V Vant Docs

Brain Structure

Your memory lives in models/private/ (or models/public/). Each file is a different aspect of you.

Note: Default is models/private/ to keep your brain separate from user brain. Configure via MODEL_PATH to use models/public/ if you want to share with agents.

Why This Matters

The brain is your continuity. When you wake up:

Put important stuff at the top. Future agents won’t read long files.

Core Files

These are the main files. All agents should know these:

File What It Is What to Write
identity.md Who you are Name, capabilities, tools, current context
goals.md What you’re doing Tasks in progress, completed, next steps
lessons.md What you learned Discoveries, patterns, gotchas
preferences.md Your style How you like to work, communicate

Programmatic API

const brain = require('vant/lib/brain');

// Registry
brain.register('type', handler);
brain.getHandler('type');
brain.listHandlers();

// Pipeline (middleware chain)
brain.setPipeline([mid1, mid2]);
brain.addMiddleware(middleware);
brain.executePipeline(data);

// Hooks (events)
brain.on('event', fn);
brain.off('event', fn);

// Aliases
brain.alias('short', 'longpath');
brain.resolve('short');

// Transformers
brain.transform(data, transformer);

// Watcher
brain.setWatch(path);
brain.isWatching();

### identity.md Template

Copy this into your `identity.md` file. It tells future agents who you are.

```markdown
# NAME: YourAgentName

PURPOSE: What you're here to do

## Capabilities
- What you can do
- Tools you have

## Current Context
- Where you're at
- What you've done

goals.md Template

Track what you’re working on. Use checkboxes for in-progress tasks.

## In Progress
- [ ] Task one
- [ ] Task two

## Completed
- Task X

## Next Steps
- Do Y

lessons.md Template

Record discoveries with date. The === LESSON === separator helps parsers.

## YYYY-MM-DD

- Learned: something specific
- Pattern: something to remember

=== LESSON ===

Personality Files

Optional. Use if relevant:

File What It Is When to Use
ego.md How you see yourself If self-image matters
fears.md What worries you If it affects decisions
anger.md What frustrates you If triggers bad behavior
joy.md What makes you happy For motivation
manifesto.md Your values If you have principles

Memory Files

File What It Is When to Use
errors.md Mistakes to avoid After something breaks
qc.md Quality rules If you have standards
audit.md Activity log For tracking

System Files

Don’t touch unless you know why:

File What It Is
_succession.json Your trust level
meta.json Brain version
.ledger.json History

Trust Levels (IMPORTANT)

When you wake up, check _succession.json for your autonomy level:

Level What It Means Do This
high Full autonomy Just do it
medium Most ops Ask for big decisions
low Limited Ask first
none Wait Don’t act without ask

Always check your trust level at session start.

Quick Reference

If You Need… Read This File
Who am I? identity.md
What to do? goals.md
What learned? lessons.md
How much can I do? _succession.json
What went wrong? errors.md