PARA — What braynee manages

Pillar 2 needs somewhere to keep memory. That place is an Obsidian vault organized with PARA — four buckets, by actionability. braynee scaffolds this structure for you during /braynee:setup; you don't build it by hand.

The domains you actually care about — clients, business, email, projects, goals — aren't separate features. They're just entries in these four buckets, mostly Areas and Projects.

The four buckets

1. Projects — active builds with an end state

A Project has a finish line. The code feature you're shipping, the PRD you're writing, the migration you're running, the landing page you're launching. Each lives under 1. Projects/ as a note or a named folder.

When you open a Claude Code session in a repo, braynee's session start finds the matching Project note and the related session history, and feeds that slice in — not the whole vault.

2. Areas — ongoing responsibilities with no end

An Area never "completes" — you just maintain it. This is where most of the domains you care about live:

• Clients — relationship context, engagement notes, status, meeting prep.
• Business — your companies, offerings, decisions, operating knowledge.
• Email — correspondence context braynee can pull when you draft replies.
• Goals — what you're steering toward, referenced across sessions.

These sit under 2. Areas/. Because they have no end state, they're the long-lived memory the second brain leans on most.

3. Resources — reference and templates

Reusable material that isn't tied to one Project or Area: API references, playbooks, templates, research, infrastructure notes. Under 3. Resources/. braynee surfaces these when a QMD search matches what you're working on.

4. Archives — done or retired

Completed Projects and inactive Areas/Resources move to 4. Archives/. They stay searchable — old decisions still matter — but they're out of your active working set so context stays sharp.

How braynee pulls a slice into context

The vault is the memory; PARA is what makes a slice of it addressable. The flow:

1. Session start detects the project/branch and loads the matching 1. Projects/ context plus recent 2. Areas/Sessions/ history.
2. During work, decisions and progress are written back into the right bucket, so memory stays current as you go.
3. Recall and skills (/braynee:recall, /braynee:clients, /braynee:tasks) query across the buckets via QMD to pull exactly the relevant Projects, Areas, or Resources — not the whole vault.

That selective pull is the entire point: PARA gives braynee a structure precise enough to answer "what's the right context for this?" instead of dumping everything.

Next