Ask about publication gaps in the rate provider review

[EndymionJkb]

The review checklist asks whether a rate provider's price source is off-chain, whether the price is volatile, and whether it is open to donation attacks. It does not ask whether the source publishes continuously. This adds that question, to both the checklist and the generated template, and makes it a required input to the review script.

A source that stops publishing on a schedule cannot be assessed with an elapsed-time bound, which is the usual way to think about freshness. Any maximum age short enough to catch a source that failed between publications also rejects every routine gap, and any bound loose enough to survive a routine gap is far too loose to catch a failure. There is no single value that does both jobs, so the question has to be asked differently rather than answered with a number.

The existing volatility question is close to this but not the same. It asks whether the price is a market price or a mostly monotonically increasing one. That is about how the value behaves while it is being published. This asks whether it is being published at all.

rate-providers/template.md now two new items under Oracles. The first records whether anything rejects a price that has stopped updating, and where that enforcement lives, since it is often in the source contract rather than the rate provider.

The second covers scheduled publication, with follow-ups for the publication pattern, the longest routine gap, whether the value is a live quote or a periodic valuation, how far the value actually moved across that gap, and what getRate returns during one.

The main question is whether the value can move discontinuously while the source is silent. That is what separates a gap that is a risk from a gap that is simply how the source works. A live market quote can gap when its venue reopens. A smooth NAV accrual cannot.

The same section has a new short guidance block, because there are subtleties in the question, and it's easy to gloss over important details. For maximum clarity, it sorts sources into three categories: 1) live quote on a venue calendar; 2) periodic valuation; and 3) continuously published, and states plainly that the asset class does not decide which one applies.

The third category is the one most often misread: an asset can be a tokenized bond or equity and still have a source that never stops publishing, so the publication record has to be read rather than inferred.