Most strategies work in one kind of market. Mean-reversion rules make money in ranges and get destroyed in trends; breakout rules do the reverse. A regime filter tries to detect which world you are in and stand aside in the other one.
Common filters: price above or below a long moving average, realised-volatility below its median, the VIX under 25, or a hurst-exponent above 0.5. All of them are lagging by construction, because they need history to compute.
Be honest about what this buys you. A filter that halves your time in the market also halves your sample. If the strategy only produced 300 trades to begin with, the filtered version is a 150-trade claim and needs out-of-sample evidence before you believe it.
Related: market-regime, signal-filter