Summary
MikroORM's identifier-quoting helper (Platform.quoteIdentifier and the postgres/mssql overrides) and its JSON-path emitters (Platform.getSearchJsonPropertyKey, quoteJsonKey) did not properly escape characters that delimit the SQL identifier or string-literal context they emit into. When application code passes attacker-influenced strings to public ORM APIs that expect an identifier or a JSON-property filter, an attacker can break out of the quoted context and inject arbitrary SQL.
Affected APIs
The vulnerability is reachable when application code passes an attacker-influenced string to any of the following documented APIs:
- Multi-tenant
schemaoption,em.fork({ schema }),qb.withSchema(name),wrap(entity).setSchema(name),em.create(Cls, data, { schema }). The schema name is concatenated into the SQL identifier and never had its dialect quote character escaped. em.find/qb.whereJSON-property filters,em.find(Entity, { jsonCol: { [userKey]: value } }). The user-supplied JSON sub-keys cannot be validated against any metadata (JSON columns are schemaless by design), and were spliced into the SQL string literal of the JSON path expression without escaping.qb.where/qb.orderBy/qb.groupBy/qb.having/qb.selectkeys, keys containing.or::bypassed the structured-where metadata validator inCriteriaNode, then flowed through the same brokenquoteIdentifier. Apps that forwarded raw filter keys from request input were already broken on authorization grounds (e.g.{ isAdmin: true }); the SQL injection here is a defence-in-depth failure on top of that.
The vulnerability does not affect documented escape-hatch APIs (raw(), the sql tagged template, qb.raw(), em.raw()), those are documented as accepting raw SQL and are the application's responsibility to sanitize.
Affected dialects
All SQL dialects supported by MikroORM. The identifier-quoting bug exists in every dialect's quoteIdentifier (the dialect's own quote character, backtick, double quote, or right bracket, was not doubled when embedded in the identifier). The JSON-path bug exists in all dialects' getSearchJsonPropertyKey/quoteJsonKey.
MongoDB driver is not affected (no SQL).
Workarounds
If users cannot upgrade immediately:
- For multi-tenant apps using
em.fork({ schema })/wrap().setSchema()/qb.withSchema(): validate the schema name against a strict allowlist (e.g.^[A-Za-z_][\w$]*$) before passing it to MikroORM. - For applications that pass
where/orderByfilters from request input: validate every key against the entity's known properties before constructing the filter; do not pass keys containing.or::from user input. - For applications that allow filtering on JSON columns from request input: validate every JSON sub-key against an allowlist (or against
^[a-zA-Z_][\w]*$) before passing it toem.find.
Credits
Reported and patched by Martin Adámek (project maintainer) during an internal security review.
Impact
- Confidentiality, read from any table/schema the database user has access to (cross-tenant data leak in multi-tenant deployments).
- Integrity, on dialects supporting multi-statement queries (MSSQL, MySQL with multi-statement enabled), execute additional arbitrary SQL statements (data modification, in-database privilege escalation).
- Availability, DROP/TRUNCATE via injected statements where the database user has the privilege.
Untrusted input alters a database query, allowing the attacker to read or modify data the query was not intended to access. Typical impact: data disclosure or modification.
CVE-2026-44680 has a CVSS score of 7.6 (High). The vector is network-reachable, low privileges required, and no user interaction. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. A fixed version is available (7.0.14, 6.6.14); upgrading removes the vulnerable code path.
Affected versions
Security releases
Kodem intelligence
Severity tells you how bad this could be in the worst case. It does not tell you whether you are exposed. Exploitability and impact are functions of runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A vulnerable package can sit in your dependency tree and never run.
Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter. Kodem's runtime-powered SCA identifies whether this CVE is reachable in your applications.
Remediation advice
- v7: upgrade to 7.0.14 or later.
- v6: upgrade to 6.6.14 or later.
Patches:
- Identifier quoting: #7653 (master) / #7654 (6.x)
- JSON-path keys: #7656 (master) / #7657 (6.x)
Frequently Asked Questions
- What is CVE-2026-44680? CVE-2026-44680 is a high-severity SQL injection vulnerability in @mikro-orm/sql (npm), affecting versions <= 7.0.13. It is fixed in 7.0.14, 6.6.14. Untrusted input alters a database query, allowing the attacker to read or modify data the query was not intended to access.
- How severe is CVE-2026-44680? CVE-2026-44680 has a CVSS score of 7.6 (High). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
- Which packages are affected by CVE-2026-44680?
@mikro-orm/sql(npm) (versions <= 7.0.13)@mikro-orm/knex(npm) (versions <= 6.6.13)
- Is there a fix for CVE-2026-44680? Yes. CVE-2026-44680 is fixed in 7.0.14, 6.6.14. Upgrade to this version or later.
- Is CVE-2026-44680 exploitable, and should I be worried? Whether CVE-2026-44680 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
- What actually determines whether CVE-2026-44680 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
- How do I fix CVE-2026-44680?
- Upgrade
@mikro-orm/sqlto 7.0.14 or later - Upgrade
@mikro-orm/knexto 6.6.14 or later
- Upgrade