/

CVE-2026-93597

Severity:

HIGH RISK

(7.7)

SSRF (CWE-918) in ArcadeDB IPv6 Transition Guard Logic

CVE-2026-93597

Deferred

Vulnerability type

Category:

Server-Side Request Forgery

CWE:

CWE-918

Impact:

Internal Network and Cloud Metadata Access

Severity:

HIGH RISK

(7.7)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A server-side request forgery vulnerability in ArcadeDB allows authenticated attackers to bypass IPv4-based SSRF protections using IPv6 transition addresses that contain private or loopback IPv4 values. Successful abuse may provide access to internal services, cloud metadata endpoints, or other network resources reachable from the ArcadeDB server.

Affected context

The issue can occur in ArcadeDB deployments where the IMPORT DATABASE feature or server commands accept attacker-influenced URLs and pass them through an SSRF guard that does not fully recognize NAT64, 6to4, or Teredo address representations. The risk is greater when the database server has network access to private services or instance metadata endpoints.

Exploit preconditions

  • Attack surface: URL-handling logic in IMPORT DATABASE and server commands - Required access: Authenticated access sufficient to invoke the affected

commands or features

  • Trigger point: Server-side retrieval of an attacker-supplied URL - Network condition: The ArcadeDB host must be able to reach the targeted

internal service or metadata endpoint

  • Why exploitation is possible: IPv6 transition addresses can encode private

or loopback IPv4 destinations while evading incomplete SSRF validation

Why this happens

The vulnerability results from incomplete normalization and validation of network addresses before the server performs outbound requests. An SSRF guard that evaluates only conventional IPv4 and IPv6 representations may fail to identify IPv4 payloads embedded in NAT64, 6to4, or Teredo addresses. This maps to CWE-918 because attacker-controlled URLs can cause the server to initiate requests to unintended internal or privileged destinations.

Potential impact

  • An attacker may access internal administrative services
  • An attacker could retrieve cloud instance metadata or temporary credentials
  • The vulnerability can expose sensitive information from network-restricted endpoints
  • An attacker may use the server as a pivot toward other reachable systems
  • Internal services could be queried or manipulated depending on their controls

General mitigation guidance

Mitigation generally focuses on centralized, canonical URL and address validation before any outbound connection is made. Secure-by-default patterns include resolving hostnames consistently, normalizing all IPv4, IPv6, and transition-address forms, rejecting private, loopback, link-local, multicast, and metadata destinations, and rechecking the final resolved address before connection. Outbound network access should be restricted through egress policies, cloud metadata services should use protective controls, and requests should run with the minimum network and application privileges required. URL inputs should be treated as untrusted even when issued through authenticated commands.

How teams detect issues like this

Detection approaches include static analysis and code review for SSRF guards that compare textual addresses without canonicalization, resolve DNS only once, or omit IPv6 transition formats. Reviewers can inspect every feature that accepts remote URLs and verify that validation occurs before redirects, resolution, and connection establishment. Runtime monitoring can alert on database-originated requests to private address ranges, loopback interfaces, metadata endpoints, and unusual IPv6 transition representations. Egress firewall policies and automated security review tools can provide additional enforcement.

Sources

SourceLink
github.comgithub.com/ArcadeData/arcadedb/security/advisories/GHSA-67m7-7w7g-mpmh
vulncheck.comvulncheck.com/advisories/arcadedb-before-26.9.1-ssrf-via-ipv6-transition-addresses

Frequently asked

questions

It can be exploitable through network-accessible ArcadeDB interfaces when an attacker has authenticated access to an affected command or feature.

The attacker generally needs credentials and permissions that allow use of IMPORT DATABASE or relevant server commands, along with a server able to reach the intended destination.

Use centralized, canonical address validation that covers IPv4, IPv6, and transition mechanisms, then enforce outbound network restrictions as a defense in depth.

Treat all URL values as untrusted, validate the final resolved destination before connecting, test private and metadata address cases, and monitor or restrict server egress.

Find out what's already
exploitable in your codebase.