Isolate eval subprocesses from local yt-dlp config and fix nDCG normalization against the judged pool.
Validation: uv run python -m unittest tests.test_evaluate_search_quality
This workspace uses GOOGLE_API_KEY as the canonical Google credential. Accept it ahead of the Gemini-specific aliases so the local evaluation harness can run without a separate GEMINI_API_KEY export.
Validation: uv run python -m unittest tests.test_env_project tests.test_evaluate_search_quality and a one-shot keychain-backed resolution check.
Add an optional local evaluator that compares a baseline revision against a candidate checkout, computes deterministic stability metrics, and can call Gemini for judged ranking metrics when configured.
The harness isolates child runs with a temporary HOME and a node-free PATH so historical revisions cannot trigger Bird browser-cookie auth during evaluation.
Validation: uv run python -m unittest and local smoke/full deterministic eval runs.
Matt Van Horn