CVE-2024-10940: 
Python vulnerability analysis and mitigation

A vulnerability has been identified in langchain-core versions >=0.1.17,<0.1.53, >=0.2.0,<0.2.43, and >=0.3.0,<0.3.15 that allows unauthorized users to read arbitrary files from the host file system. The vulnerability was discovered and reported through huntr.dev, and has been assigned CVE-2024-10940 (NVD).

The vulnerability stems from the ability to create langchain_core.prompts.ImagePromptTemplate's (and by extension langchain_core.prompts.ChatPromptTemplate's) with input variables that can read any user-specified path from the server file system. When the outputs of these prompt templates are exposed to the user, either directly or through downstream model outputs, it can lead to the exposure of sensitive information. The vulnerability has been assigned a CVSS v3.0 base score of 5.3 (Medium) with the vector string CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N (NVD).

The vulnerability allows unauthorized users to read arbitrary files from the host file system, potentially exposing sensitive information. This could lead to the disclosure of confidential data stored on the server where the langchain-core application is running (NVD).

The vulnerability can be exploited by creating ImagePromptTemplate instances with specifically crafted input variables that target sensitive files on the host system. No special privileges or user interaction is required for exploitation, making it relatively straightforward to exploit (NVD).

The vulnerability has been patched in the latest versions of langchain-core. The fix involves removing the ability to load images from local file paths, requiring users to specify images by URL instead. Users should upgrade to versions >=0.1.53, >=0.2.43, or >=0.3.15 depending on their major version (Github Commit).