I guess it depends on where you consider the "end" to be.
If it's at the user's computer, and all software inside is considered to be trusted, having one of these proxies to scan for malware on the "end computer" could actually improve security.
Or if it's at the user's company, then being able to scan all traffic incoming/outgoing might increase (the company's) security.
On the other hand, if you consider the "end" to actually be a person, with software in the computer not being 100% trusted, then you're right, this could only reduce the amount of security.
Personally, I'm not that sure about any of these two to be right. Maybe a better way would be to have some traffic proxy-able, some marked as "don't proxy", and running your browser/tabs in a VM... but there might be several different scenarios with different requirements.