Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-105039 Overview
CVE-2026-105039 is a missing authorization vulnerability [CWE-862] in EnergizeLab station-middleware Docker images built on NestJS. Versions 1.6.5 through 1.8.2 do not enforce role checks on administrative APIs. Any authenticated low-privilege user can call administrative endpoints to retrieve AI provider API keys, modify system configuration, and alter object-store contents. The flaw affects confidentiality and integrity across the deployment, and the scope change amplifies impact beyond the vulnerable component itself.
Critical Impact
Low-privilege authenticated users can exfiltrate AI provider secrets and tamper with configuration and object storage used by downstream robotics and station services.
Affected Products
- EnergizeLab station-middleware Docker image 1.6.5
- EnergizeLab station-middleware Docker images 1.6.6 through 1.8.1
- EnergizeLab station-middleware Docker image 1.8.2
Discovery Timeline
- 2026-10-15 - CVE-2026-105039 published to NVD
- 2026-10-16 - Last updated in NVD database
Technical Details for CVE-2026-105039
Vulnerability Analysis
The station-middleware service exposes NestJS controllers that handle administrative functions, including AI provider credential retrieval, configuration updates, and object-store management. These controllers authenticate the caller but do not verify that the caller holds an administrative role. The authorization layer is effectively absent, so possession of any valid user token is sufficient to reach privileged routes. The CVSS vector indicates a scope change, meaning the compromised component grants access to resources managed outside its own security authority, such as external object stores and third-party AI provider accounts.
Root Cause
The root cause is a missing authorization check [CWE-862] on administrative route handlers. NestJS supports guards and role decorators, but the affected versions do not apply a role guard to the sensitive endpoints. Authentication middleware runs, but no subsequent check compares the user's role against the required administrative role. As a result, the application fails open for authorization decisions.
Attack Vector
An attacker first obtains any valid low-privilege account on the middleware. The attacker then issues HTTP requests directly to the administrative endpoints responsible for credential disclosure, configuration mutation, or object-store writes. Because the server does not reject non-admin callers, the operations succeed. Harvested AI provider keys enable lateral abuse against external SaaS accounts and billing. Writes to the object store can replace firmware artifacts, model files, or operational configuration consumed by connected robotic stations. A public proof of concept is referenced in the GitHub PoC Repository and discussed in The Verge article on the Eiliko robot hack.
No verified exploitation code is reproduced here. See the TakeOnMe GCVE-1337 analysis for a technical walkthrough.
Detect Identity Attacks the Moment Authentication Fails
Singularity™ Identity surfaces credential misuse, lateral movement, and privilege escalation in real time, so you can contain the breach before it spreads.
Stop Identity AttacksDetection Methods for CVE-2026-105039
Indicators of Compromise
- Successful HTTP 200 responses to administrative routes (such as configuration, provider credential, or object-store endpoints) from user accounts that do not hold an administrative role.
- Outbound requests to AI provider APIs using credentials that originate from the station-middleware deployment but execute from unexpected source IPs.
- Unexpected PUT, POST, or DELETE operations against buckets or prefixes managed by the middleware's object-store integration.
Detection Strategies
- Enable verbose access logging on the NestJS application and alert on administrative route invocations grouped by caller role.
- Deploy a reverse proxy or Web Application Firewall rule that enforces role-based path restrictions independent of the application.
- Correlate authentication events with sensitive endpoint access to flag privilege mismatches.
Monitoring Recommendations
- Monitor AI provider dashboards for anomalous key usage, spend spikes, and new source geographies.
- Track object-store audit logs for writes outside normal CI/CD or operator workflows.
- Alert on new or modified Docker image tags in the 1.6.5 through 1.8.2 range still running in production.
How to Mitigate CVE-2026-105039
Immediate Actions Required
- Rotate all AI provider API keys, object-store credentials, and shared secrets accessible to the middleware.
- Restrict network reachability of administrative endpoints to trusted operator networks until a fixed image is deployed.
- Audit user accounts and revoke tokens that are not strictly required, reducing the pool of accounts that can reach the vulnerable routes.
Patch Information
No vendor patch URL is listed in the published CVE record. Operators should track EnergizeLab release notes and upgrade station-middleware to a version later than 1.8.2 once the vendor publishes a build that enforces role-based access control on administrative controllers. Validate the fix by confirming non-admin tokens receive HTTP 403 on the previously vulnerable routes.
Workarounds
- Place an authenticating reverse proxy in front of station-middleware and block administrative URL prefixes for non-admin groups.
- Remove or disable non-administrative user accounts on affected deployments until an upgrade is applied.
- Segment the middleware into a dedicated network zone with egress filtering to limit downstream abuse of stolen AI provider keys.
# Example NGINX snippet restricting admin paths to an operator subnet
location ~ ^/(admin|config|providers|objects)/ {
allow 10.20.30.0/24;
deny all;
proxy_pass http://station_middleware_upstream;
}
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

