Every protocol Stenion tracks, ordered by its live safety score — higher is safer. Ranking is derived purely from on-chain data; no protocol can pay to move up. Open a protocol to see the full factor breakdown behind its number.
Some entries are individual markets running another protocol’s contracts rather than protocols in their own right. Those are labelled on the row, and scored on their own reserves, oracle and admin like any other entry.
9 protocols we assessed and don’t score appear here too, marked as such and never ranked. See just those.
5 scored · 9 assessed, not scored
9 protocols and markets we looked at and don’t publish a score for. None of these has a safety score — not a low one, not a zero. They are unranked and cannot be compared with the entries above.
Listing a protocol here is a statement about Stenion’s coverage, not a criticism of the protocol. Open one to see what we found and how you can check it yourself.
Stenion’s adapters read trustless Stellar infrastructure — Soroban RPC and Horizon — and nothing else. A protocol whose reserves and positions live on another chain cannot be scored without reading that chain, which would change what the score means for every protocol.
Protocol names, logos and trademarks belong to their respective owners and are shown for identification only. Their presence here and any link to a protocol's own site or documentation does not imply endorsement, partnership, or any relationship with Stenion, in either direction.
These markets can be read, and four of the five factors compute normally for them. What their price feeds do not publish is a staleness tolerance or a deviation bound — the two on-chain parameters oracleSafety is anchored to, neither of which SEP-40 defines. Scored anyway, the nearest substitute rates a price that is hours old as perfectly fresh, and dropping the factor instead would rank a market higher for having an oracle we cannot inspect. Not a judgment about these protocols or their oracles: it says only that this particular number has nothing to be computed from.
These markets can be read; there is just almost nothing in them. Scored anyway, every factor would fall to its can’t-assess branch and publish 0 — a number in the danger band, meaning the opposite of what is true. The floor is a precondition on scoring, not a quality bar: it says only that a number was computed from something rather than from nothing.
Known, and on the list to evaluate. There is no mainnet deployment to read, so there is nothing to score — and nothing yet confirmed about whether it would be scorable when there is.
Nothing about these markets stops them being scored. A rulebook covers them, their contracts read normally, and the number would be exactly as true of them as of the market from the same protocol that is scored. What is missing is a slot: one scoring cycle runs inside a serverless function with a hard sixty-second ceiling, and every registered market has to get a full attempt inside it, which caps how many can be indexed at all. This is a statement about Stenion’s capacity and not a finding about these markets — it ends when the capacity changes, and nothing the protocol does affects it.
