vulnerability exists in versions 2.0 through 8.0 in the Log Reader feature of this add-on. A maliciously crafted log file can lead to arbitrary code execution when a user reads it with log reader commands. The log reading command process speech log entries in an unsafe manner. Python expressions embedded in the log may be evaluated when when speech entries are read with log reading commands. An attacker can exploit this by convincing a user to open a malicious crafted log file and to analyze it using the log reading commands. When the log is read, attacker-controlled code may execute with the privileges of the current user. This issue does not require elevated privileges and relies solely on user interaction (opening the log file). Version 9.0 contains a fix for the issue. As a workaround, avoid using log reading commands, or at least, commands to move to next/previous log message (any message or commands for each type of message). For more security, one may disable their gestures in the input gesture dialog.

Unitree Go2 firmware versions 1.1.7 through 1.1.11, when used with the Unitree Go2 Android application (com.unitree.doggo2), are vulnerable to remote code execution due to missing integrity protection and validation of user-created programmes. The Android application stores programs in a local SQLite database (unitree_go2.db, table dog_programme) and transmits the programme_text content, including the pyCode field, to the robot. The robot's actuator_manager.py executes the supplied Python as root without integrity verification or content validation. An attacker with local access to the Android device can tamper with the stored programme record to inject arbitrary Python that executes when the user triggers the program via a controller keybinding, and the malicious binding persists across reboots. Additionally, a malicious program shared through the application's community marketplace can result in arbitrary code execution on any robot that imports and runs it.

Unitree Go2 firmware versions V1.1.7 through V1.1.9 and V1.1.11 (EDU) do not implement DDS authentication or authorization for the Eclipse CycloneDDS topic rt/api/programming_actuator/request handled by actuator_manager.py. A network-adjacent, unauthenticated attacker can join DDS domain 0 and publish a crafted message (api_id=1002) containing arbitrary Python, which the robot writes to disk under /unitree/etc/programming/ and binds to a physical controller keybinding. When the keybinding is pressed, the code executes as root and the binding persists across reboots.

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.8.0, the CSV Agent node in Langflow hardcodes `allow_dangerous_code=True`, which automatically exposes LangChain’s Python REPL tool (`python_repl_ast`). As a result, an attacker can execute arbitrary Python and OS commands on the server via prompt injection, leading to full Remote Code Execution (RCE). Version 1.8.0 fixes the issue.

Agenta is an open-source LLMOps platform. In Agenta-API prior to version 0.48.1, a Python sandbox escape vulnerability existed in Agenta's custom code evaluator. Agenta used RestrictedPython as a sandboxing mechanism for user-supplied evaluator code, but incorrectly whitelisted the `numpy` package as safe within the sandbox. This allowed authenticated users to bypass the sandbox and achieve arbitrary code execution on the API server. The escape path was through `numpy.ma.core.inspect`, which exposes Python's introspection utilities — including `sys.modules` — thereby providing access to unfiltered system-level functionality like `os.system`. This vulnerability affects the Agenta self-hosted platform (API server), not the SDK when used as a standalone Python library. The custom code evaluator runs server-side within the API process. The issue is fixed in v0.48.1 by removing `numpy` from the sandbox allowlist. In later versions (v0.60+), the RestrictedPython sandbox was removed entirely and replaced with a different execution model.

psd-tools is a Python package for working with Adobe Photoshop PSD files. Prior to version 1.12.2, when a PSD file contains malformed RLE-compressed image data (e.g. a literal run that extends past the expected row size), decode_rle() raises ValueError which propagated all the way to the user, crashing psd.composite() and psd-tools export. decompress() already had a fallback that replaces failed channels with black pixels when result is None, but it never triggered because the ValueError from decode_rle() was not caught. The fix in version 1.12.2 wraps the decode_rle() call in a try/except so the existing fallback handles the error gracefully.