feat: add explicit benchmark drawdown guard

[Pigbibi]

Summary

• add a price-only, fail-closed benchmark drawdown guard with explicit lookback, thresholds, scalars, and freshness policy
• mount it only through the existing unified market-regime control, disabled unless a strategy explicitly enables and freezes every policy input
• retain shadow/no-order execution controls; no strategy or platform configuration changes are included
• align uv.lock with the already-committed Ruff 0.16.4 test dependency

Validation

• uvx ruff check src/quant_strategy_plugins/benchmark_drawdown_guard.py src/quant_strategy_plugins/init.py src/quant_strategy_plugins/market_regime_control_plugin.py src/quant_strategy_plugins/strategy_plugin_runner.py tests/test_benchmark_drawdown_guard.py tests/test_market_regime_control_plugin.py
• uv run --extra test pytest -q tests/test_benchmark_drawdown_guard.py tests/test_market_regime_control_plugin.py tests/test_strategy_plugin_runner.py tests/test_plugin_policies.py tests/test_plugin_catalog.py

The guard is a reusable research component. Any future strategy mount requires its own P1/P3 source assurance and out-of-sample backtest evidence before promotion.