Ask HN: What is your knowledge toolchain in 2018?

I've found it increasingly difficult to engage with and keep track of the sum of knowledge I encounter every day between papers, talks, podcasts, blog posts, and my own thoughts. Specifically, I care about the following tasks:

1. Active engagement: while consuming a piece of knowledge, I want to write down my thoughts/reactions with as little friction as possible. For visual knowledge (web pages and PDFs), I want to annotate my thoughts inline (preferably with support for stylus input).

2. Summarization: after finishing a piece, I want to review my annotations and pick out the important ones, as well as potentially write a standalone summary of the piece. I also want to categorize the piece via tags.

3. Linking: I want to be able to provide different kinds of links between knowledge pieces, either implicitly (e.g. by providing the same tag) or explicitly (e.g. "this web page is the HN comment section for this article"). When writing comments, I should be able to write down internal hyperlinks to any knowledge piece in my collection.

4. Contextualization: I want to compare/contrast the main points of the article with linked articles. I should be able to easily search along every dimension of my data cube (e.g. all articles from source X, all articles about topic Y, etc.) to produce different contexts.

For lack of a true memex, my question is: what tools do you use today to accomplish these tasks? From my experience, almost everyone has a different ad-hoc setup, so I'm curious to understand the diversity in HN's community. I've seen people use tools like Evernote, Scrivener, or even a custom MediaWiki to perform a subset of these tasks, but I've never seen a complete toolchain.

I'm also interested in perspectives on: should this problem be approached in Unix fashion (each task deserves a single tool that does it well?) or in Emacs fashion (the entire toolchain needs a single, over-arching, opinionated framework for maximum integration)?