feat: per-network validator leaderboard tabs and richer testnets metrics

[albert-mr]

Summary

Validator leaderboard (/validators/leaderboard):

• Asimov | Bradbury network tab pills with lazy fetch per tab
• Client-side re-ranking (1, 2, 3...) since backend returns global ranks
• URL deep-linking via ?network= query param
• ARIA tab roles for accessibility
• Empty state: "No validators on {Network} yet"

Testnets overview (/testnets):

• Total validator count from wallet API network_stats (counts all on-chain wallets, not just profiled validators)
• Status breakdown: active (green), quarantined (yellow), banned (red), inactive (gray)
• Deep-links from "Leaderboard →" to the matching network tab
• Leaderboard API calls limited to top 5 (was fetching all entries)

Test Coverage

• AI-assessed coverage: 0/14 code paths (0%) — existing test infrastructure has 4 pre-existing broken suites
• No new tests generated (component test foundation unstable)
• 14 passing tests in 2 suites unaffected by changes

Pre-Landing Review

Pre-Landing Review: No issues found.

• 3 issues auto-fixed during review: initial state shape, optional chaining on API response, removed unused imports

Plan Completion

14/14 plan items DONE. All changes match the approved design doc.

Scope Drift

Scope Check: CLEAN
Intent: Per-network validator leaderboard tabs + richer testnets metrics
Delivered: Exactly that — 2 Svelte components + package-lock

Test plan

• Frontend tests pass (14 passing, 4 pre-existing failures unrelated to changes)
• Pre-landing review clean
• Plan completion: 14/14 items addressed

🤖 Generated with Claude Code