Chapter 16: Building Your Own Knowledge Base#

Everything so far has been about single moves — a website here, a research brief there, a skill you wrote for one repeated job. This chapter is about the move that ties them together: turning one folder into a workspace that remembers who you are and what you have been working on.

I built a tool for exactly this, and I open-sourced it. It is called Command Centre, and it is the closing idea of this part of the book — the point where you stop using Claude Code as a string of clever one-offs and start running your actual work through it.

The problem it solves#

Here is the question that made me build it: do you live in Claude Code and feel like you start from zero every morning?

Every session is a cold start. You re-explain who you are, re-paste last week’s context, re-attach the same files, and only then reach the actual work. The intelligence is there, but the memory is not — Claude forgets you the moment the session ends.

Command Centre fixes that. It turns one folder into your entire work base, so the context is already there when you sit down.

The idea, kept simple#

At its core, Command Centre is three things: a CLAUDE.md that knows who you are, a set of purpose-built folders, and eleven skills built for the real rhythm of knowledge work.

You can picture it as a system with three jobs. It remembers what matters (memory), it does your repeatable work for you (skills), and it reaches your other tools (connectors). Do not overthink the framework — the point is that one folder becomes a place that gets to know you.

Installing Command Centre#

You need Claude Code working first. Then install the plugin from its marketplace.

Inside Claude Code:

/plugin marketplace add vanakrish/command-center
/reload-plugins
/plugin install command-center@command-center

When it asks for a scope, choose Project scope. That records the plugin in the folder’s .claude/settings.json, which is committed to git, so the workspace travels with the folder — copy or clone it anywhere and the skills come along.

I do not recommend User scope here. Command Centre is built to run one folder as your work base, not to hover over every project you open.

First run: /setup#

One command turns the plugin into your actual workspace.

Inside Claude Code:

/setup

It runs a short interview — about twelve questions covering who you are, your remit, the key people you work with, your voice rules, and which tools you have connected. Then it writes three things into the folder: a personalized CLAUDE.md that becomes your operating context, eleven folders (each with its own README), and a small .command-center/context.json that every skill reads from.

Once that is done, make your first move. Drop a thought in with /brain-dump testing my command centre, and the next morning run /daily-briefing to watch the system come alive.

The eleven skills#

Each skill maps to a real moment in the working day. Here is the full set.

The daily rhythm#

This is the part that compounds. The skills are not eleven separate tools — they hand off to each other across the day.

Morning: /daily-briefing pulls your calendar and overnight messages to the top of today’s log, with a top-three focus. Through the day: /brain-dump appends raw, timestamped thoughts with zero friction — no formatting, no filing, just capture.

End of day: /daily-log reads that brain dump and sorts it into Decisions, To-dos, Ideas, and Blockers, then refreshes your CLAUDE.md. Brain Dump is the raw tape; Daily Log is the processed view.

Do that for a few weeks and the folder becomes a searchable record of how you actually work. Every brief, decision, and meeting compounds into context you can build on — the opposite of starting from zero.

Connectors make it sing#

The skills get materially richer once you wire up your other tools — a calendar, a meeting-notes tool, team chat, email, web search. That is why /setup asks which ones you have.

And if you have none yet, everything still works. Skills that need a connector fall back to your local files and tell you plainly what they could not reach, rather than inventing it.

It is yours to change#

None of this is a black box. Every skill is just a SKILL.md file in the repo, written in plain language you can read.

Want to change how /daily-briefing behaves? Copy that skill folder into your project’s own .claude/skills/ and edit it — your version wins, exactly as you learned in the customization chapter. The scaffolding is universal; the contents are yours.

The payoff#

Command Centre is open source and free. The repo is https://github.com/vanakrish/command-center — clone it, read the skills, make it yours.

Start with one folder and one /brain-dump. Keep showing up, and the system gets smarter the more you use it — until the thirty minutes you used to lose re-explaining yourself every morning is simply already there, waiting.

And when you are ready to go past the quickstart: this chapter has a whole part of the book behind it now. Part 5: The Command Centre (Chapters 18–22) is the full build — the framework, the memory system, the workspace design, writing your own skills, and the daily rhythm that keeps it all alive.