Описание
Caddy Defender trusted proxy client IP bypass
Impact
Caddy Defender used r.RemoteAddr when evaluating whether a request should be blocked. RemoteAddr is the address of the immediate peer connected to Caddy.
In deployments where Caddy is behind a trusted proxy, CDN, or load balancer, the immediate peer is usually the proxy, not the original client. Caddy resolves the original client address into its client_ip request variable after applying the configured trusted_proxies policy, but Defender did not use that value.
As a result, clients from blocked IP ranges could bypass Defender when accessing Caddy through a trusted proxy whose own IP address was not blocked. This affects deployments that use Defender behind trusted proxies and expect it to enforce blocking based on the real client IP.
Patches
The issue is fixed by making Defender prefer Caddys resolved client_ip request variable when it is available. Defender falls back to RemoteAddr only when Caddy has not provided a resolved client IP.
Users should upgrade to v0.10.1 or later.
Workarounds
There is no complete workaround in affected Defender versions for deployments that rely on Caddys trusted proxy client IP resolution.
Until upgrading, affected users should enforce equivalent IP blocking at the trusted proxy, CDN, load balancer, firewall, or other edge layer before traffic reaches Caddy.
Deployments where Caddy receives traffic directly from clients, without an intermediate trusted proxy, are not affected by this bypass.
Пакеты
pkg.jsn.cam/caddy-defender
< 0.10.1
0.10.1
Связанные уязвимости
The Caddy Defender plugin is a middleware for Caddy that allows users to block or manipulate requests based on the client's IP address. Prior to version 0.10.1, Caddy Defender used `r.RemoteAddr` when evaluating whether a request should be blocked. `RemoteAddr` is the address of the immediate peer connected to Caddy. In deployments where Caddy is behind a trusted proxy, CDN, or load balancer, the immediate peer is usually the proxy, not the original client. Caddy resolves the original client address into its `client_ip` request variable after applying the configured `trusted_proxies` policy, but Defender did not use that value. As a result, clients from blocked IP ranges could bypass Defender when accessing Caddy through a trusted proxy whose own IP address was not blocked. This affects deployments that use Defender behind trusted proxies and expect it to enforce blocking based on the real client IP. The issue is fixed in version 0.10.1 by making Defender prefer Caddys resolved `client