CVE-2023-37274
HighCVSS 7.5Exploitation Probability (EPSS)
Low risk25th percentile - higher than 25% of all known CVEs
Summary
Auto-GPT is an experimental open-source application that had a vulnerability in the `execute_python_code` command before version 0.4.3, allowing for a path traversal attack. This enables overwriting .py files outside the workspace directory, potentially leading to arbitrary code execution on the host.
Risk Assessment
This vulnerability poses a significant security risk as it can lead to unauthorized code execution on the machine running Auto-GPT.
Recommendation
It is recommended to upgrade to version 0.4.3, which fixes this vulnerability. Alternatively, Auto-GPT can be run in a virtual machine to minimize the risk of file damage.
Other vulnerabilities in Auto-GPT
See all- CVE-2023-37275Low
Auto-GPT is an experimental open-source application that, before version 0.4.3, allowed malicious external resources to trigger misleading messages in the console by using ANSI escape sequences. The issue has been patched in release version 0.4.3.
- CVE-2023-37273High
Auto-GPT versions prior to 0.4.3 have a vulnerability that allows the docker-compose.yml file in the container to be overwritten, potentially leading to control over the host system. This issue has been patched in version 0.4.3.
Original NVD description (English source)
Auto-GPT is an experimental open-source application showcasing the capabilities of the GPT-4 language model. When Auto-GPT is executed directly on the host system via the provided run.sh or run.bat files, custom Python code execution is sandboxed using a temporary dedicated docker container which should not have access to any files outside of the Auto-GPT workspace directory. Before v0.4.3, the `execute_python_code` command (introduced in v0.4.1) does not sanitize the `basename` arg before writing LLM-supplied code to a file with an LLM-supplied name. This allows for a path traversal attack that can overwrite any .py file outside the workspace directory by specifying a `basename` such as `../../../main.py`. This can further be abused to achieve arbitrary code execution on the host running Auto-GPT by e.g. overwriting autogpt/main.py which will be executed outside of the docker environment meant to sandbox custom python code execution the next time Auto-GPT is started. The issue has been patched in version 0.4.3. As a workaround, the risk introduced by this vulnerability can be remediated by running Auto-GPT in a virtual machine, or another environment in which damage to files or corruption of the program is not a critical problem.