Status update: all three proposals passed but deployment paused on a Pyth/CosmWasm blocker
First, thank you. All three StoreCode proposals passed!
#1037 - factory 89.4% yes - 1901
#1038 - creator-pool 80.3% yes - 1902
#1040 - router 55.1% yes - 1903
All three `data_hash` values match the sha256 hashes published in the proposals and are reproducible from commit `91cc563`.
Unfortunately we ran into a snag when about to deploy. The problem is Bluechip prices every commit against a USD threshold using the Pyth OSMO/USD feed, reading the Pyth CosmWasm contract on osmosis-1. That contract appears to be permanently frozen and has been for 157 days. Moreover, every feed is stuck on the same timestamp.
Using a current, authenticated Hermes payload:
- `get_update_fee` still succeeds (returns `1uosmo`) — the header parses
- simulating `update_price_feeds` with that same payload fails with `GuardianSetExpired`
The contract verifies Wormhole guardian signatures, and Pyth replaced Wormhole with a 5-router 3-of-5 quorum after 2026-07-31. The guardian set it trusts has expired, so it will reject every update regardless of feed or payload.
Code 142 has exactly one instantiation (the frozen contract), its admin is itself so no third party can migrate it, and Pyth’s documentation states “Pyth Core no longer supports CosmWasm as of August 26, 2026.”
Hermes now requires an API key, and OSMO/USD seems to be behind a pretty hefty pay wall (I did message pyth as they are taking requests according to their docs)
Our factory reads the oracle through a *configurable* contract address and a simple query interface, so a compatible on-chain price source can be swapped in via the standard 48-hour config timelock without changing the contracts this community already approved.
Not deploying until the oracle is real. Our contracts fail closed on a stale price by design, so instantiate would refuse anyway. We are not going to ship something that cannot price a deposit.
This did not change the stored code is unaffected and still reproducible from `91cc563`. Nothing about the audit, the testnet validation, or the contract logic has changed. The oracle is a configuration dependency, not a flaw in the approved code. It is unfortuantely a hard dependency, and we will not launch around it.
We will report back as soon as we hear from Pyth or land on a path. Questions welcome here or at