CVE-2026-17351: pgAdmin 4: AI Assistant read-only transaction bypass via sqlparse/PostgreSQL lexer disagreement (incomplete fix for CVE-2026-12045)
The fix for CVE-2026-12045 in pgAdmin 4 9.16 required the LLM-supplied query passed to the AI Assistant's executesqlquery tool to parse, via sqlparse, as exactly one non-transaction-control statement before running it inside a BEGIN TRANSACTION READ ONLY wrapper. sqlparse's string-literal lexing can disagree with PostgreSQL's own parser: under standardconformingstrings = on (PostgreSQL's default since 9.1), a backslash immediately before a quote is an ordinary character to PostgreSQL, but sqlparse treats it as escaping the quote. A payload such as SELECT '\';COMMIT;CREATE TABLE pwn(x int);SELECT 1 --' therefore parses as a single SELECT to sqlparse's validator, while PostgreSQL executes it as four statements: the smuggled COMMIT ends the wrapping read-only transaction, and the trailing ROLLBACK becomes a no-op. This reintroduces the same write/RCE bypass CVE-2026-12045 was meant to close, reachable via the same indirect prompt-injection delivery (an attacker plants the payload in any object the AI Assistant may read; the LLM emits it as a tool call).
An initial candidate fix ran the query with psycopg's execute(..., prepare=True), intending to force PostgreSQL's own Parse step (extended query protocol) to reject multi-statement text regardless of sqlparse's classification. This candidate fix does not work as submitted: psycopg3's PrepareManager silently ignores the prepare argument whenever the connection's preparethreshold is None, which is pgAdmin's default for every server connection (the per-server "Prepare threshold" field is blank unless an administrator explicitly sets it) -- psycopg3 falls back to the simple query protocol, the same multi-statement-capable path the bypass exploits, so the candidate fix closes nothing on any real-world default configuration.
The corrected fix sets conn.preparethreshold = 0 directly on the dedicated, single-use read-only connection the AI Assistant tool opens, structurally forcing the extended query protocol independent of any server-level configuration. Verified against a live PostgreSQL 18 instance: the payload executes successfully under the preparethreshold=None (default) behavior, and is rejected with "cannot insert multiple commands into a prepared statement" once preparethreshold=0 is set on that connection.
This issue affects pgAdmin 4: from 9.13 before 9.17.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Configuration
Set conn.prepare_threshold = 0 on the dedicated, single-use read-only connection used by the AI Assistant tool, so PostgreSQL uses the extended query protocol and rejects multi-statement text (fix verified: payload is rejected with "cannot insert multiple commands into a prepared statement" when prepare_threshold=0, and executes successfully when prepare_threshold=None).
pgAdmin 4 AI Assistant tool (execute_sql_query / PostgreSQL connection) prepare_threshold = 0
Event History
Frequently Asked Questions
What is the severity of CVE-2026-17351?
CVE-2026-17351 has a critical severity rating of 9.
How do I fix CVE-2026-17351?
Fixing CVE-2026-17351 involves upgrading to the latest version of pgAdmin 4 that includes the complete fix.
What is the impact of CVE-2026-17351?
CVE-2026-17351 allows a read-only transaction bypass, potentially leading to unauthorized access to sensitive data.
What software is affected by CVE-2026-17351?
CVE-2026-17351 affects pgAdmin 4, specifically related to its AI Assistant functionality.
Is CVE-2026-17351 related to CVE-2026-12045?
Yes, CVE-2026-17351 is an incomplete fix for the issue originally identified in CVE-2026-12045.