test(message_receiver): pin receive() propagation-source threading via a message_validator test-utils feature

[petarjuki7]

Goal

Pin the behavioral half of the per-peer duplicate fix from #1131 / PR #1203: NetworkMessageReceiver::receive must pass Some(propagation_source) (not None) into Validator::validate (message_receiver/src/manager.rs:83). If that regresses to None, received_from.is_some() is false for every inbound message, so every role-8 duplicate classifies as the Ignore-class relay and the whole existing suite still passes.

Why it isn't covered today

The per-peer matrix tests in #1203 call validate_partial_signature_message directly with their own PeerId, so they never exercise the receive seam. A receive-level guard needs a real role-8 SignedSSVMessage that survives full validation — including RSA verification against the operator pubkey seeded into NetworkState. That requires a matched pair: the message signed by the same key whose pubkey is in the DB. The fixtures that build this (create_signed_proposer_preferences_message, create_operator_pub_keys, MockDutiesProvider, create_committee_info) are pub(crate) inside message_validator's #[cfg(test)] module and are unreachable from message_receiver; message_validator has no test-utils feature to expose them.

Suggested approach

1. Add a test-utils feature to message_validator that exposes the role-8 message / committee / duties fixtures, mirroring database's existing test-utils feature.
2. Add a message_receiver test (dev-dep on message_validator with test-utils and database test-utils): construct a NetworkMessageReceiver, spawn a processor, is_synced = true, seed a committee/share into NetworkState, then call receive twice from the same peer with the same message and assert the second Outcome.action is Reject.

Acceptance criteria

• The new test fails if manager.rs:83 is changed to None.
• No production logic change beyond gating existing test helpers behind the test-utils feature.

Notes

• Follow-up to fix(message_validator): classify proposer-preferences duplicate roots per peer (SIP-94 §7) #1131 / PR fix(message_validator): classify proposer-preferences duplicate roots per peer #1203, split out to keep that PR focused (per review discussion).