In order to reduce signalling, traders may resort to limiting access to dark venues and imposing limits on minimum fill sizes they are willing to trade. However, doing this also restricts the liquidity available to the trader since an ever increasing quantity of orders are traded by algos in clips. An alternative is to attempt to monitor signalling in real time and dynamically make adjustments to the dark liquidity accessed. In practice, price slippage against the order is commonly taken as an indication of signalling. However, estimating slippage is difficult and requires a large number of fills to reliably detect it. Ultimately, even if detected, it fails to capture an important element of causality between dark fills and lit prints - a signature of information leakage. In the extreme, this can lead to scaling back trading at a time when slippage is caused by a competing trader consuming liquidity, and the appropriate action would be to scale trading up -- not down -- in order to capture good prices. In this paper we describe a methodology aimed to address this dichotomy of trading objectives, allowing to maximally capture available liquidity while at the same time protecting the trader from excessive signalling. The method is designed to profile dark liquidity in a dynamic fashion, on a per fill basis, in contrast to historical venue analyses based on estimated slippage. This allows for a dynamic and real-time control of the desired liquidity exposure.