Summary
Marten's full-text search APIs interpolated the user-supplied regConfig parameter directly into the generated SQL without parameterization or validation, making every code path that exposes regConfig to untrusted input a SQL injection sink.
Affected APIs
IQuerySession.SearchAsync<T>(string searchTerm, string regConfig, ...)IQuerySession.PlainTextSearchAsync<T>(...)IQuerySession.PhraseSearchAsync<T>(...)IQuerySession.WebStyleSearchAsync<T>(...)IQuerySession.PrefixSearchAsync<T>(...)IQueryable<T>.Where(x => x.Search(term, regConfig))and the matchingPlainTextSearch/PhraseSearch/WebStyleSearch/PrefixSearchextension methods
Details
In the affected versions, FullTextWhereFragment renders the WHERE-clause SQL by string interpolation:
private string Sql =>
$"to_tsvector('{_regConfig}'::regconfig, {_dataConfig}) @@ {_searchFunction}('{_regConfig}'::regconfig, ?)";
_regConfig arrives unchanged from the public API surface above. Any value containing a single quote terminates the SQL literal and lets an attacker append arbitrary PostgreSQL.
Confirmed exploit shapes (with regConfig set to attacker-controlled input)
| Goal | Payload |
|---|---|
| Time-based blind | english'::text); SELECT pg_sleep(5); -- |
| Information disclosure | english'; SELECT version(); -- |
| DDL execution | english'; DROP TABLE mt_doc_article; -- |
All five overloads listed above produced SQL containing the verbatim payload.
Workarounds
If users cannot upgrade immediately, do one of the following at the application boundary:
- Hard-code
regConfigto a compile-time constant ("english","simple", …) and never accept it from request input. - Validate any externally-sourced
regConfigvalue before passing it to Marten, e.g. against the same regex as the patch (^[a-zA-Z_][a-zA-Z0-9_]*(\.[a-zA-Z_][a-zA-Z0-9_]*)?$) or against an allowlist of PostgreSQL configurations the application actually uses. - Drop the
regConfigargument from the call site so Marten falls back to the safe default.
Resources
- Patch PR: JasperFx/marten#4343
- Patched file:
FullTextWhereFragment.cs - Regression tests:
full_text_regconfig_sql_injection.cs - CWE-89: https://cwe.mitre.org/data/definitions/89.html
Credit
Reported privately to the JasperFx team with a working proof of concept covering all five affected overloads.
Impact
- Confidentiality: an attacker can append arbitrary
SELECTstatements and exfiltrate database contents through error channels, response timing, or, if the application surfaces query results, directly. - Integrity / Availability: DDL,
UPDATE,DELETE, andpg_sleep-style denial-of-service payloads succeed under the same vector. Concrete impact depends on the database role used by the Marten connection string. - Precondition: the calling application must forward attacker-controlled input into the
regConfigparameter (e.g. a?lang=query string mapped toregConfig). Applications that hard-coderegConfigto a compile-time constant are not exploitable.
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-2026-45288 has a CVSS score of 9.8 (Critical). The vector is network-reachable, no 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 (8.37.0); 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
Fixed in Marten 8.36.1 (and forward) by #4343.
FullTextWhereFragment now validates regConfig against ^[a-zA-Z_][a-zA-Z0-9_]*(\.[a-zA-Z_][a-zA-Z0-9_]*)?$ (a simple PostgreSQL identifier, optionally schema-qualified, capped at NAMEDATALEN-1 per side) and throws ArgumentException for anything else. The default value ("english"), schema-qualified configs ("pg_catalog.english"), and the standard PostgreSQL text-search configurations all continue to work.
Frequently Asked Questions
- What is CVE-2026-45288? CVE-2026-45288 is a critical-severity SQL injection vulnerability in Marten (nuget), affecting versions <= 8.36. It is fixed in 8.37.0. 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-2026-45288? CVE-2026-45288 has a CVSS score of 9.8 (Critical). 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 versions of Marten are affected by CVE-2026-45288? Marten (nuget) versions <= 8.36 is affected.
- Is there a fix for CVE-2026-45288? Yes. CVE-2026-45288 is fixed in 8.37.0. Upgrade to this version or later.
- Is CVE-2026-45288 exploitable, and should I be worried? Whether CVE-2026-45288 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-2026-45288 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-2026-45288? Upgrade
Martento 8.37.0 or later.