All vulnerabilities
CVE-2025-1750GHSA-6mcc-q3mg-3qg6PYSEC-2025-245PYSEC-2026-4242

llama-index-vector-stores-duckdb: LlamaIndex DuckDBVectorStore vulnerable to SQL injection through ref_doc_id

run-llama llama-index-vector-stores-duckdbpipLibraries & packages
Open in ThreatlineRecord JSON
CVSS
9.8Critical
ExploitationPublic exploit codeCISA's assessment says proof of concept code is public, assessed 2 Jun 2025. Not in the KEV catalogCISA via CVE List
Exploit odds (EPSS)0.82%Chance of exploitation activity in the next 30 days, higher than 56% of CVEs. Scored 23 Oct 2026FIRST EPSS
Fixed in0.12.21Also 0.3.1GitHub
Published2 Jun 2025Updated 15 Oct 2026CVE List

What fixes it

  • 0.12.21llama-index-vector-stores-duckdb, pipGitHub
  • 0.3.1run-llama/llama_indexCVE List
  • 0.12.21llama-index-vector-stores-duckdb, PyPIOSVpip install "llama-index-vector-stores-duckdb>=0.12.21"

A fixed version on one release line is not the upgrade target for every line. Check your exact version and any vendor backports.

Description

An SQL injection vulnerability exists in the delete function of DuckDBVectorStore in run-llama/llama_index version v0.12.19. This vulnerability allows an attacker to manipulate the ref_doc_id parameter, enabling them to read and write arbitrary files on the server, potentially leading to remote code execution (RCE).

Text from CVE List

Affected versions 3

  • llama-index-vector-stores-duckdb, pip>= 0.12.19, < 0.12.21GitHub
  • run-llama/llama_indexaffected: unspecified before 0.3.1CVE List
  • llama-index-vector-stores-duckdb, PyPIfrom 0.12.19, fixed in 0.12.21OSV

References 6