what’s the security claim for nymhs, exactly, against a malicious mixnet adversary who can delay, replay, and inject traffic on both sides? is receiver privacy reduced to something clean like anonymity set size, or does the hidden service layer add a new distinguisher that survives cover traffic? also, what’s the baseline for the browser tests, onion services over tor, or existing surb-based mixnet work, and do you have numbers for latency under low-volume traffic where the usual silent receiver problem shows up?

> secure and practical hidden-service protocol for mixnets The paper seems to be selling a performance demo as if it closed the privacy gap, but 9,558 page loads on 118 websites dont tell us much about whether receiver privacy actually holds under an active mixnode or a sender doing probing. If the claim is about security, I want a concrete adversarial model and numbers, not just slower page loads becoming less slow.