CVE-2026-15829 Google CVE debrief

Who should care

Defenders responsible for securing BigQuery forecasting tool deployments should assess exposure and apply patches. This includes administrators, security teams, and developers who use or manage the googleapis/mcp-toolbox.

Why it matters

CVE-2026-15829 is a high-severity vulnerability in the googleapis/mcp-toolbox that allows unauthorized users to bypass security boundaries and access sensitive BigQuery data. Defenders should assess exposure, apply patches, and verify allowedDatasets mechanism configuration.

• Potential unauthorized access to sensitive BigQuery data
• Possible disruption of BigQuery forecasting tool service
• Need for urgent patching and validation of allowedDatasets mechanism
• Potential for lateral movement to other datasets

Technical summary

The prebuilt BigQuery forecasting tool in googleapis/mcp-toolbox is vulnerable to SQL injection and security boundary bypass. The tool accepts client-controlled parameters as plain strings and interpolates them unescaped into a generated AI.FORECAST table-valued SELECT statement. This allows an unauthorized user to bypass the operator-configured allowedDatasets boundary and read arbitrary BigQuery tables.

Defensive priority

High priority for defenders to assess exposure and apply patches

Recommended defensive actions

• Assess exposure of BigQuery forecasting tool to unauthorized users
• Apply patches to update the tool to a version that validates and sanitizes user input
• Verify allowedDatasets mechanism is properly configured and enforced
• Monitor for suspicious activity and implement compensating controls
• Review compensating controls for exposed systems while remediation is scheduled and verified
• Check relevant monitoring, detection, and logs for exposed assets that need extra review
• Track exceptions, retest remediated assets, and close the item only after evidence is documented

Evidence notes

The vulnerability is caused by the lack of input validation and sanitization in the prebuilt BigQuery forecasting tool. The tool's allowedDatasets mechanism only validates the history_data parameter, but not the final assembled query.

Sources and references

Methodology and review provenance

AI-assisted synthesis based on stored public vulnerability evidence. System validation, approval state, and publication status do not by themselves establish human review of this revision. PatchSiren helps prioritize defensive review and does not prove exposure or remediation on any system.