Back github.com

Mem0 Issues 4073

github.com • October 23, 2026

The Databricks vector store implementation uses f-string interpolation for SQL queries, creating SQL injection vulnerabilities.

File: mem0/vector_stores/databricks.py

Comparison with Safe Implementations

Other vector stores in the same codebase correctly use parameterized queries:

pgvector: cur.execute(f"DELETE FROM ... WHERE id = %s", (vector_id,)) ✅

azure_mysql: Uses %s placeholders ✅

Databricks: Uses f-string interpolation ❌

A malicious memory_id like '; DELETE FROM table; -- would execute:

Use the Databricks Statement Execution API's parameterized queries:

I attempted to use GitHub's private vulnerability reporting but it appears to be disabled. Consider enabling it at Settings → Code security → Private vulnerability reporting.

Discovered during security audit by Lighthouse Research Project ( )

TopangaLudwitt added a commit that references this issue on Feb 20, 2026 fix: prevent SQL injection in Databricks vector store This issue will close once commit da5d45d is merged into the 'main' branch. da5d45d

fix: prevent SQL injection in Databricks vector store

Anandesh-Sharma added a commit that references this issue on Feb 26, 2026 fix: use parameterized queries in Databricks vector store to prevent … This issue will close once commit 11569e6 is merged into the 'main' branch. 11569e6

fix: use parameterized queries in Databricks vector store to prevent …

blazingretinol added bug Something isn't working Something isn't working security Security vulnerability Security vulnerability P0-critical Data loss, security vuln, blocks all users Data loss, security vuln, blocks all users on Mar 24, 2026

utkarsh240799 added a commit that references this issue on Mar 26, 2026 fix: prevent SQL injection in Databricks vector store ( #4073 ) c6cc8f1

fix: prevent SQL injection in Databricks vector store ( #4073 )

blazingretinol closed this as completed in #4558 on Mar 26, 2026

harshgupta-mem0 assigned blazingretinol on Apr 14, 2026

oyaah added a commit that references this issue on Jun 7, 2026 fix: prevent SQL injection in Databricks vector store ( mem0ai#4073 ) 3743d30

fix: prevent SQL injection in Databricks vector store ( mem0ai#4073 )

The full story

This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.

SQL Injection Vulnerability in mem0 AI's Databricks Integration A SQL injection vulnerability (CVE-2026-115324) was identified in mem0 version 0.1.117 before 1.0.8, affecting the Databricks vector store implementation. The flaw allows attackers to inject malicious SQL through unescaped f-string interpolation in the insert() and update() methods, potentially leading to data exposure or disruption. The vulnerability was disclosed on October 22, 2026, with a CVSS score of 9.2, marking it as. Users of the Databricks vector store integration are particularly at risk, especially where untrusted content can be supplied. The recommended mitigation includes upgrading to the fixed version, disabling the integration, and restricting access to the warehouse. No exploitation in the wild has been reported yet, but the complexity of the attack may limit immediate threats. Users are advised to review logs for suspicious activity until a patch is applied. 1 sources 78

Extracted Entities

Mem0 Issues 4073

The full story

Extracted Entities

Attack Types (1)

CWE Weaknesses (1)

MITRE ATT&CK (1)