Know every vulnerabilitybefore it knows you.
DevGuard continuously monitors your dependencies and alerts you when CVEs like this one affect your stack — with real-time threat intelligence built for developers.
GHSA-5jh9-2h63-pw4q
Summary
CC-Tweaked's HTTP API (http.request, http.websocket) blocks requests to private network ranges to prevent server-side request forgery (SSRF). This protection can be bypassed on IPv6-capable servers using NAT64 well-known prefix addresses (64:ff9b::/96). An attacker who can execute Lua code can reach any internal IPv4 service that the filter is intended to block, by addressing it as http://[64:ff9b::<ipv4-as-hex>]/ instead of its direct IPv4 address. This affects any CC-Tweaked deployment on a network with NAT64 routing — a configuration that is standard on AWS, GCP, and other cloud platforms when using IPv6-only subnets.
Details
The IP filter in PrivatePattern.matches() (AddressPredicate.java#L121–L130) blocks private network ranges by calling Java's standard InetAddress classification methods:
public boolean matches(InetAddress socketAddress) {
return socketAddress.isAnyLocalAddress()
|| socketAddress.isLoopbackAddress()
|| socketAddress.isLinkLocalAddress()
|| socketAddress.isSiteLocalAddress()
|| socketAddress.isMulticastAddress()
|| isUniqueLocalAddress(socketAddress)
|| isCarrierGradeNatAddress(socketAddress)
|| additionalAddresses.contains(socketAddress);
}
When a NAT64 address such as 64:ff9b::c0a8:0101 (encoding 192.168.1.1) is resolved via new InetSocketAddress("64:ff9b::c0a8:0101", 80), Java returns an Inet6Address. Every method above returns false for this address — the 64:ff9b::/96 prefix is not recognised by any of Java's built-in classification methods. The address passes the filter and CC-Tweaked opens a connection.
On a network with a 64:ff9b::/96 → NAT Gateway route, that connection is translated at the network level: the embedded IPv4 address is extracted and the packet is forwarded to 192.168.1.1:80. The internal service receives a normal IPv4 TCP connection.
The existing 6to4 (2002::/16) mitigation in AddressRule.java#L74–L75 does not cover the NAT64 prefix. No other check catches 64:ff9b::/96.
PoC
Preconditions (all three required):
- The server running CC-Tweaked has an IPv6 address assigned
- The network has a NAT Gateway (or equivalent)
- The route table contains
64:ff9b::/96 → NAT Gateway— the standard AWS/GCP configuration for IPv6-only subnets with outbound IPv4 access (AWS documentation)
Lua payload (targets an internal service at 10.0.1.17:8888):
-- 10.0.1.17 = 0x0a000111, expressed as NAT64: 64:ff9b::0a00:0111
local res = http.request("http://[64:ff9b::0a00:0111]:8888/")
if res then print(res.readAll()) end
Conversion formula — for any blocked IPv4 a.b.c.d, the bypass address is:
64:ff9b::<hex(a)><hex(b)>:<hex(c)><hex(d)>
Impact
This is a server-side request forgery (SSRF) vulnerability. Any user able to execute Lua code on a CC-Tweaked computer — including players on a public server — can use it to send HTTP requests to internal IPv4 services that the HTTP filter is designed to block. On cloud-hosted servers (AWS, GCP, Azure) using IPv6-only subnets with NAT64, which is an increasingly common configuration following AWS's February 2024 public IPv4 pricing change, this includes other instances in the VPC, internal databases, and cloud management APIs.
Suggested fix: add a check for the NAT64 well-known prefix in PrivatePattern.matches():
|| isNAT64Address(socketAddress)
private boolean isNAT64Address(InetAddress address) {
if (!(address instanceof Inet6Address)) return false;
byte[] b = address.getAddress();
// 64:ff9b::/96 — NAT64 well-known prefix (RFC 6052)
return b[0] == 0x00 && b[1] == 0x64 &&
b[2] == (byte) 0xff && b[3] == (byte) 0x9b &&
b[4] == 0 && b[5] == 0 && b[6] == 0 && b[7] == 0 &&
b[8] == 0 && b[9] == 0 && b[10] == 0 && b[11] == 0;
}
This vulnerability does not seem to work on servers running on modern versions of MacOS.
The vulnerability can be exploited over the network without needing physical access. It is easy for an attacker to exploit this vulnerability. An attacker does not need any special privileges or access rights. No user interaction is needed for the attacker to exploit this vulnerability.
Limited exploitation activity has been observed. Close monitoring and planned remediation are recommended.
Probability that this vulnerability will be exploited in the wild within the next 30 days.
We did not find any exploit available. Neither in GitHub repositories nor in the Exploit-Database.
Browse More
Continuously monitor your dependencies and get alerted when vulnerabilities like this one affect your stack.
Checkout DevGuard