Browse / Direct Prompt Injection / CVE-2023-39661
PandasAI vulnerable to arbitrary code execution
- Class
- Direct Prompt Injection (filed by a model that read the advisory)
- Technique
- Code and command execution
- Also involves
- Improper Output Handling
- Affected
- pandasai (PyPI)
- Gets past
- Input filter or sanitiser
- CVSS
- CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
- EPSS
- 1.4% chance of exploitation in 30 days
How it works, in short
Natural-language prompts can evade the jailbreak keyword check, so model-generated code runs on the host. (Our summary, written by a model. The source text is below.)
Details
An issue in pandas-ai v.0.8.1 and before allows a remote attacker to execute arbitrary code via the _is_jailbreak function.
References
- https://nvd.nist.gov/vuln/detail/CVE-2023-39661
- https://github.com/gventuri/pandas-ai/issues/410
- https://github.com/gventuri/pandas-ai
- https://github.com/pypa/advisory-database/tree/main/vulns/pandasai/PYSEC-2026-447.yaml
- https://pypi.org/project/pandasai
Source: GitHub Advisory Database, used under CC BY 4.0.