CVE-2024-12366
Published: 11 February 2025
Description
Adversaries may exploit software vulnerabilities in client applications to execute code.
Security Summary
CVE-2024-12366 is a critical vulnerability in PandasAI, an AI-enhanced library for data analysis using large language models (LLMs). The issue resides in an interactive prompt function susceptible to prompt injection attacks, enabling attackers to execute arbitrary Python code and achieve remote code execution (RCE). Instead of providing the intended explanation of natural language processing queries via the LLM, the function can be hijacked. Published on 2025-02-11, it carries a CVSS v3.1 base score of 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H).
Unauthenticated remote attackers can exploit this vulnerability over the network with low attack complexity and no user interaction required. Exploitation allows full control over the affected system through arbitrary Python code execution, resulting in high impacts to confidentiality, integrity, and availability.
Mitigation guidance is available in vendor documentation and advisories, including https://docs.getpanda.ai/v3/privacy-security, https://docs.pandas-ai.com/advanced-security-agent, and the CERT vulnerability note at https://www.kb.cert.org/vuls/id/148244.
This vulnerability underscores prompt injection risks in AI/ML tools that combine LLMs with code execution capabilities, such as PandasAI's integration for natural language-driven data processing.
Details
- CWE(s)
AI Security Analysis
- AI Category
- Data Processing Libraries
- Risk Domain
- LLM/Generative AI Risks
- OWASP Top 10 for LLMs 2025
- MITRE ATLAS Techniques
- Classification Reason
- PandasAI is an AI extension to the Pandas data processing library, enabling natural language queries via LLMs on dataframes, making it primarily a data processing library with AI integration.
MITRE ATT&CK Enterprise Techniques
Why these techniques?
The prompt injection vulnerability enables remote code execution via arbitrary Python code, mapping to T1059.006 (Command and Scripting Interpreter: Python) for execution and T1203 (Exploitation for Client Execution) as it exploits a software vulnerability in PandasAI.