How I build context in my Obsidian notes

I build context so the note that owns the work shows exactly which notes, sections, and external sources are relevant.

When it comes to context, less is more:

• more tokens solving the problem, less on retrieval
• more focus on what matters to produce the best result
• less chance that unrelated material pulls the work in the wrong direction
• less time rebuilding the same context for every delegation

What I am trying to avoid

I do not want AI guessing from memory.

I do not want to keep iterating in a long chat thread to prevent losing context.

I do not want to dump my whole vault into a chat and hope the model notices what matters.

I do not want context to be rebuilt for every delegation.

I do not want similar notes to be mistaken for relevant notes.

What I do instead

I make context visible in the note before delegating.

Depending on the work, that looks like:

• links under a Context heading
• a Smart Context codeblock
• named context packages for reusing the same context across multiple outcome notes

This page is about building and attaching context.

When the note is ready to be copied as an assignment, use How I use my notes as assignments for delegating work to AI in Obsidian.

The rule I use

A source belongs in context when it helps produce the desired result.

A source does not belong just because it is related.

My default pattern:

## Context

- [[specific useful note]]
	- Use for: [why this matters to the result]
- [[another useful note]]
	- Use only for: [specific section, constraint, source, or example]

The Use for line is important.

It tells future me, and AI, why the source belongs.

The exact workflow

1. Add a Context section

I start with a plain Context heading.

## Context

- [[specific useful note]]
	- Use for: [why this note matters]
- [[another useful note]]
	- Use only for: [specific section, constraint, source, or example]
- [external source]
	- Use for: [why this source belongs]

The goal is to make relevance visible.

2. Insert a Smart Context codeblock

When I need a lot of context, I insert a Smart Context codeblock. The context builder makes it easy to create larger contexts.

Run:

Smart Context: Insert codeblock

That adds a context block inside the note.

## Context

```ctx
```

3. Open the Context Builder

From the codeblock menu, open the Context Builder.

The Builder is where I add the notes, blocks, existing named contexts, and external sources that should be attached to the note.

4. Add what matters

I add sources based on what the delegated work needs to know.

The useful question is:

What context would help produce the best result?

After adding context, the codeblock shows how much context is attached directly in your note.

5. Name context when it should be reused

If the same context package will help across multiple outcome notes, I save it as a named context.

A named context is useful when it prevents rebuilding the same context again.

Named contexts can also be used to clean-up the context codeblock. For example, if you added 20 context items to the codeblock then it will have 20 lines inside the codeblock. Converting to a named context will replace those 20 lines with a single line.

Related