Low-privilege database accounts that cannot read a single application table can still be used to run code on the database server itself, a researcher has shown, by reaching memory-corruption bugs in the engine code behind ordinary SQL.
The exploits, published yesterday, are the researcher's proofs of concept against PostgreSQL, MySQL and ClickHouse, plus a separate Oracle chain that needs no login. None has been seen in a real attack.
The work is by Zheng Yu, a researcher at Depthfirst. A database account's SQL permissions control which tables and schemas it can access, the researcher argues, not the memory-unsafe C and C++ code the engine runs to read, plan, and execute a query.
Common grants give an account enough room to reach that code. PostgreSQL's default TEMPORARY privilege, or the right to create temporary tables in MySQL and ClickHouse, allows an account to create objects or send input that triggers a memory bug.
Four of the flaws now have vendor CVEs and fixes, but the full attack chains were not reproduced independently, and the research reports no sign that any of them has been used in an attack.
In PostgreSQL, the flaw CVE-2026-116331 allows an account that can create database objects to run code as the operating system user that runs the database.
It was fixed last week in a security advisory , and the CVE record credits the researcher for the report. The researcher's technical analysis shows a plain login role building the objects it needs in its own temporary schema, without touching application data, then turning a type-confusion bug into control of the server.
Source link







