Skip to main content

Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection')

CVE-2026-25879

Low
0/10

Summary

SQLChatAgent executes SQL produced by an LLM, which is influenceable by prompt injection. When configured with a database role that has privileges enabling code execution or filesystem access (e.g., PostgreSQL pg_execute_server_program, MySQL FILE, MSSQL xp_cmdshell), an attacker who can shape the agent's input -- including indirectly via data returned to the LLM -- can coerce execution of dialect-specific primitives such as "COPY ... FROM PROGRAM", achieving RCE on the database host. Fixed in vX.Y by defaulting SQLChatAgent to a SELECT-only sqlglot-parsed statement allowlist with a dialect-aware dangerous-pattern blocklist; allow_dangerous_operations=True restores the previous unrestricted behavior for trusted deployments. This issue affects versions prior to 0.63.0.

  • LOW
  • NETWORK
  • HIGH
  • UNCHANGED
  • NONE
  • NONE
  • HIGH
  • HIGH

CWE-89 - SQL Injection

Structured Query Language (SQL) injection attacks are one of the most common types of vulnerabilities. They exploit weaknesses in vulnerable applications to gain unauthorized access to backend databases. This often occurs when an attacker enters unexpected SQL syntax in an input field. The resulting SQL statement behaves in the background in an unintended manner, which allows the possibility of unauthorized data retrieval, data modification, execution of database administration operations, and execution of commands on the operating system.

References

Advisory Timeline

  • Published