Strict checking for '/multistream/1.0.0' prefix in remote peer protocol negotiation is missing when rust-libp2p acts as dialer

Summary

When rust-libp2p acts as the connection initiator (dialer), it appears that it does not strictly check whether the remote peer's (responder) protocol negotiation message always starts with "/multistream/1.0.0". From the implementation here, the code does not add a tag to show that "/multistream/1.0.0" has been sent from the responder; instead, it just continues with "AwaitProtocol". I also tested this and found that it does not check for "/multistream/1.0.0" during any of the protocol negotiation stages.

Expected behavior

When rust-libp2p acts as the connection initiator (dialer), it should strictly check whether the remote peer's (responder) protocol negotiation message always starts with "/multistream/1.0.0".

Actual behavior

When rust-libp2p acts as the connection initiator (dialer), it does not strictly check whether the remote peer's (responder) protocol negotiation message always starts with "/multistream/1.0.0".

Relevant log output

No response

Possible Solution

No response

Version

No response

Would you like to work on fixing this bug ?

Yes

Start in misc/multistream-select/src/dialer_select.rs at lines 179-195 and trace the dialer’s protocol-negotiation stages. Confirm that the responder’s message is rejected unless it starts with /multistream/1.0.0, then add or update coverage for the dialer path.