Summary
MOOS core through version 10.4.0 can expose an optional MOOSDB HTTP server that accepts variable-write requests without authenticating clients. An attacker who can reach the server port may modify MOOS variables, including actuator or override commands, potentially affecting connected robotic or autonomous systems.
Affected context
The vulnerability occurs when the optional MOOSDB HTTP server is enabled and reachable by an untrusted client. In affected deployments, the server may process requests that write MOOS variables without first verifying the client's identity or authorization, allowing network-reachable clients to influence application and control state.
Exploit preconditions
- Attack surface: The optional MOOSDB HTTP server and its listening network port
- Required access: Network connectivity to the HTTP server; valid credentials may not be required
- Trigger point: An HTTP request that asks the server to change a MOOS variable
- Why exploitation is possible: The request-processing path permits variable writes without enforcing authentication or authorization
Why this happens
The underlying issue is a missing authentication control on a state-changing HTTP interface. The server accepts requests that modify MOOS variables without establishing the identity of the caller or checking whether that caller is permitted to perform the requested operation. This behavior maps to CWE-306: Missing Authentication for Critical Function, because a security-sensitive function is exposed without an effective authentication requirement.
Potential impact
- An unauthenticated client may modify operational or configuration-related MOOS variables
- Unauthorized actuator or override commands could be issued
- Connected vehicles, robots, or automated processes could behave unexpectedly
- Safety, availability, or mission integrity could be degraded
- Attackers could cause data manipulation or support follow-on compromise of connected components
General mitigation guidance
Mitigation generally focuses on ensuring that every state-changing HTTP operation has an explicit authentication and authorization boundary.
Common defensive patterns include:
- Disabling the optional HTTP server when it is not required
- Restricting the service to trusted interfaces and isolated management networks
- Requiring strong authentication before accepting variable-write operations
- Applying per-operation authorization and least-privilege access controls
- Separating read-only monitoring interfaces from command and control interfaces
- Using network filtering, service identity controls, and encrypted transport where appropriate
- Treating actuator and override variables as high-impact operations requiring additional policy checks
How teams detect issues like this
Organizations can look for this weakness through:
- Static analysis and code review checks for HTTP handlers that mutate state without authentication middleware
- Tests confirming that unauthenticated requests cannot write MOOS variables
- Review of server binding, firewall, and network-segmentation policies
- Monitoring for unexpected variable changes, actuator commands, or override activity
- API security policy enforcement requiring authentication and authorization for all state-changing endpoints
- Automated security testing that inventories exposed services and evaluates access-control behavior

