Sources
Raw materials — the input layer of your Second Brain.
What Goes Here
- Articles and blog posts (saved as markdown)
- Academic papers (markdown summary + PDF reference)
- Transcripts (meeting notes, podcast transcripts)
- Screenshots and exported data (converted to markdown)
- Any external information you want the agent to synthesize
Rules
- Agent reads, never writes. Files here are immutable once created.
- Every file needs frontmatter. Copy
_template.mdas a starting point. - Source status tracks processing:
unread → reading → processed. - File naming: use lowercase, hyphens, descriptive. Example:
2024-ai-agents-survey.md.
Frontmatter
---
type: source
title: "Descriptive Title"
created: YYYY-MM-DD
updated: YYYY-MM-DD
source_url: "https://..."
author: "Name or Organization"
status: unread | reading | processed
tags: [tag1, tag2]
---
Workflow
- Drop a source file here.
- Run
/ingest(Claude Code) or tell your agent to process it. - The agent reads the source, updates status, creates wiki pages.
- Status moves to
processedwhen fully synthesized.