The flaw allows an authenticated pgAdmin user to inject arbitrary SQL into the named restore point endpoint by submitting a crafted 'value' field. Because the input is splice‑directly into an SQL call using str.format, the attacker can execute any statement that the database role already permits. The injection does not cross a privilege boundary—the attacker already has full SQL rights for the connected role through the Query Tool—so the vulnerability merely duplicates an existing attack surface. The primary consequence is the potential for unintended data exposure or modification at the level of the authenticated role.

The issue affects pgAdmin 4 versions from 1.0 through the pre‑9.16 releases. The product is sold by pgadmin.org and is distributed under the open‑source pgAdmin 4 label.

The CVSS score of 5.3 reflects a moderate impact. EPSS data are not available and the vulnerability is not listed in the CISA KEV catalog, indicating no known exploitation reports. The attack vector is a web‑based POST request to /browser/server/restore_point/{gid}/{sid} and requires the user to be authenticated with a PostgreSQL session. Since the injected SQL runs under the user’s own database role, the likelihood of a catastrophic compromise is low, but an attacker can still leverage the endpoint to run arbitrary queries they could already run via the Query Tool.