Remote SQL injection in drogonframework drogon ORM

Note: This data is for reference and cybersecurity research purposes only.CyStack advises users not to use this information for unlawful purposes.

What is CVE-2026-94144?

CVE-2026-94144 is a vulnerability classified as Improper Neutralization of Special Elements used in an SQL Command and Improper Neutralization of Special Elements in Output Used by a Downstream Component, affecting drogon (affected versions: 1.9.0, 1.9.1, and other affected versions). This vulnerability is rated Medium, with a CVSS score of 6.9. Public exploit code or evidence is available for this vulnerability, but that does not confirm exploitation in the wild.

Overview

Original source data

A flaw has been found in drogonframework drogon up to 1.9.13. This affects the function makeCriteria in the library orm_lib/src/Criteria.cc of the component ORM. Executing a manipulation of the argument filter can lead to sql injection. The attack may be performed from remote. The exploit has been published and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

Affected products and scope

  • The CNA record confirms that drogonframework drogon versions 1.9.0 through 1.9.13 are affected.
  • A public technical report broadens the vulnerable-pattern claim from v1.0.0-beta8 through v1.9.13 and states that later versions should be treated as potentially affected until a fix is confirmed. This is a researcher report and is not reflected in the CNA's structured affected-version list.
  • No confirmed fixed-version boundary is identified. Do not assume that a version after 1.9.13 is safe without a confirmed fix statement.

Technical details

The flaw is in orm_lib/src/Criteria.cc, in Criteria::Criteria(const Json::Value &json) noexcept(false), which is reached by makeCriteria when ORM criteria are processed. The public technical report states that generated RESTful controllers can parse a JSON filter field and pass it through RestfulController to this constructor. Column names may be checked or masqueraded, and values are parameterized, but the operator element is appended to the SQL expression without an allowlist. This is a CWE-89 condition in which attacker-controlled input can alter part of the SQL predicate. Exposure requires an application to accept externally controlled filter data and pass it to the vulnerable ORM path. The exact set of exposed applications cannot be determined from the library version alone because the application must also use the affected RESTful or equivalent call path.

Exploitability

The issue is remotely reachable; the normalized record states that authentication and user interaction are not required and that attack complexity is low. In practice, the service must expose an endpoint that accepts attacker-controlled JSON filter data and routes it into the affected ORM code. A public proof of concept has been published, and the CNA source identifies the exploitation status as proof of concept. This establishes that public exploit code exists, but it does not establish widespread exploitation in the wild. Using drogon without the affected filter-processing path is not by itself evidence that an application is exposed.

Technical impact

The flaw can allow an attacker to bypass WHERE conditions, alter the returned row set, and in some deployments weaken application-level row authorization. If the service's database account has broad privileges, controlled SQL predicate changes could lead to data disclosure or modification; the actual result depends on those privileges, the generated query, and the database dialect. The issue does not by itself establish server-side code execution or compromise of other systems. Exploitability is also limited by whether an endpoint is publicly reachable and whether the deployment uses the affected filter-processing path.

Business impact

  • Public APIs using the affected filter path may return data outside the requesting user's intended scope.
  • Bypassing or changing WHERE conditions can weaken row-level authorization and cause applications to return more records than intended.
  • Confidentiality, integrity, and availability consequences depend on the database account's permissions, the generated query, and the endpoint's response behavior.
  • Because a public proof of concept exists, organizations should prioritize version and endpoint inventory, but the available evidence does not identify a specific victim or breach.

Remediation

  1. Deploy a vendor-confirmed fixed release when one is published. The available evidence does not identify a specific patch or fixed version.
  2. Until a fix is available, disable or restrict endpoints that accept untrusted JSON filters where the functionality is not required; in particular, avoid exposing those endpoints publicly without access controls.
  3. If the functionality must remain enabled, enforce a strict allowlist for filter operators before SQL construction, continue parameterizing values, and review the complete path from JSON filter through RestfulController to Criteria.
  4. Apply appropriate authentication and authorization to generated REST endpoints and limit the database account to the service's required privileges.
  5. Do not assume that releases after 1.9.13 are fixed until a release note or technical advisory confirms the affected branch and the correction.

Detection

  1. Inventory drogon deployments and identify systems using versions from 1.9.0 through 1.9.13; also review older branches reported publicly to contain the same vulnerable pattern.
  2. Inspect application code and drogon_ctl generated code for RESTful controllers that accept a JSON filter field and pass it to makeCriteria.
  3. Review orm_lib/src/Criteria.cc, RestfulController, and the RESTful templates to determine whether filter operators are checked against an allowlist before SQL construction.
  4. Identify filter endpoints reachable from the Internet or other untrusted networks, especially endpoints without appropriate authentication and authorization.
  5. Review HTTP and database telemetry for unexpected SQL-like operator content in filter fields or unusual query results. This is precautionary monitoring, not a confirmed IOC; absence of suspicious logs does not prove safety.
Sources (16)
Learn more

Run an in-depth assessment with complete web risk management

CyStack VulnScan continuously discovers assets, validates vulnerabilities, and helps security teams prioritize remediation across the organization.