Obsidian Vault Intelligence via MCP

When to Use This Skill

Invoke this skill when you need to:

•Search user's Obsidian vault for project context, decisions, or strategic information
•Retrieve specific notes about companies, projects, or technical approaches
•Query vault for patterns, themes, or cross-referenced information
•Create or update notes with new insights or action items
•Find backlinks to understand note relationships
•Extract knowledge for research, planning, or decision-making

Core Principles

•Progressive Search: Start broad with intelligent_search, narrow with specific tools
•Context Preservation: Capture full note content when relevant for decision-making
•Cross-Reference Intelligence: Use backlinks to understand note relationships
•Smart Queries: Use natural language for vault queries, specific patterns for search
•Note Hygiene: When creating/updating, follow existing vault conventions

Available MCP Tools

Primary Tools

•query_vault: Natural language queries about vault contents (best for: "what do I know about X?")
•intelligent_search: Advanced search with link analysis and tag hierarchies (best for: finding related notes)
•search_notes: Keyword/phrase search in filenames or content (best for: finding specific terms)
•get_note: Retrieve full note content by path (best for: reading specific notes)
•get_backlinks: Find all notes linking to a target note (best for: understanding relationships)

Secondary Tools

•write_note: Overwrite note with new content (use sparingly - destructive)
•create_note: Create new note with frontmatter (requires title, path)
•append_to_note: Add content to end of existing note (best for: adding updates)
•update_note_section: Update specific heading section (best for: targeted updates)
•list_directories: List vault structure (best for: understanding organization)

Advanced Tools

•guided_path: Generate narrative tour through linked notes from a seed note
•audit_recent_notes: Find recently modified notes missing required structure
•contextual_companions: Suggest related notes based on topic or seed note
•fresh_energy: Find recent notes lacking backlinks (integration opportunities)
•initiative_bridge: Identify initiative-tagged notes with outstanding tasks
•pattern_echo: Find notes reusing specific phrasing or patterns
•synthesis_ready: Flag note clusters that need synthesis/summary note

Search Strategy Patterns

Pattern 1: Discovery (What do I know about X?)

code

Use: mcp__obsidian__query_vault
Query: "What do I know about Partner Project's protein interaction prediction technology?"

Result: Natural language summary with relevant context

When to use: Starting point for unfamiliar topics, broad context gathering

Pattern 2: Specific Term Search

code

Use: mcp__obsidian__search_notes
SearchTerm: "protein-protein interaction"
SearchType: "both" (filename and content)

Result: List of notes containing the exact term

When to use: Finding notes mentioning specific technical terms, company names, or concepts

Pattern 3: Intelligent Context Search

code

Use: mcp__obsidian__intelligent_search
Query: "LightForge Works email marketing campaign with Clay and Attio"

Result: Ranked results with link analysis and structural context

When to use: Multi-concept searches, finding related notes across different areas

Pattern 4: Relationship Mapping

code

# Step 1: Find seed note
Use: mcp__obsidian__search_notes
SearchTerm: "Partner Project Master Command Center"

# Step 2: Get backlinks
Use: mcp__obsidian__get_backlinks
NotePath: "Areas/Companies/Partner Project/Partner Project Master Command Center.md"

Result: All notes linking to command center (related projects, meetings, action items)

When to use: Understanding how a topic connects to other work, finding related contexts

Pattern 5: Recent Activity Analysis

code

Use: mcp__obsidian__fresh_energy
HoursBack: 48
MinWords: 80
Limit: 10

Result: Recently updated substantive notes that lack integration (backlinks)

When to use: Weekly reviews, finding recent work that needs cross-referencing

Common Use Case Workflows

Use Case: Strategic Context for Meeting

Goal: Prepare for Partner Project partnership discussion

markdown

Step 1: Query for overview
mcp__obsidian__query_vault("Partner Project partnership status, technology details, and strategic next steps")

Step 2: Get command center details
mcp__obsidian__get_note(notePath: "Areas/Companies/Partner Project/Partner Project Master Command Center.md")

Step 3: Find related action items
mcp__obsidian__search_notes(searchTerm: "Partner Project", searchType: "content")
Filter results for notes in "next_up/Action Items/"

Step 4: Check recent meeting notes
mcp__obsidian__intelligent_search("Partner Project meetings 2025")

Step 5: Get market research prompt
mcp__obsidian__get_note(notePath: "Areas/Companies/Partner Project/market_research/Partner Project Market Analysis Research Prompt.md")

Use Case: Project Discovery

Goal: Find all information about a specific project

markdown

Step 1: Intelligent search for project
mcp__obsidian__intelligent_search("LightForge Works micro-application development business")

Step 2: Get project note
mcp__obsidian__get_note(notePath: "[path from search results]")

Step 3: Find related notes
mcp__obsidian__get_backlinks(notePath: "[project note path]")

Step 4: Check for recent updates
mcp__obsidian__fresh_energy(hoursBack: 168)  # Last week
Filter for project-related notes

Step 5: Find action items
mcp__obsidian__search_notes(searchTerm: "@paia", searchType: "content")
Filter for project-specific paia tags

Use Case: Research Synthesis

Goal: Extract knowledge for market research report

markdown

Step 1: Query for thematic overview
mcp__obsidian__query_vault("What research have I done on pharmaceutical drug discovery market and protein interaction prediction competitors?")

Step 2: Pattern search for specific insights
mcp__obsidian__pattern_echo(snippet: "AlphaFold accuracy")

Step 3: Find synthesis opportunities
mcp__obsidian__synthesis_ready(minClusterSize: 3)
Identify if research notes need a synthesis document

Step 4: Get detailed notes
For each relevant note from steps 1-3:
  mcp__obsidian__get_note(notePath: "[note path]")

Step 5: Create synthesis note
mcp__obsidian__create_note(
  notePath: "Areas/Companies/Partner Project/market_research/Competitive Landscape Synthesis.md",
  title: "Competitive Landscape Synthesis",
  content: "[synthesized content]"
)

Use Case: Task Management Integration

Goal: Find outstanding tasks across projects

markdown

Step 1: Find initiative-tagged tasks
mcp__obsidian__initiative_bridge(
  initiative: "Partner Project",
  frontmatterField: "project"
)

Step 2: Search for @paia tags
mcp__obsidian__search_notes(searchTerm: "@paia-updater", searchType: "content")

Step 3: Check recent notes for unprocessed items
mcp__obsidian__audit_recent_notes(
  hoursBack: 72,
  requiredFields: ["status", "priority"]
)

Step 4: Get specific action plan
mcp__obsidian__get_note(notePath: "next_up/Action Items/[specific action plan].md")

Use Case: Knowledge Connection

Goal: Find related notes to create strategic connections

markdown

Step 1: Find contextual companions for a topic
mcp__obsidian__contextual_companions(
  topic: "multi-agent AI coordination",
  limit: 5
)

Step 2: Generate narrative tour
mcp__obsidian__guided_path(
  notePath: "Areas/Tech/Development/Claude Code Multi-Agent Patterns.md",
  supportingLimit: 3,
  counterpointLimit: 2
)

Step 3: Update note with cross-references
mcp__obsidian__update_note_section(
  notePath: "[target note]",
  sectionHeading: "Related Concepts",
  newContent: "[connections from steps 1-2]"
)

Note Operations Best Practices

Creating Notes

code

Use: mcp__obsidian__create_note

Requirements:
- notePath: Full path relative to vault root
- title: Clear, descriptive title
- content: Well-structured markdown
- tags: Relevant tags array (optional)

Pattern:
notePath: "Areas/[Category]/[Subcategory]/[Title].md"
title: "Descriptive Title"
tags: ["tag1", "tag2"]
content: |
  # Title

  ## Overview
  Context and purpose

  ## Key Points
  - Point 1
  - Point 2

  ## Related
  - [[Related Note 1]]
  - [[Related Note 2]]

Updating Notes

code

PREFER: mcp__obsidian__update_note_section
Over: write_note (destructive) or append_to_note (less precise)

Pattern:
1. Read current note: get_note
2. Identify target section heading
3. Update only that section: update_note_section

Example:
notePath: "Areas/Companies/Partner Project/Partner Project Master Command Center.md"
sectionHeading: "Next Steps"
newContent: |
  - Execute comprehensive market research
  - Schedule follow-up with partner contact
  - Prepare partnership proposal

Appending Updates

code

Use: mcp__obsidian__append_to_note
When: Adding new information to end of note (meeting updates, action items)

Pattern:
notePath: "[existing note path]"
content: |

  ## Update [Date]
  New information or action items
  - Action 1
  - Action 2

Vault Structure Intelligence

user's vault organization (common paths):

code

Areas/
  ├── Companies/              # Client and partnership companies
  │   ├── Partner Project/
  │   ├── Company B/
  │   └── LightForge/
  ├── Tech/                   # Technical notes and development
  │   ├── Development/
  │   └── Tools/
  └── Projects/               # Active projects

next_up/
  ├── Action Items/           # Actionable tasks
  ├── Opportunities/          # Business opportunities
  │   └── Active/
  └── Research/               # Ongoing research

people/                       # Contact and relationship notes

daily/                        # Daily notes (date-based)

Search Query Optimization

Natural Language Queries (query_vault)

Good: "What are my next steps for the Partner Project partnership?" Better: "What are my next steps for Partner Project partnership development, including market research tasks and strategic priorities?"

Keyword Searches (search_notes)

Good: "LightForge" Better: "LightForge Works" (more specific) Best: Use searchType: "both" to search filenames and content

Intelligent Searches (intelligent_search)

Good: "protein prediction" Better: "protein-protein interaction prediction market pharmaceutical" Best: Include context words that appear in related notes

Integration with Other Skills

With coordinating-sub-agents

code

Coordinator Claude workflow:
1. Query vault for strategic context: mcp__obsidian__query_vault
2. Get specific notes: mcp__obsidian__get_note
3. Delegate task to sub-agent with context
4. After completion: append_to_note with results

With syncing-task-completions

code

Sync workflow:
1. Search for task notes: mcp__obsidian__search_notes(searchTerm: "TODO")
2. For each project, find action items: intelligent_search
3. Update master task list: update_note_section
4. Append completion dates: append_to_note

With processing-paia-tags

code

PAIA processing workflow:
1. Search for @paia tags: search_notes(searchTerm: "@paia")
2. Get full note content: get_note
3. Analyze cross-references: get_backlinks
4. Execute @paia instruction
5. Update note status: update_note_section

With learning-from-outcomes

code

Learning workflow:
1. Query vault for past similar situations: query_vault
2. Find pattern echoes: pattern_echo
3. Create synthesis: create_note
4. Link to related notes: update_note_section with [[links]]

Performance Optimization

Minimize API Calls

Bad: Multiple get_note calls for discovery

code

get_note(path1)
get_note(path2)
get_note(path3)

Good: Use search/query to identify, then get_note only what you need

code

search_results = search_notes("Partner Project")
# Filter in memory to top 2 most relevant
get_note(top_result_1)
get_note(top_result_2)

Batch Related Operations

Bad: Search, read, search, read Good: Search all, read all relevant

Cache Context

Bad: Re-query vault for same information Good: Query once, use results for entire workflow

Error Handling

python

# When note path might not exist
try:
    note = mcp__obsidian__get_note(notePath: "Areas/Companies/Unknown/Note.md")
except:
    # Fall back to search
    results = mcp__obsidian__search_notes(searchTerm: "Unknown Company", searchType: "both")

# When search returns no results
results = mcp__obsidian__search_notes(searchTerm: "obscure-term")
if not results or len(results) == 0:
    # Try broader query
    results = mcp__obsidian__query_vault("What do I know related to obscure-term?")

Vault Query Checklist

Before querying Obsidian vault:

• Search Strategy: Have I chosen the right tool (query_vault vs search_notes vs intelligent_search)?
• Query Clarity: Is my query specific enough to get relevant results?
• Breadth-First: Am I starting broad (search/query) before narrow (get_note)?
• Relationship Mapping: Do I need backlinks to understand context?
• Result Filtering: Can I filter results programmatically to reduce API calls?
• Update Plan: If I'm updating notes, am I using update_note_section over write_note?
• Cross-References: Should I link this work to related notes?

Common Patterns Summary

Goal	Primary Tool	Secondary Tool	Notes
"What do I know about X?"	query_vault	intelligent_search	Start here for discovery
Find specific term	search_notes	grep content	Use searchType: "both"
Related notes	get_backlinks	contextual_companions	Understand relationships
Read specific note	get_note	-	Only after identifying path
Recent activity	fresh_energy	audit_recent_notes	Weekly review pattern
Create new note	create_note	-	Use full paths
Update existing	update_note_section	append_to_note	Prefer section updates
Find clusters	synthesis_ready	intelligent_search	Identify synthesis needs
Task tracking	initiative_bridge	search_notes @paia	Project management

Obsidian Vault as Strategic Intelligence

Think of the Obsidian vault as user's:

•Strategic Memory: Past decisions, approaches, learnings
•Project Context: Current initiatives, next steps, blockers
•Relationship Map: Connections between concepts, projects, people
•Task System: @paia tags, action items, initiative tracking
•Research Archive: Market analysis, competitive intelligence, technical explorations

Use it proactively to:

•Inform delegation decisions with past context
•Avoid re-solving already-solved problems
•Understand user's preferences and approaches
•Connect new work to existing strategic priorities
•Capture learnings and insights for future sessions