Tasks: Quality & Diversity Baseline

T1: Add "brief" built-in ranking profile

File: tidal/src/ranking/builtins.rs

Add constants: BRIEF_VIEW_GATE_THRESHOLD, BRIEF_COMPLETION_GATE_THRESHOLD, BRIEF_MAX_PER_CREATOR, BRIEF_FORMAT_MIX_MAX_FRACTION, BRIEF_EXPLORATION
Add fn brief() -> RankingProfile following the existing skeleton pattern
Configure: Sort::Hot { gravity: 1.5 }, two gates (view >= 3, completion >= 1), three boosts (view, like, completion), diversity (max_per_creator: 2, format_mix: 0.6), exploration: 0.0 (disabled — bypasses quality gates)
Register in register_builtins() after date_saved()
Update the builtin count test
Fix format enrichment in query/executor/mod.rs: populate ScoredCandidate.format from item metadata
Fix metadata loading trigger: extend needs_metadata_for_creator_grouping to check format_mix_max_fraction
Run cargo test --manifest-path tidal/Cargo.toml --lib to verify no regressions

File: tidal/tests/p1_quality_diversity.rs (new)

Create test helpers: schema with view, completion, like signals; item writer with configurable signal counts
Test quality_gate_excludes_low_view_items: 10 items, 5 with < 3 views. Assert only high-quality items in results.
Test quality_gate_excludes_zero_completion_items: 10 items, 5 with 0 completions. Assert only items with completions in results.
Test creator_diversity_enforced: 30 items from 5 creators. Assert no single creator dominates top 10.
Test format_diversity_enforced: 12 items (6 video, 6 podcast), all different creators. Assert no format exceeds 60%.
Test brief_profile_registered: Assert profile resolvable, empty results on empty DB.
Test combined_quality_and_diversity: 13 items mixed quality/creators/formats. Assert all constraints hold.

Run cargo test --manifest-path tidal/Cargo.toml --lib (1352 tests)
Run cargo test --manifest-path tidal/Cargo.toml --test p1_quality_diversity (6 tests)
Run cargo clippy --manifest-path tidal/Cargo.toml -- -D warnings
Run cargo fmt --manifest-path tidal/Cargo.toml -- --check