Why not testing? Approached the right way, you get to use a holistic skill set while remaining far more hands on than you do if you drift into project management. A good tester needs to quickly understand a system from top to bottom to understand where the risks are, and start digging where the bodies are most likely buried. Being able to quickly find the issues that matter, articulate them in appropriate language for both the dev side and the business side, and, even better, highlight potential issues during design and planning turns out to be a highly desirable skill set. Being able to quickly knock up a complex test network, code some automated integration tests, do some SQL queries, do some network analysis with wireshark through to getting code reviewed, merged in and deployed to production is all part of the job. In terms of consultancy, you can quickly get a reputation as a very useful person to have around.
There is a specialist aspect in that you do need a slightly evil brain to be a good tester, and if you're working with anything GUI/front end based, you do need to have more empathy for end users than your average dev/sysadmin, but this is more about attitude than skills. While not for everyone, it's a specialism that the right sort of generalist can dive into and thrive. One word of warning, there's still a lot of old school 'write a detailed manual test script in excel, then brainlessly follow it' testing out there, especially in the corporate outsourcing world. You don't want that. No one wants that.