Summary
ArcadeDB contains a server-side request forgery vulnerability in its OpenCypher LOAD CSV implementation, which does not adequately validate HTTP and HTTPS destinations. An authenticated attacker who can submit queries may cause the database server to request internal services or cloud metadata endpoints and return sensitive responses.
Affected context
The vulnerability occurs in ArcadeDB deployments that expose OpenCypher query execution and support the LOAD CSV feature. It becomes risky when authenticated users can submit or influence queries and the database server has network access to internal services, administrative interfaces, or cloud metadata endpoints.
Exploit preconditions
- Attack surface: OpenCypher query execution through an ArcadeDB server
- Required access: Valid authentication and permission to execute queries using LOAD CSV
- Trigger point: Processing a LOAD CSV request containing an attacker-controlled HTTP or HTTPS URL
- Why exploitation is possible: Destination URLs are not sufficiently restricted, allowing the server to make requests to otherwise inaccessible internal or cloud-resident services
Why this happens
The underlying issue is insufficient validation and policy enforcement for server-side outbound requests. The LOAD CSV implementation treats user-influenced HTTP or HTTPS URLs as acceptable fetch targets without reliably restricting private, loopback, link-local, or cloud metadata destinations. This behavior maps to CWE-918: Server-Side Request Forgery, where an attacker causes a trusted server to access resources on the attacker's behalf.
Potential impact
- An attacker may retrieve sensitive data from internal HTTP services
- Cloud instance metadata or temporary credentials could be exposed
- Internal administrative endpoints could be probed or accessed
- Network segmentation controls could be bypassed from the database host
- Service availability could be affected through unwanted outbound requests
General mitigation guidance
Mitigation generally focuses on reducing trust in query-supplied URLs and applying explicit outbound request controls.
Common defensive patterns include:
- Restricting LOAD CSV destinations to an allowlisted set of approved hosts and schemes
- Blocking loopback, private, link-local, multicast, and cloud metadata address ranges
- Resolving hostnames and validating the resulting addresses before connecting
- Applying outbound network policies and egress filtering at the host or deployment layer
- Limiting query privileges so only trusted roles can use network-fetching features
- Returning only the minimum data required and avoiding unnecessary server-side URL retrieval
How teams detect issues like this
Organizations can identify SSRF exposure through:
- Static analysis and code review for network requests built from query or user-controlled values
- Review signals involving URL parsing without scheme, hostname, port, redirect, or resolved-address validation
- Tests covering private, loopback, link-local, and metadata-service destinations
- Monitoring database-originated outbound connections to unusual internal hosts or cloud metadata addresses
- Network policy enforcement that records or blocks unexpected egress from database workloads
- Automated security review tools that flag server-side fetch functionality lacking destination controls

