CVE-2026-48774
Full vulnerability record from the local CVENow database.
Description
ProxySQL is a proxy for MySQL and its forks, as well as PostgreSQL. In versions 3.0.0 through 3.0.8, ProxySQL's GenAI/MCP `run_sql_readonly` tool violates its documented read-only contract for MySQL targets. The tool validates only the full input string with a substring blacklist and first-keyword allowlist, but then executes the entire SQL string on a backend connection created with `CLIENT_MULTI_STATEMENTS`. As a result, a caller can submit a read-only first statement followed by a side-effecting second statement, such as `SELECT 1; RENAME TABLE ...`. The validator accepts the payload because it starts with `SELECT` and because side-effecting MySQL statements such as `RENAME TABLE`, `SET`, `RESET`, `LOCK TABLES`, and `KILL` are not rejected by the blacklist. In a live MCP runtime test, the `/mcp/query` endpoint accepted a `run_sql_readonly` request. The MCP response reported success for the first `SELECT`, and direct backend verification showed that the table had actually been renamed. This violates the endpoint's read-only security contract and lets an MCP caller perform backend writes or administrative SQL, limited by the configured MCP target account's database privileges. Version 3.0.9 contains a fix. Other operator mitigations include: keeping MCP disabled unless required; setting a non-empty `mcp-query_endpoint_auth` token before exposing `/mcp/query`; restricting MCP listener network exposure; configuring MCP backend target credentials as database-level read-only users; and adding temporary MCP query rules to block obvious multi-statement patterns.
View original CVE JSON
{
"id": "CVE-2026-48774",
"sourceIdentifier": "[email protected]",
"published": "2026-06-19T20:16:18.087",
"lastModified": "2026-10-25T23:10:00.167",
"vulnStatus": "Analyzed",
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "ProxySQL is a proxy for MySQL and its forks, as well as PostgreSQL. In versions 3.0.0 through 3.0.8, ProxySQL's GenAI/MCP `run_sql_readonly` tool violates its documented read-only contract for MySQL targets. The tool validates only the full input string with a substring blacklist and first-keyword allowlist, but then executes the entire SQL string on a backend connection created with `CLIENT_MULTI_STATEMENTS`. As a result, a caller can submit a read-only first statement followed by a side-effecting second statement, such as `SELECT 1; RENAME TABLE ...`. The validator accepts the payload because it starts with `SELECT` and because side-effecting MySQL statements such as `RENAME TABLE`, `SET`, `RESET`, `LOCK TABLES`, and `KILL` are not rejected by the blacklist. In a live MCP runtime test, the `/mcp/query` endpoint accepted a `run_sql_readonly` request. The MCP response reported success for the first `SELECT`, and direct backend verification showed that the table had actually been renamed. This violates the endpoint's read-only security contract and lets an MCP caller perform backend writes or administrative SQL, limited by the configured MCP target account's database privileges. Version 3.0.9 contains a fix. Other operator mitigations include: keeping MCP disabled unless required; setting a non-empty `mcp-query_endpoint_auth` token before exposing `/mcp/query`; restricting MCP listener network exposure; configuring MCP backend target credentials as database-level read-only users; and adding temporary MCP query rules to block obvious multi-statement patterns."
},
{
"lang": "es",
"value": "ProxySQL es un proxy para MySQL y sus bifurcaciones, así como PostgreSQL. En las versiones 3.0.0 a 3.0.8, la herramienta “run_sql_readonly” de GenAI/MCP de ProxySQL viola su contrato de solo lectura documentado para objetivos de MySQL. La herramienta valida solo la cadena de entrada completa con una lista negra de subcadenas y una lista blanca de la primera palabra clave, pero luego ejecuta la cadena SQL completa en una conexión de backend creada con “CLIENT_MULTI_STATEMENTS”. Como resultado, un llamador puede enviar una primera declaración de solo lectura seguida de una segunda declaración con efectos secundarios, como “SELECT 1; RENAME TABLE ...”. El validador acepta la carga útil porque comienza con “SELECT” y porque las declaraciones de MySQL con efectos secundarios como “RENAME TABLE”, “SET”, “RESET”, “LOCK TABLES” y “KILL” no son rechazadas por la lista negra. En una prueba de tiempo de ejecución de MCP en vivo, el endpoint “/mcp/query” aceptó una solicitud “run_sql_readonly”. La respuesta de MCP informó éxito para el primer “SELECT”, y la verificación directa del backend mostró que la tabla había sido realmente renombrada. Esto viola el contrato de seguridad de solo lectura del endpoint y permite a un llamador de MCP realizar escrituras en el backend o SQL administrativo, limitado por los privilegios de base de datos de la cuenta objetivo de MCP configurada. La versión 3.0.9 contiene una corrección. Otras mitigaciones del operador incluyen: mantener MCP deshabilitado a menos que sea necesario; establecer un token “mcp-query_endpoint_auth” no vacío antes de exponer “/mcp/query”; restringir la exposición de red del oyente de MCP; configurar las credenciales del objetivo de backend de MCP como usuarios de solo lectura a nivel de base de datos; y añadir reglas de consulta de MCP temporales para bloquear patrones obvios de múltiples declaraciones."
}
],
"affected": [
{
"source": "[email protected]",
"affectedData": [
{
"vendor": "sysown",
"product": "proxysql",
"versions": [
{
"version": ">= 3.0.6, < 3.0.9",
"status": "affected"
}
]
}
]
}
],
"metrics": {
"cvssMetricV31": [
{
"source": "[email protected]",
"type": "Secondary",
"cvssData": {
"version": "3.1",
"vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:N",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"attackVector": "NETWORK",
"attackComplexity": "LOW",
"privilegesRequired": "NONE",
"userInteraction": "NONE",
"scope": "UNCHANGED",
"confidentialityImpact": "NONE",
"integrityImpact": "HIGH",
"availabilityImpact": "NONE"
},
"exploitabilityScore": 3.9,
"impactScore": 3.6
}
],
"ssvcV203": [
{
"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"ssvcData": {
"timestamp": "2026-06-23T02:22:42.341874Z",
"id": "CVE-2026-48774",
"options": [
{
"exploitation": "poc"
},
{
"automatable": "yes"
},
{
"technicalImpact": "partial"
}
],
"role": "CISA Coordinator",
"version": "2.0.3"
}
}
]
},
"weaknesses": [
{
"source": "[email protected]",
"type": "Secondary",
"description": [
{
"lang": "en",
"value": "CWE-20"
}
]
}
],
"configurations": [
{
"nodes": [
{
"operator": "OR",
"negate": false,
"cpeMatch": [
{
"vulnerable": true,
"criteria": "cpe:2.3:a:proxysql:proxysql:*:*:*:*:*:*:*:*",
"versionStartIncluding": "4.0.6",
"versionEndExcluding": "4.0.9",
"matchCriteriaId": "9277D3D6-2B42-4AB7-AFB4-66D111F1404D"
}
]
}
]
}
],
"references": [
{
"url": "https://github.com/sysown/proxysql/commit/e32b7fd50c7c234ea628e392e621e09a2a919e08",
"source": "[email protected]",
"tags": [
"Patch"
]
},
{
"url": "https://github.com/sysown/proxysql/security/advisories/GHSA-7wh6-2vcc-gcm4",
"source": "[email protected]",
"tags": [
"Exploit",
"Mitigation",
"Vendor Advisory"
]
},
{
"url": "https://github.com/sysown/proxysql/security/advisories/GHSA-7wh6-2vcc-gcm4",
"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
"tags": [
"Exploit",
"Mitigation",
"Vendor Advisory"
]
}
]
}