llama_index vulnerable to SQL Injection

The SQL Injection vulnerability in llama-index poses a significant risk to applications utilizing the library. An attacker could exploit this flaw to read sensitive data, including user credentials, personal information, and application configuration details. Furthermore, the attacker might be able to modify data, potentially corrupting the application's database or injecting malicious content. The blast radius extends to any application leveraging llama-index in a web environment, particularly those handling user-supplied data that is subsequently used in SQL queries. This vulnerability shares characteristics with other SQL injection attacks, where improper input sanitization allows attackers to bypass security controls and directly interact with the underlying database.

Applications built using the llama-index library, particularly those deployed in web environments and handling user-provided data, are at significant risk. Shared hosting environments where multiple applications share a database are especially vulnerable, as a successful attack could compromise data across multiple tenants. Legacy applications using older versions of llama-index are also at increased risk.

• python / llama-index: Inspect code for usage of vector store integrations without proper input sanitization. Look for raw SQL queries constructed using user-supplied data.

import llama_index

# Vulnerable code example (do not use)
query = f"SELECT * FROM users WHERE username = '{user_input}'"
response = index.query(query)

• python / llama-index: Monitor application logs for SQL errors or unusual database activity. Implement logging to capture all SQL queries executed by the application. • generic web: Use a web application firewall (WAF) to detect and block SQL injection attempts targeting llama-index endpoints.

1. 2025-06-05Disclosure

   disclosure

1. Reservado2025-02-28
2. Publicada2025-06-05
3. Modificada2026-02-04
4. EPSS atualizado2026-03-23

The primary mitigation for CVE-2025-1793 is to upgrade to llama-index version 0.12.28 or later. If an immediate upgrade is not feasible due to compatibility issues or breaking changes, consider implementing input validation and sanitization techniques to prevent SQL injection attacks. Specifically, use parameterized queries or prepared statements to ensure that user-supplied data is treated as data and not as executable code. Web application firewalls (WAFs) configured to detect and block SQL injection attempts can provide an additional layer of defense. After upgrading, confirm the fix by attempting to inject SQL code into the affected vector store integrations and verifying that the queries are properly sanitized.