Critical RCE Vulnerability in MindsDB Exposes Sensitive Data via Unauthenticated API Endpoints
Overview
A critical unauthenticated remote code execution (RCE) vulnerability in MindsDB Minds Platform 26.1.0 and earlier allows attackers to run arbitrary OS commands, granting access to SSH keys, credentials, and environment secrets with no authentication required. This flaw stems from improperly secured API endpoints that bypass sandboxing protections.
Technical Analysis
The vulnerability arises from two unauthenticated endpoints: PUT /api/v1/settings/ enables attackers to inject a controlled LLM API key, and POST /api/v1/responses/ allows crafted prompts to trigger the Anton agent’s scratchpad tool, which executes arbitrary Python code via exec(). The lack of input validation and sandboxing creates a direct path for privilege escalation and system compromise.
Enterprise & DIB Impact
Defense Industrial Base (DIB) and enterprise systems using MindsDB face severe risks, including unauthorized access to classified data, potential insider threat scenarios, and disruption of mission-critical workflows. Exploitation could bypass multi-factor authentication and compromise air-gapped environments through lateral movement.
Recommended Actions
- Verify affected MindsDB versions and apply updates immediately
- disable or restrict access to /api/v1/settings/ and /api/v1/responses/ endpoints
- implement network segmentation for internal agents
- monitor for anomalous API usage patterns
- and enforce least-privilege execution environments for agent tools
Need Help Assessing Your Exposure?
Axiom Cyber Research provides vulnerability assessment, red team operations, and security advisory services to the Defense Industrial Base and regulated sectors.
Schedule a ConsultationFull security advisory on 247alerts.net →