CVE-2026-13751: Snowflake CLI Server-Side Request Forgery via Remote SQL References
CVE-2026-13751 is a server-side request forgery weakness (CWE-829) in Snowflake CLI versions from 1.1.0 before 3.19. The SQL statement reader's !source and !load directives can fetch remote URLs without sufficient restriction on the destination, and the record says Snowflake fixed it in version 3.19. NVD rates it 9.6 (critical) under CVSS 3.1, but the record contains no confirmed exploitation.
Overview
The advisory describes a server-side request forgery weakness in Snowflake CLI. The SQL statement reader's !source and !load directives can reference remote URLs that are retrieved at runtime, and the record says this retrieval was not restricted enough by request destination. The weakness is classified as CWE-829 (inclusion of functionality from untrusted control sphere).
Impact
- The record states that an attacker could cause unintended outbound requests toward internal or non-public network locations.
- The record states that remote SQL content could be retrieved and executed in the context of the victim user's session.
- The record states that exploitation requires the victim to process attacker-controlled content through a vulnerable command path, and that impact is limited by the privileges of that session and environment.
- NVD lists a CVSS 3.1 base score of 9.6 (critical). The record also lists a second CVSS 3.1 score of 4.1 from the CNA, and does not explain the difference.
Affected products and versions
- Product: Snowflake Snowflake Cli
- Affected range in the record: from 1.1.0 and below 3.19.0
- The record includes no separate fixed-version field, but its description states that the fix is available in version 3.19.
Exploitation status
- The CISA KEV catalog does not list this CVE in the record.
- No confirmed exploitation in the wild appears in the sources consulted for this analysis.
- EPSS is 0.00204 (percentile 0.09683), dated 2026-10-13. This is a statistical estimate of exploitation likelihood, not evidence of observed attacks.
Remediation and mitigation
- Upgrade Snowflake CLI to version 3.19 or later, which the record identifies as the release containing the fix.
- Review the Snowflake vendor advisory for the option that disables remote URL retrieval, and apply it where upgrading is not immediately possible.
- Until upgrading, limit the SQL content processed by the CLI to trusted sources and avoid running scripts from untrusted origins.
- Restrict outbound network access from systems running the CLI, reducing the reach of any unintended request.
Detection guidance
- Inventory hosts and automation pipelines that run Snowflake CLI and record their installed versions.
- Monitor for unexpected outbound connections from CLI runtime environments, especially to internal address ranges or metadata services.
- Review logs of SQL file processing for !source or !load directives that reference unfamiliar remote locations.
Limitations of the available information
- The record does not give a detailed description of the vulnerable command path, so its scope beyond the stated directives is unknown.
- The record does not provide the CVSS vector for the 4.1 score, nor an explanation of why NVD and the CNA score differ.
- The record does not name the specific fixed build beyond version 3.19, and the fixes list is empty.
- Sources consulted for this analysis do not report confirmed exploitation.
Frequently asked questions
Which Snowflake CLI versions are affected by CVE-2026-13751?
The record lists versions from 1.1.0 up to, but not including, 3.19.0 as affected. The description states that version 3.19 contains the fix.
Is CVE-2026-13751 known to be actively exploited?
No. The record shows the CVE is not listed in the CISA KEV catalog, and no confirmed exploitation appears in the sources consulted.
What is the recommended remediation?
Upgrade Snowflake CLI to version 3.19 or later. Where that is not immediately possible, use the remote URL retrieval option described in the vendor advisory and restrict outbound network access from affected systems.
This analysis was drafted with AI assistance from verified source data and automatically checked against it.