Detections
Analytics
CVE-2026-24904
medium
Description
TrustTunnel is an open-source VPN protocol with a rule bypass issue in versions prior to 0.9.115. In `tls_listener.rs`, `TlsListener::listen()` peeks 1024 bytes and calls `extract_client_random(...)`. If `parse_tls_plaintext` fails (for example, a fragmented/partial ClientHello split across TCP writes), `extract_client_random` returns `None`. In `rules.rs`, `RulesEngine::evaluate` only evaluates `client_random_prefix` when `client_random` is `Some(...)`. As a result, when extraction fails (`client_random == None`), any rule that relies on `client_random_prefix` matching is skipped and evaluation falls through to later rules. As an important semantics note: `client_random_prefix` is a match condition only. It does not mean "block non-matching prefixes" by itself. A rule with `client_random_prefix = ...` triggers its `action` only when the prefix matches (and the field is available to evaluate). Non-matches (or `None`) simply do not match that rule and continue to fall through. The vulnerability is fixed in version 0.9.115.
References
https://github.com/TrustTunnel/TrustTunnel/security/advisories/GHSA-fqh7-r5gf-3r87
https://github.com/TrustTunnel/TrustTunnel/commit/aa5060145506952b9431b0ed3edb52bb6c08d9a6