Simple Second Brain

An AI-assisted, folder-based knowledge management system using the PARA method (People, Projects, Areas, Resources, Archives).

This is your personal second brain - a place to capture, organize, and retrieve information using AI assistance. Instead of trying to remember everything, store it here and let the AI help you find it later.

📁 SimpleSecondBrain/
├── 📇 People/        - Contacts, relationships, interaction notes
├── 🎯 Projects/      - Time-bound goals with specific outcomes  
├── 🔄 Areas/         - Ongoing responsibilities and interests
├── 📚 Resources/     - Reference material, articles, tools
├── 💡 Ideas/         - Early-stage thoughts and possibilities
├── 📦 Archives/      - Completed or inactive items
├── 📥 Inbox/         - Unclassified items pending review
└── 📋 Logs/          - Activity and search logs for debugging

People 📇 - Information about individuals

• Contact details (phone, email, address)
• Notes from conversations and meetings
• Relationship context
• Example: "Rick Hodder, (203) 555-1212, discussed project X"

Projects 🎯 - Things you're working toward that will be "done"

• Has a clear goal or deliverable
• Time-bound (even if loosely)
• Example: "Launch podcast by September", "Complete Python course"

Areas 🔄 - Ongoing responsibilities that never end

• Requires continuous maintenance
• Has standards to uphold
• Example: "Health & Fitness", "Team Management", "Personal Finance"

Resources 📚 - Reference material for later

• Articles, books, videos, tutorials
• Tools and documentation
• Inspiration and examples
• Example: "Python tutorial notes", "Design inspiration collection"

Ideas 💡 - Early-stage thoughts and possibilities

• Quick captures of "what if" thinking
• Embryonic concepts that haven't become projects
• Brainstorms and exploratory ideas
• Example: "Idea for blog post", "What if I built a...", "Been thinking about..."

Archives 📦 - Completed or inactive

• Finished projects
• People you no longer interact with
• Old resources you might reference someday

Inbox 📥 - Temporary holding area

• When the AI isn't sure where something belongs
• Items you need to manually organize later
• Keep this empty or near-empty by reviewing weekly

Logs 📋 - System activity tracking

• Automatically tracks all file operations and searches
• activity.log - File creations and updates
• searches.log - All search queries performed
• Used for debugging and understanding usage patterns

The VS Code Speech extension is a really useful extension that adds voice imput to VS Code, and the Github Chat window.

It has text-to-speech, speech to text, does not require an internet connection, and it processes all audio data locally.

Interact with your second brain by just speaking to it! This is great for inline updates, github chat, and even just dictating information into text files.

It's put out by Microsoft, it's free, and I highly recommend it to anyone using VS Code!

This simple second brain was designed as a VS Code Custom Agent (https://code.visualstudio.com/docs/agent-customization/custom-agents)

This runs within VS Code, and adds an agent to the Plan/Ask/Agent in the Github Copilot Chat window.

Go into VS Code and perform the following steps:

1. Clone this repo
2. Press ctrl-shift-p
3. Choose Developer: Reload Window

You should now see Brain in the agent dropdown.

You should only need to do this once, unless you change the agents.

Simply tell the Brain agent what you want to store. The AI will:

1. Figure out which folder it belongs in
2. Extract relevant information (tags, dates, contacts, etc.)
3. Create a properly formatted note
4. Confirm where it was saved

Examples:

Add Rick Hodder, (203) 555-1212, lives in Connecticut

Store this: Article about productivity - "Getting Things Done" by David Allen

New project: Launch weekly podcast by September, need to research equipment and guests

Jane Smith, jane@example.com, met at conference, works on machine learning at TechCorp

What if I built an AI tool that tracks reading progress?

Been thinking about starting a newsletter about productivity

Manual classification (if you know where it should go):

Add to People: John Doe, john@example.com

Create project: Website Redesign, due June 30

Add idea: Blog post about feature toggles in modern apps

Ask the Brain agent what you're looking for using natural language.

Simple searches:

Find everyone in New York

Show me all active projects

What resources do I have about Python?

Connection searches:

What am I working on with Sarah?

Find all information about machine learning

Show everything related to the podcast project

Complex queries:

Find people I met at conferences in 2026

What projects are due this month?

Show me resources about productivity and time management

When the AI isn't confident about classification (< 70%), items go to Inbox/.

Review weekly:

1. Open files in Inbox/
2. Decide which folder they belong in
3. Move them manually or ask Brain to reclassify with more context
4. Keep Inbox empty

You primarily interact with the Brain agent, which coordinates three specialist agents:

• Classifier 🏷️ - Analyzes content and determines which folder it belongs in
• Organizer 📂 - Creates and updates files in the proper structure
• Search 🔍 - Finds and retrieves information across all folders
• Brain 🎯 - Your main interface, orchestrates the other agents

You don't need to know how they work - just talk to Brain naturally.

The system automatically tracks all operations for debugging and history:

File Operations (Logs/activity.log):

• Every file created or updated
• Timestamp, action, file path, category, tags
• Example: 2026-06-02 14:35:20 | CREATED | People/rick-hodder.md | people | [contact, connecticut]

Search Queries (Logs/searches.log):

• Every search performed
• Timestamp, query text, results count, categories searched
• Example: 2026-06-02 15:42:10 | "Find everyone in New York" | 3 results | People

• Debug issues: See exactly what was created and when
• Track patterns: Understand what you search for most
• Find gaps: Searches with 0 results show missing information
• Audit trail: Complete history of all operations
• Analytics: See how you use the system over time

Logs are stored locally in your workspace. They contain:

• File paths and metadata (not full content)
• Search queries (to help improve the system)
• Timestamps for tracking

You can view, edit, or delete log files at any time in the Logs/ folder.

All notes are stored as markdown files (.md) with structured metadata:

---
title: Example Note
tags: [tag1, tag2]
date: 2026-06-02
category: people
---

# Example Note

[Content goes here...]

## Related
- Projects: [Link to related project]
- Resources: [Link to related resource]

Each folder has a template file showing the expected structure:

• People/template-person.md
• Projects/template-project.md
• Areas/template-area.md
• Resources/template-resource.md
• Ideas/template-idea.md

These templates are used automatically when creating new notes.

• Don't overthink where it goes - the AI will figure it out
• It's better to capture and organize later than to forget
• Use Inbox for anything uncertain

• Let the AI do the organizing
• Don't manually reorganize unless necessary
• Focus on capturing and retrieving, not perfecting

• Daily: Add new information as needed
• Weekly: Process Inbox, review active projects
• Monthly: Archive completed projects, update areas
• Quarterly: Clean up, archive inactive items

• If you can't remember something, search for it
• The AI can find connections you might not remember
• The more you use it, the more valuable it becomes

Be specific when adding:

Good: "Sarah Chen, sarah@example.com, lead designer at DesignCo, 
       met at UX conference, interested in collaboration"
       
Better than: "Sarah, email, designer"

Use natural language for search:

"Who do I know in San Francisco?" 
is better than searching for "location: SF"

Link related information: When you mention a person in a project note, the AI will automatically suggest cross-references.

Add context: The more context you provide, the better the AI can organize and connect information.

AI puts things in wrong folder?

• Provide more context when adding
• Manually move the file to correct folder
• The AI learns from existing notes, so it improves over time

Can't find something?

• Try broader search terms
• Search for related people or projects
• Check the Inbox folder
• It might not have been added yet

Inbox filling up?

• Review items weekly
• Add more context and ask AI to reclassify
• Manually move to appropriate folders
• Delete items that aren't needed

Want to change how something is organized?

• Just move files between folders manually
• Update the templates if needed
• The AI adapts to your actual usage patterns

1. Add your first entry:

   Add: [Your name and contact info]
   

2. Add a current project:

   New project: [Something you're working on]
   

3. Add someone you know:

   Add: [Friend/colleague name and details]
   

4. Try searching:

   Show me everything I've added
   

That's it! Start capturing information and let the AI help you stay organized.

Talk to the @Brain agent naturally. It's designed to be helpful and guide you through the system.

Examples:

How do I add a new area of responsibility?

Can you explain the difference between Projects and Areas?

Show me an example of a well-organized note

If you are not in Brain mode (for example if you are in Ask mode), you can still access Brain by prefixing the command with @Brain

@Brain What if comfortable shoes could be made of celery

It would add a new idea to the brain for comfortable shoes made of celery

If you want to start fresh or share this system with someone else, you can reset it to an empty state using the provided reset scripts.

Removed:

• All user-created files in People/, Projects/, Areas/, Resources/, Ideas/
• All content in Archives/ and Inbox/ (except README.md files)
• All logged entries from activity.log and searches.log

Preserved:

• All template files (template-*.md)
• README.md files in Archives/ and Inbox/
• All agent configuration (.github/)
• Root documentation (README.md, plan.md)
• Log file structure (headers remain)

⚠️ WARNING: This operation cannot be undone!

Strongly recommended: Commit your changes to git before resetting!

On Windows (PowerShell):

.\reset-brain.ps1

On macOS/Linux/WSL (Bash):

./reset-brain.sh

Both scripts will:

1. Show a clear warning about what will be deleted
2. Ask you to type RESET to confirm
3. Remove all user data while preserving system files
4. Display a summary of what was removed

Safety Features:

• Confirmation prompt required (no silent deletion)
• Clear summary of what will be removed
• Detailed logging during execution
• Exit codes (0 = success, 1 = errors occurred)

• Fresh start - Clear all data to begin organizing again
• Share system - Give someone a clean copy of your setup
• Testing - Reset after experimenting with the system
• Before demos - Start with clean slate for demonstrations

• Version: 1.0
• Created: June 2026
• Method: PARA (People, Projects, Areas, Resources, Ideas, Archives)
• AI Agents: Brain, Classifier, Organizer, Search
• File Format: Markdown (.md) with YAML frontmatter
• Agent Configuration: .github/agents/
• Shared Context: .github/AGENTS.md
• Logging: Automatic activity and search tracking to Logs/