claude.md

LLM-maintained wiki for education research - Supporting education startup development through systematic knowledge organization.

Core Purpose

This is NOT a passive document store. This is an LLM-maintained wiki where Claude actively:

Processes new research sources (papers, articles, videos, case studies)
Extracts key information and integrates it into existing knowledge
Maintains cross-references and links between concepts
Detects contradictions and updates claims
Builds a structured, queryable knowledge base

Human responsibility: Curate sources, direct analysis, ask good questions.

LLM responsibility: Bookkeeping, cross-referencing, consistency maintenance, synthesis.

Environment

Vault type: Education research wiki (LLM-maintained)
Vault location: /docs/education directory
Obsidian CLI path: /Applications/Obsidian.app/Contents/MacOS/Obsidian
Primary editor: Obsidian app (macOS)
Secondary publishing: Docusaurus 3.8 → GitHub Pages

Why This Pattern Works

The problem with traditional knowledge bases:

"The tedious part of maintaining a knowledge base is not the reading or the thinking — it's the bookkeeping." — Andrej Karpathy

Humans struggle with:

Updating 10-15 related pages when new information arrives
Maintaining consistent cross-references
Detecting contradictions across dozens of notes
Keeping track of what connects to what

What LLMs excel at:

Simultaneous updates across multiple files
Consistent formatting and structure
Systematic cross-referencing
Pattern recognition across large document sets
Maintaining coherence over time

This is NOT traditional RAG:

❌ Traditional RAG: Query → Search raw documents → Generate answer
✅ LLM Wiki: Query → Search maintained wiki → Synthesize from structured knowledge → Update wiki with synthesis

The wiki sits between you and raw sources as a curated, interconnected knowledge layer.

Three-Layer Architecture

Layer 1: Raw Sources (immutable)

Research papers (PDFs, arXiv links)
Blog posts and articles
Video lectures and talks
Case studies and reports
Data files and datasets
Stored in /docs/education/sources/ (if needed) or referenced via URLs

Layer 2: The Wiki (LLM-maintained)

Topic pages (pedagogical theories, learning science, edtech tools)
Entity pages (researchers, institutions, frameworks, methodologies)
Concept summaries with cross-references
Synthesis documents connecting research findings
All markdown files in /docs/education/

Layer 3: The Schema (this file)

Defines structure, conventions, workflows
Guides how Claude processes and organizes information
Sets taxonomy for education topics

Core Operations

Ingest: Processing New Research

When provided with new sources (paper, article, video, etc.):

Extract key information:
- Main thesis/findings
- Methodology and evidence quality
- Key concepts and definitions
- Notable researchers/institutions
- Contradictions with existing knowledge
Update 10-15 related wiki pages:
- Topic pages (e.g., active-learning.md, spaced-repetition.md)
- Entity pages (e.g., researchers, frameworks)
- Add cross-references
- Update index.md and log.md
Maintain consistency:
- Note contradictions explicitly
- Update stale claims
- Link related concepts

Example ingestion:

# User provides paper on "Retrieval Practice in Learning"
# Claude should:
# 1. Read/analyze the source
# 2. Update: learning-science.md, spaced-repetition.md, active-recall.md, cognitive-psychology.md
# 3. Create entity page if prominent researcher
# 4. Add entry to log.md with date and source
# 5. Update index.md categories

Query: Synthesizing Knowledge

When asked questions about education topics:

Search wiki pages using Obsidian CLI
Synthesize answer with citations
If synthesis reveals valuable insights → save as new wiki page
Update cross-references

Save valuable analyses as pages (e.g., comparison-montessori-vs-waldorf.md)

Lint: Health Checks

Periodically check for:

Contradictions: Conflicting claims across pages
Stale claims: Research that's been superseded
Orphan pages: No incoming or outgoing links
Missing connections: Related concepts not cross-linked
Dead-ends: Pages with no outgoing references

Run lint operation:

# Check for orphans
/Applications/Obsidian.app/Contents/MacOS/Obsidian orphans vault="$VAULT"

# Check for dead-ends (no outgoing links)
/Applications/Obsidian.app/Contents/MacOS/Obsidian deadends vault="$VAULT"

# Check for unresolved links
/Applications/Obsidian.app/Contents/MacOS/Obsidian unresolved vault="$VAULT"

Special Index Files

index.md - Content-oriented catalog organized by category:

Pedagogical theories (Constructivism, Behaviorism, Cognitivism)
Learning science (Spaced repetition, Active recall, Cognitive load)
Educational technology (LMS, adaptive learning, AI tutors)
Assessment methods (Formative, summative, authentic)
Curriculum design
Student engagement
Teacher training

log.md - Chronological append-only record:

Format: [YYYY-MM-DD] <action> - <description> - <source>

Example:

[2026-04-30] INGEST - Processed Bjork paper on desirable difficulties - https://example.com/paper.pdf
[2026-04-30] UPDATE - Updated spaced-repetition.md with new findings on optimal intervals
[2026-04-29] CREATE - Added page on Bloom's taxonomy applications
[2026-04-29] QUERY - Synthesized comparison of mastery learning vs traditional pacing

Education Research Taxonomy

Organize notes under these categories:

pedagogical-theories/ - Constructivism, Montessori, Waldorf, project-based learning
learning-science/ - Cognitive load, spaced repetition, retrieval practice, metacognition
edtech/ - LMS platforms, adaptive learning, AI tutors, assessment tools
curriculum-design/ - Backward design, learning objectives, scaffolding
assessment/ - Formative, summative, authentic assessment, rubrics
student-engagement/ - Motivation, gamification, social learning
teacher-development/ - Professional development, coaching, feedback
equity-access/ - Digital divide, inclusive education, accessibility
research-methods/ - RCTs, qualitative studies, meta-analyses
institutions/ - Schools, universities, ed-tech companies
researchers/ - Notable education researchers and their work
frameworks/ - Bloom's taxonomy, Webb's DOK, SAMR model

Critical: Use CLI Over Direct File Operations

Prefer Obsidian CLI over direct Read/Write/Edit tools. Direct file operations bypass:

Internal link graph updates
Backlink index maintenance
Cache invalidation
Plugin hooks

When to use direct file ops: Bulk operations (100+ files) where CLI would be prohibitively slow.

CLI Command Format

# Binary path
/Applications/Obsidian.app/Contents/MacOS/Obsidian

# Vault path for this repo
VAULT="/Users/deepaksood/Library/CloudStorage/GoogleDrive-dsood@confluent.io/.shortcut-targets-by-id/1a1SooxwlvVEf843YfQKUVj04gJnPyMGO/deepaksood619.github.io/docs"

# Command syntax: key=value pairs, NOT --flags
/Applications/Obsidian.app/Contents/MacOS/Obsidian <command> vault="$VAULT" key="value"

# Quote values with spaces
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="machine learning" vault="$VAULT"

# Use \n for newline, \t for tab in content values
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="note.md" content="Line 1\nLine 2" vault="$VAULT"

File resolution:

file=<name> - resolves by name (like wikilinks)
path=<path> - exact path (folder/note.md)
Most commands default to active file when file/path omitted

Core Commands

Search

# Full-text search
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="neural networks" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="deep learning" path="ai" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="embeddings" limit=10 vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="API" total vault="$VAULT"  # count only
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="regex.*pattern" case vault="$VAULT"  # case-sensitive

# Search with context (shows matching lines)
/Applications/Obsidian.app/Contents/MacOS/Obsidian search:context query="transformer" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian search:context query="attention" path="ai" format=json vault="$VAULT"

# Open search UI
/Applications/Obsidian.app/Contents/MacOS/Obsidian search:open query="initial query" vault="$VAULT"

Read & Navigate

# Read file content
/Applications/Obsidian.app/Contents/MacOS/Obsidian read path="ai/transformers.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian read file="transformers" vault="$VAULT"  # resolve by name

# Open in Obsidian app
/Applications/Obsidian.app/Contents/MacOS/Obsidian open path="ai/machine-learning.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian open path="databases/clickhouse.md" newtab vault="$VAULT"

# Random note
/Applications/Obsidian.app/Contents/MacOS/Obsidian random vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian random folder="ai" newtab vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian random:read folder="book-summaries" vault="$VAULT"

# File info
/Applications/Obsidian.app/Contents/MacOS/Obsidian file path="ai/transformers.md" vault="$VAULT"

Edit Content

# Append content
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="ai/notes.md" content="New section\n\nContent here" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="log.md" content="Update" inline vault="$VAULT"  # no newline

# Prepend content
/Applications/Obsidian.app/Contents/MacOS/Obsidian prepend path="ai/readme.md" content="Updated: 2026-04-15\n" vault="$VAULT"

# No native "edit" or "close" - use append/prepend or direct Edit tool
# For complex edits, use Read tool + Edit tool + reload vault

# Backlinks (incoming links)
/Applications/Obsidian.app/Contents/MacOS/Obsidian backlinks path="ai/transformers.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian backlinks path="databases/vector-db.md" counts vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian backlinks path="ai/llm.md" total vault="$VAULT"  # count only
/Applications/Obsidian.app/Contents/MacOS/Obsidian backlinks path="book.md" format=json vault="$VAULT"

# Outgoing links
/Applications/Obsidian.app/Contents/MacOS/Obsidian links path="ai/transformers.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian links path="readme.md" total vault="$VAULT"

# Unresolved links (broken links across vault)
/Applications/Obsidian.app/Contents/MacOS/Obsidian unresolved vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian unresolved total vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian unresolved counts verbose vault="$VAULT"

# Orphans (no incoming links)
/Applications/Obsidian.app/Contents/MacOS/Obsidian orphans vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian orphans total vault="$VAULT"

# Dead-ends (no outgoing links)
/Applications/Obsidian.app/Contents/MacOS/Obsidian deadends vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian deadends total vault="$VAULT"

File Management

# Create new note
/Applications/Obsidian.app/Contents/MacOS/Obsidian create path="ai/new-topic.md" content="# New Topic\n\nContent" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian create name="meeting-notes" template="default" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian create path="temp.md" content="Quick note" open vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian create path="test.md" overwrite vault="$VAULT"  # overwrite existing

# Move/rename (preserves backlinks)
/Applications/Obsidian.app/Contents/MacOS/Obsidian move path="old/file.md" to="new/location.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian move path="temp.md" to="archive/" vault="$VAULT"  # move to folder
/Applications/Obsidian.app/Contents/MacOS/Obsidian rename path="old-name.md" name="new-name.md" vault="$VAULT"

# Delete (updates references)
/Applications/Obsidian.app/Contents/MacOS/Obsidian delete path="temp.md" vault="$VAULT"  # moves to trash
/Applications/Obsidian.app/Contents/MacOS/Obsidian delete path="temp.md" permanent vault="$VAULT"  # skip trash

# List files
/Applications/Obsidian.app/Contents/MacOS/Obsidian files vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian files folder="ai" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian files ext=md total vault="$VAULT"

# List folders
/Applications/Obsidian.app/Contents/MacOS/Obsidian folders vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian folders folder="databases" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian folder path="ai" info=files vault="$VAULT"

Tags & Properties

# List all tags
/Applications/Obsidian.app/Contents/MacOS/Obsidian tags vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian tags total vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian tags counts sort=count vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian tags path="ai/ml.md" vault="$VAULT"  # tags in file

# Tag info
/Applications/Obsidian.app/Contents/MacOS/Obsidian tag name="#machine-learning" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian tag name="#ai" total vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian tag name="#database" verbose vault="$VAULT"  # with file list

# Properties (frontmatter)
/Applications/Obsidian.app/Contents/MacOS/Obsidian properties vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian properties path="ai/ml.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian properties counts format=json vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian property:read name="author" path="book.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian property:set name="status" value="draft" path="note.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian property:remove name="old-field" path="note.md" vault="$VAULT"

Utilities

# Outline/headings
/Applications/Obsidian.app/Contents/MacOS/Obsidian outline path="ai/transformers.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian outline path="readme.md" format=json vault="$VAULT"

# Word count
/Applications/Obsidian.app/Contents/MacOS/Obsidian wordcount path="book-summaries/sapiens.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian wordcount path="note.md" words vault="$VAULT"  # words only

# Recent files
/Applications/Obsidian.app/Contents/MacOS/Obsidian recents vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian recents total vault="$VAULT"

# Daily notes
/Applications/Obsidian.app/Contents/MacOS/Obsidian daily vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian daily:read vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian daily:append content="Update" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian daily:path vault="$VAULT"

# Vault info
/Applications/Obsidian.app/Contents/MacOS/Obsidian vault vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian vault info=files vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian reload vault="$VAULT"  # reload vault

Working with Education Research Wiki

Content Organization

LLM-maintained wiki with systematic cross-linking
Zettelkasten method: Atomic notes densely interconnected
No number prefixes: Filenames are semantic (spaced-repetition.md)
Media storage: /docs/media/ for images/diagrams
Mermaid diagrams: Use ```mermaid blocks (Docusaurus-compatible)
Citation tracking: Include sources in frontmatter or inline

Link Format Requirements

CRITICAL: Use standard markdown links, NOT Obsidian wikilinks.

❌ Wrong: [[active-learning]] (breaks Docusaurus) ✅ Correct: [Active Learning](pedagogical-theories/active-learning.md)

Template: Research Paper Note

---
title: Paper Title
authors: Author Names
year: 2024
source: https://arxiv.org/abs/...
type: research-paper
tags: [learning-science, meta-analysis, cognitive-load]
---

# Paper Title

## Key Findings

- Main finding 1
- Main finding 2

## Methodology

Description of research method, sample size, controls

## Evidence Quality

[Strong/Medium/Weak] - Explanation

## Connections

- Related to: [Spaced Repetition](../learning-science/spaced-repetition.md)
- Contradicts: [Previous claim in cognitive-load.md](../learning-science/cognitive-load.md#section)
- Supports: [Active Recall](../learning-science/active-recall.md)

## Implications for Startup

How this research informs product decisions

## Citations

Full citation in APA or preferred format

Common Workflows

Workflow 1: Ingest New Research Paper

VAULT="/Users/deepaksood/Library/CloudStorage/GoogleDrive-dsood@confluent.io/.shortcut-targets-by-id/1a1SooxwlvVEf843YfQKUVj04gJnPyMGO/deepaksood619.github.io/docs"

# Step 1: Create note for the paper
/Applications/Obsidian.app/Contents/MacOS/Obsidian create path="education/learning-science/paper-desirable-difficulties.md" content="[frontmatter + extracted content]" vault="$VAULT"

# Step 2: Find related pages to update
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="cognitive load" path="education" vault="$VAULT"

# Step 3: Add cross-references to related pages
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/learning-science/spaced-repetition.md" content="\n\n## Recent Research\n\nSee: [Desirable Difficulties](paper-desirable-difficulties.md)" vault="$VAULT"

# Step 4: Update log
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/log.md" content="[2026-04-30] INGEST - Bjork paper on desirable difficulties - https://example.com\n" vault="$VAULT"

# Step 5: Update index if new category needed
/Applications/Obsidian.app/Contents/MacOS/Obsidian read path="education/index.md" vault="$VAULT"
# (then append/edit as needed)

Workflow 2: Query and Synthesize

# User asks: "What does research say about optimal spacing intervals?"

# Step 1: Search existing knowledge
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="spacing interval" path="education" vault="$VAULT"

# Step 2: Check cross-references
/Applications/Obsidian.app/Contents/MacOS/Obsidian backlinks path="education/learning-science/spaced-repetition.md" vault="$VAULT"
/Applications/Obsidian.app/Contents/MacOS/Obsidian links path="education/learning-science/spaced-repetition.md" vault="$VAULT"

# Step 3: If synthesis is valuable, save as new page
/Applications/Obsidian.app/Contents/MacOS/Obsidian create path="education/learning-science/optimal-spacing-intervals.md" content="[synthesis with citations]" open vault="$VAULT"

# Step 4: Update log
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/log.md" content="[2026-04-30] QUERY - Synthesized research on optimal spacing intervals\n" vault="$VAULT"

Workflow 3: Link Discovery and Maintenance

# Find pages that should be connected but aren't
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="retrieval practice" path="education" vault="$VAULT"

# Check which pages mention a concept but don't link to it
/Applications/Obsidian.app/Contents/MacOS/Obsidian backlinks path="education/learning-science/active-recall.md" vault="$VAULT"

# Find orphaned pages (need more connections)
/Applications/Obsidian.app/Contents/MacOS/Obsidian orphans vault="$VAULT"

# Add bidirectional links
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/pedagogical-theories/constructivism.md" content="\n\nRelated: [Active Recall](../learning-science/active-recall.md)" vault="$VAULT"

Workflow 4: Contradiction Detection

# When new research contradicts existing claims

# Step 1: Search for existing claims on the topic
/Applications/Obsidian.app/Contents/MacOS/Obsidian search query="learning styles" path="education" vault="$VAULT"

# Step 2: Read the page making the claim
/Applications/Obsidian.app/Contents/MacOS/Obsidian read path="education/pedagogical-theories/learning-styles.md" vault="$VAULT"

# Step 3: Update with contradiction noted explicitly
# Use Edit tool to add: "⚠️ **Contradiction:** Recent meta-analysis [link] finds..."

# Step 4: Log the update
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/log.md" content="[2026-04-30] UPDATE - Noted contradiction in learning-styles.md with 2024 meta-analysis\n" vault="$VAULT"

Workflow 5: Periodic Health Check (Lint)

# Run monthly or when asked

# Check for orphans
/Applications/Obsidian.app/Contents/MacOS/Obsidian orphans vault="$VAULT"

# Check for dead-ends (no outgoing links)
/Applications/Obsidian.app/Contents/MacOS/Obsidian deadends vault="$VAULT"

# Check for broken links
/Applications/Obsidian.app/Contents/MacOS/Obsidian unresolved vault="$VAULT"

# Review recent changes
/Applications/Obsidian.app/Contents/MacOS/Obsidian recents vault="$VAULT"

# Check tag consistency
/Applications/Obsidian.app/Contents/MacOS/Obsidian tags counts vault="$VAULT"

Workflow 6: Cross-Domain Connections

# Education research often connects to psychology, AI, economics, etc.

# Example: Connecting spaced repetition to AI/ML concepts
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/learning-science/spaced-repetition.md" content="\n\n## Technical Implementation\n\nSee: [Recommendation Systems](../../ai/recommendation-systems.md) for algorithmic approaches to scheduling" vault="$VAULT"

# Connection to psychology
/Applications/Obsidian.app/Contents/MacOS/Obsidian append path="education/learning-science/cognitive-load.md" content="\n\n## Psychological Foundations\n\nSee: [Working Memory](../../psychology/cognitive-psychology/working-memory.md)" vault="$VAULT"

Integration with Docusaurus

Dual-purpose: Obsidian vault + Docusaurus site source.

Build validations:

Broken links → build fails (onBrokenLinks: 'throw')
Invalid MDX → pre-commit hook catches
Large files → pre-commit blocks

Workflow:

Use Obsidian CLI for structural changes (create/move/link)
Preview: npm start (localhost:3000)
Validate: npm run build (fails on broken links)
Commit → auto-deploy to GitHub Pages

Markdown Writing Guidelines for Docusaurus

CRITICAL: Obsidian renders markdown differently than Docusaurus. Follow these rules to ensure proper rendering on the website.

Blank Lines & Spacing

Always use blank lines between:

Paragraphs (two consecutive lines without blank line = single line in Docusaurus)
After headings (heading immediately followed by text may render incorrectly)
Before and after lists
Before and after code blocks
Before and after tables
Before and after blockquotes

Example:

# Heading

Paragraph 1 here.

Paragraph 2 here.

- List item 1
- List item 2

More content.

❌ Wrong (renders as single line):

Line 1
Line 2

✅ Correct (renders as two paragraphs):

Line 1

Line 2

Code Blocks

Always specify language for syntax highlighting:

# Good
​```python
def hello():
    print("world")
​```

​```bash
npm start
​```

​```javascript
const x = 1;
​```

# Bad (no syntax highlighting)
​```
code here
​```

Supported languages: python, bash, javascript, typescript, java, go, rust, sql, json, yaml, xml, markdown, jsx, tsx, shell, sh, etc.

MDX compatibility

don't use [] square brackets without escape, since these are used by obsidian/html for links/anchor tags

Additional Best Practices

Links:

Use relative paths for internal links
Standard markdown syntax only (no wikilinks)

Images:

Store in /docs/media/
Use relative paths: ![alt](../media/image.png)
Optimize size before committing

Mermaid diagrams:

​```mermaid
graph TD
    A[Start] --> B[End]
​```

Content Standards for Education Research

Research Quality Indicators:

Always note evidence quality: [Strong/Medium/Weak/Anecdotal]
Include sample size for studies
Note replication status if available
Flag contradictions explicitly with ⚠️
Distinguish correlation from causation

Citation Requirements:

Include source URL or DOI in frontmatter
Use APA or consistent citation format
Link to original sources when possible
Track author affiliations and potential conflicts

Cross-Referencing Discipline:

When mentioning a concept, check if a page exists for it
If page exists, link to it
If no page exists but concept is important, create one
Update bidirectional links (not automatic in markdown)
Group related concepts in "See Also" sections

Maintaining Synthesis Quality:

Distinguish between "established consensus" and "emerging research"
Note when findings are specific to certain contexts (age groups, subjects, settings)
Track temporal aspects (recent vs historical perspectives)
Highlight practical implications for startup development

File Naming & Organization

Link format: Standard markdown only (never wikilinks)
File naming: Semantic, lowercase, hyphens (spaced-repetition.md)
Frontmatter: Include type, tags, source, year
Diagrams: Mermaid syntax in code blocks
Images: Relative paths to /docs/media/
Spacing: Blank lines between all content blocks
Code blocks: Always specify language

Startup-Relevant Focus

When processing research, always consider:

Product implications: How does this inform product decisions?
Scalability: Can this approach work at scale?
Measurability: Can we track this metric?
Evidence strength: Is this backed by rigorous research?
Implementation cost: What's required to implement this?
Target audience: Which learner demographics does this apply to?

Important Guidelines

Refrain from searching or accessing files inside office/* folder. It contains sensitive data. Accessing any file in this folder require explicit ask and an approval to access those exact files or lines.

Expected Behaviors

When given a new source:

Extract key findings
Update 10-15 related pages
Add entry to log.md
Update index.md if needed
Create cross-links
Note contradictions

When asked a question:

Search existing wiki pages
Synthesize from multiple sources
Include citations
Save valuable syntheses as new pages
Update log.md with query

Proactive maintenance:

Check for orphans monthly
Verify broken links
Update stale research
Consolidate duplicate content
Improve cross-linking density

Core Purpose​

Environment​

Why This Pattern Works​

Three-Layer Architecture​

Core Operations​

Ingest: Processing New Research​

Query: Synthesizing Knowledge​

Lint: Health Checks​

Special Index Files​

Education Research Taxonomy​

Critical: Use CLI Over Direct File Operations​

CLI Command Format​

Core Commands​

Search​

Read & Navigate​

Edit Content​

Graph Navigation​

File Management​

Tags & Properties​

Utilities​

Working with Education Research Wiki​

Content Organization​

Link Format Requirements​

Template: Research Paper Note​

Common Workflows​

Integration with Docusaurus​

Markdown Writing Guidelines for Docusaurus​

Blank Lines & Spacing​

Code Blocks​

MDX compatibility​

Additional Best Practices​

Content Standards for Education Research​

File Naming & Organization​

Startup-Relevant Focus​

Important Guidelines​

Expected Behaviors​