Building a Second Brain That Understands How You Work

Most of what happens in a workday disappears.

Not dramatically. It just fades. The Slack thread where a client revealed something about how they actually make decisions. The offhand comment in a meeting that told you exactly how someone wants to be communicated with. The thing you figured out on Tuesday that you’ll spend 30 minutes re-figuring out in six weeks because you never wrote it down.

Nothing was getting lost in a way I could point to. But nothing was adding up either.

The Problem With Scattered Context

I was keeping track of everything in the usual places: Slack, meeting notes, calendar, whatever felt right that week. Each one had a piece of the picture.

The problem isn’t that any single tool is bad. The problem is that none of them talk to each other, and none of them understand the relationships between the things you’re tracking. A client isn’t just a name. A project isn’t just a task list. A person isn’t just a contact. Everything at work is connected, and a pile of separate notes doesn’t reflect that.

What I wanted was a system that understood the structure of how I actually work: clients connected to projects, projects connected to people, daily logs that link back to both. Context that compounds instead of scatters.

What I Built

The foundation is Obsidian: everything stored as plain Markdown files, linked together with wikilinks, and queryable through Dataview.

The first thing I created was About Me.md. Not a task list or a project log. A document that tells Claude Code who I am: what I value, how I think about my career, what kinds of advice actually fit my situation and what kinds don’t. That file loads into every session. The system doesn’t just know what I’ve done. It knows who I am.

From there, the vault is built around eight interconnected note types: Daily Logs, Big Wins, People, Projects, Companies, Goals, Action Items, and Decisions. Everything links. A log entry from a client meeting links to the company, the project, and the people in the room via Obsidian wikilinks. A Big Win links to the goal it advanced. An action item traces back to the conversation that created it. A Decision, written as a proper Architecture Decision Record with context, alternatives considered, and consequences, links to the project and the people who were involved.

That last one is underrated. Technical decisions are some of the most valuable context a team loses. Writing them down as structured ADRs, linked to the work they came from, means I can trace why something was built the way it was. Months later, in a different conversation, with a different client.

Two Dataview dashboards aggregate across the vault automatically. The daily dashboard shows open action items, active goals, and a two-week energy trend. The strategy dashboard shows something more interesting: which goals haven’t appeared in a daily log in fourteen days, and which people with a regular contact cadence are overdue for a conversation. It surfaces the things I’d otherwise let slip without noticing.

That second layer is what makes this a genuine second brain rather than an organized archive.

The Memory Layer

Here’s what that actually means in practice.

When I sit down to work on something, drafting a message to a client, thinking through a project approach, preparing for a conversation, Claude Code can read the relevant Obsidian files before we start. It’s not starting from scratch. It has the context from every previous session that touched that person or project: what I’ve learned, what’s worked, what hasn’t.

New context captured today becomes a memory. That memory influences how I work tomorrow. Over time the system doesn’t just record what happened. It shapes how I approach what comes next. That’s the difference between a notes app and a second brain.

The Daily Capture Skill

The thing that makes the system actually work is a Claude Code skill I built for daily capture.

I drop things in throughout the day. Meeting notes, observations, something someone said out loud that matters, a thought I don’t want to lose. No formatting required, no decisions about where it belongs. It’s less a daily log and more a self-organizing scratch-pad: whatever goes in gets filed into the right place automatically. The skill reads raw input and turns it into structured entries. Which projects came up, which goals moved forward, who was involved, what got accomplished. Everything lands in Obsidian as linked, structured Markdown.

Two things it does that I designed deliberately.

First: it catches action items I would have let slide. Not because I didn’t know about them, but because they got buried in everything else. The skill flags them and asks me to confirm before anything gets created. Nothing gets written without my approval.

Second: it surfaces Big Wins I would have skipped past. I am genuinely humble to a fault. Left to myself, I’d log the work and move on without noting that something went well. The skill reads the entry and, if something sounds like a real win, proposes a Big Win record linked to the goal it advanced. Over time that becomes a real log of what I’ve actually accomplished, not just what I’ve done.

Getting those inference rules right required thought. What counts as a win? What rises to the level of an action item versus a passing mention? Those aren’t things a tool can decide. I defined them. The skill executes against them.

What Six Months Actually Shows

Problems I’ve already solved keep showing up in new contexts. Having a structured record means I recognize them faster and reach for something that already works instead of treating each one as new.

Blind spots in how I communicate with people. The Obsidian notes for each person accumulate over time: how someone prefers to receive feedback, what they care about, how they make decisions. When you read those back across six months of interactions, patterns emerge that you’d never notice in the moment. I’ve changed how I approach specific people because of what the data showed me.

That’s the context I didn’t know I was missing. Not any single piece of it. The aggregate.

The System Evolves

One thing I was deliberate about from the start: the schema isn’t fixed. Obsidian notes are plain text. The Claude Code skill is a Markdown file I can update whenever the inference rules need to change. As my work shifts, the system follows.

That’s the right way to build something like this. A second brain built around last year’s priorities isn’t useful. The structure should reflect how you actually work today, not how you worked when you set it up.

The best tools aren’t the ones with the most features. They’re the ones that fit how you actually think.