It's not that it would degrade your health, it's that it wouldn't significantly impact it. Doing dedicated grip training, for example, will radically increase your grip strength (and I found it super fun to do/have), but it is very doubtful that it will increase your longevity. These correlations/proxies are based on aggregates of untrained people. Once you train it, it is no longer a good metric.

For things like walking speed, it's a little harder to game it without actually increasing your cardiovascular health. But without knowing why something is a proxy, it might be useless/less-useful. For example, if fast walking speed is due to a healthy cardiovascular system, and you notice that your walking speed is low, you might train your walking speed. With effort you might get your walking speed up to 'very fast for non-athletes'. But if you had poor CV health for some underlying reason, you wouldn't have addressed that reason, you would just have masked the proxy. Of course, training walking speed is likely to have positive effects on your CV health, so I'm not saying "don't bother walking faster", I'm just saying that it may cause it to be a much more flawed indicator.