NIXPKGS-2025-0004

published on 1 Nov 2025

Permalink CVE-2025-10854

8.1 HIGH

CVSS version: 3.1
Attack vector (AV): NETWORK
Attack complexity (AC): HIGH
Privileges required (PR): NONE
User interaction (UI): NONE
Scope (S): UNCHANGED
Confidentiality impact (C): HIGH
Integrity impact (I): HIGH
Availability impact (A): HIGH

updated 4 months, 3 weeks ago by @LeSuisse Activity log

Created automatic suggestion 5 months, 3 weeks ago
@LeSuisse removed
2 packages
- python312Packages.llama-index-readers-txtai
- python313Packages.llama-index-readers-txtai
4 months, 3 weeks ago
@LeSuisse accepted 4 months, 3 weeks ago
@LeSuisse published on GitHub 4 months, 3 weeks ago

Symlink Following in txtai leads to arbitrary file write when loading untrusted embedding indices

The txtai framework allows the loading of compressed tar files as embedding indices. While the validate function is intended to prevent path traversal vulnerabilities by ensuring safe filenames, it does not account for symbolic links within the tar file. An attacker is able to write a file anywhere in the filesystem when txtai is used to load untrusted embedding indices