#dusk @Dusk
Woke up today and honestly the market felt half asleep, half awake. That weird in between mood where nothing's moving but nothing's dead either. So instead of doom scrolling charts I ended up back in Citadel 2's docs, don't ask me why, just felt like reading something technical.

I was trying to find the exact point where Citadel actually decides if a user gets access. Kept scrolling, kept expecting to hit it. Never did. Turns out that was never Citadel's job.

User proves they hold a valid license without revealing who they are, the license contract on Dusk verifies it, a nullifier gets appended, session cookie gets issued. Reads like access granted, right. But the last step in the flow is the one nobody mentions. The Service Provider fetches that session, checks the cookie locally, and then makes its own call on whether to actually grant service. Citadel just confirms the credential is real and hasn't been reused already. It says nothing about whether that credential is even good enough for what the SP wants.

Small distinction but it changes what the whole thing is actually solving. A valid session tells you the proof was generated correctly by someone holding a genuine license, no double spending involved. It does not tell you whether that attribute set matches what this particular SP requires today. One's a math problem, Dusk solves it once and it's done. The other is a business call, every SP makes it themselves, and it can change tomorrow without touching the proof system at all.

What I'm keeping an eye on isn't how fast proofs generate on Dusk. It's whether SPs actually build real policy logic on top of Citadel, or whether verification quietly becomes just another pass fail gate with extra steps.

If a system can prove something is true but can't decide what that truth means, feels like the real bottleneck was never the proof. It was always the party asking the question.

$DUSK

$TUT

$PORTAL


Where should Citadel 2's access model go next?