What is the CVSS score of CVE-2026-46373?

CVE-2026-46373 has a CVSS 3.1 base score of 7.5 (High). CVSS vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H. EPSS exploitation probability: 0.0%.

Which versions of SQLFluff are affected by CVE-2026-46373?

SQLFluff, a Python SQL linting tool widely used in development pipelines, contains a remote denial-of-service vulnerability affecting versions below 4.1.0.. A patch is available.

How to fix or mitigate CVE-2026-46373?

Within 24 hours: Inventory all systems running SQLFluff below version 4.1.0 and assess exposure (internal vs. external-facing). Within 7 days: Implement input validation to enforce maximum SQL nesting depth, apply rate limiting to SQLFluff submission endpoints, and restrict network access to trusted sources. Within 30 days: Subscribe to SQLFluff vendor advisories for patch availability and prepare an upgrade procedure.

SQLFluff CVE-2026-46373

HIGH

Uncontrolled Recursion (CWE-674)

2026-05-19 https://github.com/sqlfluff/sqlfluff

GHSA-wmhf-fqc8-vxhh

Denial Of Service

7.5

CVSS 3.1 · Vendor: https://github.com/sqlfluff/sqlfluff

Severity by source

Vendor (https://github.com/sqlfluff/sqlfluff) PRIMARY

7.5 HIGH

AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Primary rating from Vendor (https://github.com/sqlfluff/sqlfluff) · only source for this CVE.

CVSS VectorVendor: https://github.com/sqlfluff/sqlfluff

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Attack Vector

Network

Attack Complexity

Low

Privileges Required

None

User Interaction

None

Scope

Unchanged

Confidentiality

None

Integrity

None

Availability

High

Lifecycle Timeline

Source Code Evidence Fetched

May 19, 2026 - 20:31 vuln.today

Analysis Generated

May 19, 2026 - 20:31 vuln.today

Blast Radius

ecosystem impact

† from your stack dependencies † transitive graph · vuln.today resolves 4-path depth

4 pypi packages depend on sqlfluff (4 direct, 0 indirect)

Ecosystem-wide dependent count for version 4.1.0.

DescriptionCVE.org

Impact

In deployments where untrusted users can provide SQL queries to be linted, an untrusted user can submit a malicious query with deliberate excessive nesting to any application using the parser to trigger a Denial of Service through resource exhaustion.

Patches

Versions 4.1.0 and up contain a configurable recursion limit, which is enabled by default, to prevent this manner of exploit.

Credit

Ori Nakar from Imperva Threat Research Team.

AnalysisAI

Denial of service in SQLFluff (Python SQL linter/parser) below version 4.1.0 allows remote attackers to exhaust server resources by submitting SQL queries with deliberately excessive nesting, triggering uncontrolled recursion in the parser. The flaw (CWE-674) affects any application that accepts untrusted SQL input for linting and carries a CVSS 7.5 (AV:N/AC:L/PR:N/UI:N/A:H); no public exploit identified at time of analysis and EPSS data was not provided.

Technical ContextAI

SQLFluff is a popular open-source SQL linter and parser distributed via PyPI (pkg:pip/sqlfluff) and used in CI pipelines, IDE plugins, and hosted linting services. The underlying weakness is CWE-674 (Uncontrolled Recursion): the parser's grammar walks deeply nested SQL constructs recursively without a hard depth limit, so a query with deliberately deep nesting causes the Python call stack and CPU/memory to be consumed until the process crashes or becomes unresponsive. Because SQLFluff is often embedded as a library inside web services or developer-platform integrations that expose 'lint this query' endpoints, the recursion issue becomes a network-reachable DoS rather than a purely local one.

RemediationAI

Vendor-released patch: SQLFluff 4.1.0. Upgrade the sqlfluff PyPI package to 4.1.0 or later (for example, pip install --upgrade 'sqlfluff>=4.1.0') in any application, CI pipeline, or service that lints SQL, and pin the minimum version in requirements files and lockfiles to prevent regressions. Version 4.1.0 introduces a configurable recursion limit that is enabled by default; review the configuration if your workloads include legitimate deeply-nested SQL and tune the limit rather than disabling it. If immediate upgrade is impossible for services exposing linting to untrusted users, compensating controls include rejecting or truncating SQL inputs above a conservative byte/character size, capping parenthesis or keyword nesting depth at the application layer before invoking the parser, running the linter in an isolated subprocess with strict CPU and memory ulimits (so a crash does not affect the parent service), and rate-limiting the lint endpoint per source - each of these reduces availability risk but may reject some legitimate complex queries. Full advisory: https://github.com/sqlfluff/sqlfluff/security/advisories/GHSA-wmhf-fqc8-vxhh.