fix: don't compare bounty rewards across mixed USDC/XLM assets

[smartcalsystems-sketch]

Problem

filterBounties's minReward/maxReward bounds and sortBounties's reward-desc/reward-asc sorts compare bounty.reward as a raw number regardless of asset. Since USDC and XLM aren't 1-for-1 (1 XLM is roughly $0.10, not $1), a 500 XLM bounty (~$50) sorts above a 400 USDC one, and minReward=100 keeps a 100 XLM bounty as if it were $100. The asset facet is separate from the reward bound/sort, so the default "any asset" view silently mixes units.

Fix

Added a hasMixedAssets check: when the asset facet is "all" and the candidate set actually spans more than one asset, the reward range filter is ignored and reward-based sorting is skipped (falls back to the existing order) rather than producing a misleading cross-currency comparison. When a single asset is selected, or the data happens to be single-asset anyway, behavior is unchanged — this only changes the specific case where comparison would be meaningless.

I considered the alternative in the issue (normalize via a displayed conversion rate) but there's no exchange-rate source anywhere in this codebase today, and introducing one just for comparison purposes felt like a bigger, separate change (and a staleness/accuracy liability) versus just not comparing incomparable things. Happy to go that route instead if there's already a rate source I missed.

Testing

• Added 4 new tests covering the mixed-asset case for both filterBounties and applyBountyQuery's sort, plus a same-asset-data regression check.
• Full bounty-query.test.ts suite: 26/26 passing (22 pre-existing + 4 new), zero regressions.
• tsc --noEmit and eslint clean on both changed files.

Closes #613

[vercel]

@smartcalsystems-sketch is attempting to deploy a commit to the chonilius' projects Team on Vercel.

A member of the Team first needs to authorize it.