CVE-2026-46373: SQLFluff DoS Vulnerability – Resource Exhaustion via Nested Queries

Technical summary

CVE-2026-46373 is a resource exhaustion vulnerability in SQLFluff's SQL parser triggered by deeply nested query structures. The vulnerability stems from inadequate handling of recursive parsing operations, allowing crafted input to consume excessive CPU and memory. The attack requires no authentication or user interaction, making it trivially exploitable in multi-tenant or user-facing SQL analysis platforms. SQLFluff versions prior to 4.1.0 are vulnerable; version 4.1.0 includes fixes for parser recursion limits and resource consumption.

Business impact

Organizations using SQLFluff to validate or lint user-supplied SQL queries face service disruption risks. SaaS platforms offering SQL formatting or linting as a feature, internal analytics tools, and CI/CD pipelines that accept arbitrary SQL are at risk of availability outages. The impact is limited to denial of service—no data breach or code execution occurs—but service unavailability can disrupt workflows and erode user trust. Organizations with strict uptime requirements should prioritize patching.

Affected systems

SQLFluff versions before 4.1.0 are vulnerable. This includes both direct SQLFluff installations and any application or service that embeds or depends on affected SQLFluff versions. Risk is highest in scenarios where the parser processes untrusted SQL input: web-based SQL formatters, linting-as-a-service offerings, multi-tenant analytics platforms, and automated CI/CD linting systems that accept user-contributed queries.

Exploitability

Exploitability is high. The attack requires only the ability to submit a SQL query to an affected system—no credentials, no user interaction, and no complex setup. A single crafted query with excessive nesting can trigger denial of service. Exploitation difficulty is low; an attacker needs only to understand basic SQL syntax and nesting mechanics. This makes the vulnerability practically exploitable by anyone with network access to a vulnerable application.

Remediation

Upgrade SQLFluff to version 4.1.0 or later. This update addresses parser resource limits and prevents excessive recursion from malicious input. For applications that cannot immediately patch, implement input validation and query complexity limits at the application level—for example, rejecting queries exceeding a reasonable nesting depth threshold or enforcing query size limits before passing them to the parser. Run-time resource limits (e.g., memory caps, CPU timeouts) can mitigate impact but do not eliminate the vulnerability.

Patch guidance

Verify your SQLFluff installation version using `sqlfluff --version`. Update to version 4.1.0 or later via your package manager (e.g., `pip install --upgrade sqlfluff` for Python environments). Test the patched version in a staging environment with existing SQL linting workflows to ensure compatibility. If you embed SQLFluff as a library, update your dependency declaration and rebuild/redeploy applications. Consult the official SQLFluff release notes to confirm patching has completed and no related regressions were introduced.

Detection guidance

Monitor for signs of parser-related resource exhaustion: CPU spikes correlated with SQL linting requests, memory growth that does not stabilize, and timeout errors in linting logs. Inspect application logs for queries with unusual nesting depth or excessively long query structures submitted shortly before resource exhaustion events. Network-based detection is limited; focus on application-level monitoring. If deploying SQLFluff behind a web service, track request latency and failure rates for linting endpoints. Implement request size and complexity limits and log rejections for forensic review.

Why prioritize this

Despite the 7.5 CVSS score, this vulnerability should be prioritized based on deployment context. If your organization exposes SQL linting to untrusted users or the internet, patch immediately—exploitability is trivial and impact is certain. Internal-only SQL linting deployments are lower priority but should still be addressed in a regular maintenance window. The lack of KEV status indicates this has not yet been widely exploited in the wild, providing a window to patch before attacker tooling matures.

Risk score, explained

The CVSS 3.1 score of 7.5 (HIGH) reflects high exploitability (network-accessible, no auth, no user interaction) balanced against impact limited to availability (no confidentiality or integrity loss). The score accurately captures the practical threat level for most organizations, but business context matters: a SaaS SQL formatter faces higher risk than an internal tool with restricted access. Organizations in regulated industries with strict availability mandates should treat this as critical.

Frequently asked questions