I don't know that there's going to be a single book that describes what you want, however, I'm going to make some suggestions which may help you on your way. I'm a little biased towards CS and research-heavy MSc, so I think I can be of some help here.
A master's degree typically is going to consist of two phases: One of breadth: where you try to find what you're interested in enough to specialize, and one of depth: Where you start to push towards the edge of human knowledge on that subject.
If you're already reading papers, then you should be finding the kind of papers that hold your interest the most, those which have you tracking down their references.
Once you've already got a topic, start to find the papers from the best conferences and journals (in CS, good conferences are often as good or better than Journal papers) in the area. Creative use of Google can often track down full-text even of papers which are officially 'paywalled'.
The goal of your reading here is to identify a gap in the understanding. This might be as simple as working on comparative evaluations of techniques, or as complicated as trying to create a new technique for yourself. What you'd like to do is answer a question the community is interested in.
While you're doing your reading and trying to find your topic or get up to speed on the state-of-the-art, you should also be building up your background. There's always room to grow your background to support research. In CS, there are certain foundational topics which are (almost) always going to pay dividends when doing research: Algorithms and algorithmic analysis, statistics and probability, and utility programming (Data munging, shell-scriptng, etc.).
As you get further into a specialization you'll notice more techniques being deployed which will require more background --- i.e. your specialization might require more calculus than you know, which then will involve learning the calculus required.
I know you asked for a methodology, but in my experience, research is a process of creativity and perseverance. At the highest of levels, you're going to come up with an idea, work on it, and find out it doesn't work. Rinse and repeat until you find out something that works --- in a special case. If the special case is interesting or common, you might be able to publish it! Otherwise, you'll work on generalizing the idea until you can.
To be recognized as a member of the research community, eventually you'll have to publish. This means learning to write.
I'd suggest learning to write informally first -- Start a blog, and document your learning. If you're not a native English speaker, use this time to practice writing in English. Get everyone and anyone you can to read your blog, and tear it apart on style, grammar, content, concision.
After your confident in informal English, start to transition into Academic writing. Read good papers. Take notes on what makes them good in your mind. Read about writing good papers. Write papers. Take previous projects, and start writing them up like you would submit them. As you write, you'll find the holes and flaws. Finish it anyhow, then sit down, and edit them: Why wouldn't they be accepted. Keep doing this! As far as format and outline, it's going to vary from field to field, and even venue to venue to some extent.
This is going to be a very difficult process to do alone, with little guidance. Ideally you'll seek out and find a mentor. This might involve actually signing up with a school for a Master's program.
Most of the pieces I wrote about above will be an order of magnitude easier with an academic guide.