GHSA-QQF5-X7MJ-V43P

Vulnerability from github – Published: 2026-06-18 17:24 – Updated: 2026-06-18 17:24
VLAI
Summary
budibase: Database Connector SQL Injections in PostgreSQL, MS SQL, and MySQL
Details

Summary

This advisory covers three distinct SQL Injection vulnerabilities within Budibase's database connectors (PostgreSQL, Microsoft SQL Server, and MySQL). Because user-controlled schema and table configurations are interpolated directly into raw SQL queries without proper escaping or parameterization during database introspection, an authenticated administrator can break out of string delimiters. This allows for arbitrary DDL/DML execution, database compromise, and potential underlying OS command execution (e.g., via MS SQL xp_cmdshell).

Details

Vulnerability Type & Title

PostgreSQL SET search_path SQL Injection

Description & Root Cause

The schema datasource config field is interpolated directly into a raw SQL statement without proper escaping. Double quotes inside the schema name are not escaped, allowing an attacker to break out of the string literal and inject arbitrary SQL.

Vulnerable Code: File: packages/server/src/integrations/postgres.ts, lines 355–358

const search_path = this.config.schema
  .split(",")
  .map(item => `"${item.trim()}"`)                // NO escaping of embedded "
await this.client.query(`SET search_path TO ${search_path.join(",")};`)

node-postgres sends this via the simple query protocol, which supports multi-statement execution with semicolons.

Step-by-Step Reproduction

  1. Edit a PostgreSQL datasource configuration.
  2. Set the schema field to: public"; CREATE TABLE pwned AS SELECT usename, passwd FROM pg_shadow; --
  3. Save or trigger a connection test.
  4. The query executes as: SET search_path TO "public"; CREATE TABLE pwned AS SELECT usename, passwd FROM pg_shadow; --;
  5. PostgreSQL executes both statements.

Impact

  • Full database compromise. The attacker can read pg_shadow hashes, call pg_read_file(), or execute any DDL/DML.

Vulnerability Type & Title

Microsoft SQL Server Schema Introspection SQL Injection

Description & Root Cause

Three methods used during schema introspection (buildSchema) interpolate user-controlled values directly into SQL strings using single-quote delimiters with no escaping.

Vulnerable Code: File: packages/server/src/integrations/microsoftSqlServer.ts, lines 388–414

getDefinitionSQL(tableName: string, schemaName: string) {
  return `select * from INFORMATION_SCHEMA.COLUMNS
          where TABLE_NAME='${tableName}' AND TABLE_SCHEMA='${schemaName}'`
}

schemaName comes directly from this.config.schema (user config).

Step-by-Step Reproduction

  1. Edit an MS SQL Server datasource configuration.
  2. Set the schema field to: dbo'; EXEC xp_cmdshell('whoami'); --
  3. Trigger schema introspection (fetch tables).
  4. The OS command executes on the SQL server if xp_cmdshell is enabled.

Impact

  • Arbitrary SQL execution, potentially leading to OS command execution via xp_cmdshell.

Vulnerability Type & Title

MySQL multipleStatements: true + DESCRIBE Backtick Injection

Description & Root Cause

The MySQL integration enables multipleStatements: true, allowing semicolon-separated multi-statement execution. When introspecting tables, table names are interpolated into a DESCRIBE query wrapped in backticks, but the backticks are not escaped.

Vulnerable Code: File: packages/server/src/integrations/mysql.ts, lines 172, 305

this.config = { ...config, multipleStatements: true, ... }  // line 172
...
{ sql: `DESCRIBE \`${tableName}\`;` }  // line 305 — backtick NOT escaped

Step-by-Step Reproduction

  1. An attacker (or malicious database user) creates a table named foo`; DROP TABLE users; --.
  2. In Budibase, an admin triggers schema introspection for the database.
  3. Budibase reads the malicious table name from INFORMATION_SCHEMA.TABLES and inserts it into the DESCRIBE query.
  4. The backtick breaks out, and the secondary DROP TABLE payload executes.

Impact

  • Arbitrary SQL execution triggered during schema discovery. Requires prior database catalog manipulation.
Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "budibase"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "3.39.19"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [],
  "database_specific": {
    "cwe_ids": [
      "CWE-89"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-18T17:24:47Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "### Summary\nThis advisory covers three distinct SQL Injection vulnerabilities within Budibase\u0027s database connectors (PostgreSQL, Microsoft SQL Server, and MySQL). Because user-controlled schema and table configurations are interpolated directly into raw SQL queries without proper escaping or parameterization during database introspection, an authenticated administrator can break out of string delimiters. This allows for arbitrary DDL/DML execution, database compromise, and potential underlying OS command execution (e.g., via MS SQL `xp_cmdshell`).\n\n### Details\n\n### Vulnerability Type \u0026 Title\n**PostgreSQL `SET search_path` SQL Injection**\n\n### Description \u0026 Root Cause\nThe `schema` datasource config field is interpolated directly into a raw SQL statement without proper escaping. Double quotes inside the schema name are not escaped, allowing an attacker to break out of the string literal and inject arbitrary SQL.\n\n**Vulnerable Code:**\n**File:** `packages/server/src/integrations/postgres.ts`, lines 355\u2013358\n\n```typescript\nconst search_path = this.config.schema\n  .split(\",\")\n  .map(item =\u003e `\"${item.trim()}\"`)                // NO escaping of embedded \"\nawait this.client.query(`SET search_path TO ${search_path.join(\",\")};`)\n```\n\n`node-postgres` sends this via the **simple query protocol**, which supports multi-statement execution with semicolons.\n\n### Step-by-Step Reproduction\n1. Edit a PostgreSQL datasource configuration.\n2. Set the schema field to:\n   `public\"; CREATE TABLE pwned AS SELECT usename, passwd FROM pg_shadow; --`\n3. Save or trigger a connection test.\n4. The query executes as:\n   `SET search_path TO \"public\"; CREATE TABLE pwned AS SELECT usename, passwd FROM pg_shadow; --;`\n5. PostgreSQL executes both statements.\n\n### Impact\n- Full database compromise. The attacker can read `pg_shadow` hashes, call `pg_read_file()`, or execute any DDL/DML.\n\n---\n\n### Vulnerability Type \u0026 Title\n**Microsoft SQL Server Schema Introspection SQL Injection**\n\n### Description \u0026 Root Cause\nThree methods used during schema introspection (`buildSchema`) interpolate user-controlled values directly into SQL strings using single-quote delimiters with no escaping.\n\n**Vulnerable Code:**\n**File:** `packages/server/src/integrations/microsoftSqlServer.ts`, lines 388\u2013414\n\n```typescript\ngetDefinitionSQL(tableName: string, schemaName: string) {\n  return `select * from INFORMATION_SCHEMA.COLUMNS\n          where TABLE_NAME=\u0027${tableName}\u0027 AND TABLE_SCHEMA=\u0027${schemaName}\u0027`\n}\n```\n\n`schemaName` comes directly from `this.config.schema` (user config).\n\n### Step-by-Step Reproduction\n1. Edit an MS SQL Server datasource configuration.\n2. Set the schema field to: `dbo\u0027; EXEC xp_cmdshell(\u0027whoami\u0027); --`\n3. Trigger schema introspection (fetch tables).\n4. The OS command executes on the SQL server if `xp_cmdshell` is enabled.\n\n### Impact\n- Arbitrary SQL execution, potentially leading to OS command execution via `xp_cmdshell`.\n\n---\n\n### Vulnerability Type \u0026 Title\n**MySQL `multipleStatements: true` + `DESCRIBE` Backtick Injection**\n\n### Description \u0026 Root Cause\nThe MySQL integration enables `multipleStatements: true`, allowing semicolon-separated multi-statement execution. When introspecting tables, table names are interpolated into a `DESCRIBE` query wrapped in backticks, but the backticks are not escaped.\n\n**Vulnerable Code:**\n**File:** `packages/server/src/integrations/mysql.ts`, lines 172, 305\n\n```typescript\nthis.config = { ...config, multipleStatements: true, ... }  // line 172\n...\n{ sql: `DESCRIBE \\`${tableName}\\`;` }  // line 305 \u2014 backtick NOT escaped\n```\n\n### Step-by-Step Reproduction\n1. An attacker (or malicious database user) creates a table named ``foo`; DROP TABLE users; --``.\n2. In Budibase, an admin triggers schema introspection for the database.\n3. Budibase reads the malicious table name from `INFORMATION_SCHEMA.TABLES` and inserts it into the `DESCRIBE` query.\n4. The backtick breaks out, and the secondary `DROP TABLE` payload executes.\n\n### Impact\n- Arbitrary SQL execution triggered during schema discovery. Requires prior database catalog manipulation.",
  "id": "GHSA-qqf5-x7mj-v43p",
  "modified": "2026-06-18T17:24:47Z",
  "published": "2026-06-18T17:24:47Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/Budibase/budibase/security/advisories/GHSA-qqf5-x7mj-v43p"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/Budibase/budibase"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:H/UI:R/S:C/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "budibase: Database Connector SQL Injections in PostgreSQL, MS SQL, and MySQL"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…