fix: enforce patient's phase preference in API call, not LLM guidance

perf: rank trials by phase, cap at 15, strip bloat, tighten defaults (#25)

test(critic-fixes): address brittleness, over-mocking, and missing edge cases (#20)

test(wave-4): LLM orchestration tests for eligibility, intake, and research (#19)

test(wave-3): deterministic eligibility and trials_api tests (#18)

test(wave-2): add pure-function tests for models and prompts (#17)

test(wave-1): add test infrastructure and conftest.py (#16)