fix: update Morpho GraphQL queries for upstream schema changes (MOO-391)

[lyoungblood]

Summary

The Morpho API shipped breaking schema changes that caused the SDK's Morpho GraphQL queries to fail with HTTP 400 — and because getGraphQL swallows errors silently, every Morpho vault has been reporting an empty rewards array and rewardsApy: 0 through the legacy (non-Lunar-indexer) data path. This is the root cause of MOO-391: the yield-backend (SDK 0.9.15, legacy path) returns an empty rewards object for mwcbBTC (and all other vaults).

Upstream changes handled:

• Market.uniqueKey renamed to marketId (the uniqueKey_in where-filter is unchanged and still valid)
• PublicAllocatorSharedLiquidity.allocationMarket replaced by withdrawMarket — aliased back to allocationMarket in the query so the SDK's public types are unchanged
• VaultState.rewards deprecated in favor of allRewards (removal 2026-09-30) — migrated
• amountPerSuppliedToken / amountPerBorrowedToken deprecated with removal 2026-06-24 — dropped; reward token amounts are now reported as 0 since the API no longer exposes a replacement

Also: Morpho GraphQL errors are now logged with console.error outside browser contexts, so server-side consumers (e.g. Cloudflare Workers) see these failures in logs instead of silently getting empty rewards.

Verification

• pnpm typecheck, pnpm lint, pnpm build pass
• pnpm test: 34 failed / 515 passed — identical to clean main baseline (pre-existing live-network failures in getUserVotingPowers, getMorphoVaultUserPosition, getMorphoVault, getMorphoVaults)
• Live smoke test of the fixed query path: getMorphoVaultsRewards for the cbBTC V1 vault 0x5432...a796 returns WELL supplyApr 0.758 (the expected ~0.77% from MOO-391)
• All three fixed/affected queries validated against https://api.morpho.org/graphql with real data

🤖 Generated with Claude Code

[github-actions]

Quality Gate: ✅ PASS

Metric	Baseline	Current	Δ	Verdict
coverage.lines	81.6%	83.9%	+2.33pp	✅ PASS
coverage.statements	81.6%	83.9%	+2.33pp	✅ PASS
coverage.functions	51.1%	57.8%	+6.77pp	✅ PASS
coverage.branches	81.8%	84.0%	+2.25pp	✅ PASS
lint.errors	0	0	0.00	✅ PASS
lint.warnings	0	0	0.00	✅ PASS
audit.critical	0	0	0.00	✅ PASS
audit.high	0	0	0.00	✅ PASS
duplication.percentage	9.3%	8.9%	-0.34pp	✅ PASS
duplication.clones	101	95	-6.00	✅ PASS

Informational (non-gating)

Metric	Baseline	Current	Δ
audit.moderate	1	1	0.00
audit.low	0	0	0.00
duplication.duplicatedLines	1980	1919	-61.00

Notes

All gated metrics held or improved. If this PR raised the floor (e.g., added tests), consider running pnpm gate:update-baseline after merge to lock the new floor in.