CVE-2026-58474
Published:August 26, 2026
Updated:September 24, 2026
whichllm before 0.5.16 contains a code injection vulnerability in the run and snippet commands that allows a remote attacker who controls a HuggingFace repository to achieve arbitrary code execution by crafting a malicious GGUF filename containing double quotes or other special characters. The script generation function in cli.py interpolates HuggingFace-derived values, including GGUF variant filenames from the Hub API siblings rfilename field, directly into Python source code without escaping, allowing the crafted filename to break out of the generated string literal and execute injected code on the user's machine before any model download occurs.
Affected Packages
https://github.com/Andyyyy64/whichllm.git (GITHUB):
Affected version(s) >=v0.5.0 <v0.5.16Fix Suggestion:
Update to version v0.5.16whichllm (PYTHON):
Affected version(s) >=0.1.0 <0.5.16Fix Suggestion:
Update to version 0.5.16Related Resources (7)
Do you need more information?
Contact UsCVSS v4
Base Score:
8.6
Attack Vector
NETWORK
Attack Complexity
LOW
Attack Requirements
NONE
Privileges Required
NONE
User Interaction
ACTIVE
Vulnerable System Confidentiality
HIGH
Vulnerable System Integrity
HIGH
Vulnerable System Availability
HIGH
Subsequent System Confidentiality
NONE
Subsequent System Integrity
NONE
Subsequent System Availability
NONE
CVSS v3
Base Score:
8.8
Attack Vector
NETWORK
Attack Complexity
LOW
Privileges Required
NONE
User Interaction
REQUIRED
Scope
UNCHANGED
Confidentiality
HIGH
Integrity
HIGH
Availability
HIGH
Weakness Type (CWE)
Improper Control of Generation of Code ('Code Injection')
EPSS
Base Score:
0.46