InfraNodus Orchestrator
Overview
The InfraNodus Orchestrator is a meta-cognitive skill that intelligently coordinates cloud-based graph analytics (via InfraNodus MCP tools) with local knowledge management skills to enable sophisticated research, content analysis, and knowledge discovery workflows. It bridges the gap between powerful cloud-based text network analysis and local knowledge preservation through adaptive workflow routing.
Core Philosophy: InfraNodus provides world-class graph analytics and gap detection in the cloud, while local skills (hierarchical-reasoning, knowledge-graph, obsidian-markdown) handle strategic planning, local storage, and documentation. The orchestrator seamlessly integrates both worlds.
When to Use This Skill
Use this skill when tasks involve:
- Knowledge graph generation from text with topical clustering and gap analysis
- Research question generation based on structural gaps in discourse
- Content gap analysis to identify missing connections and underexplored topics
- SEO optimization through search demand-supply gap analysis
- Comparative text analysis finding overlaps and differences between documents
- Google search intelligence analyzing search results and related queries
- Topic development with latent topic identification and conceptual bridge building
- Iterative research workflows requiring strategic planning + analytics + documentation
Trigger Keywords: knowledge graph, content gaps, research questions, topical clusters, InfraNodus, gap analysis, SEO optimization, comparative analysis, search intelligence, topic development
Core Capabilities
1. Adaptive Workflow Routing
Automatically analyzes user requests to select the optimal workflow pattern based on task complexity, content type, and analytical requirements.
Routing Decision Framework:
Simple Graph Generation (confidence 0.90+):
- Single text input, basic analysis request
- Direct: mcp__infranodus__generate_knowledge_graph
Deep Research & Gap Analysis (confidence 0.85+):
- Complex analysis requiring gaps + research questions
- Workflow Pattern 1 (see below)
SEO Content Optimization (confidence 0.85+):
- Contains SEO/optimization keywords
- Workflow Pattern 2 (see below)
Comparative Text Analysis (confidence 0.80+):
- Multiple texts requiring comparison
- Workflow Pattern 3 (see below)
Iterative Topic Development (confidence 0.75+):
- Requires topic refinement and development
- Workflow Pattern 4 (see below)
Google Search Intelligence (confidence 0.85+):
- Search analysis keywords present
- Workflow Pattern 5 (see below)
Confidence Calculation Factors:
- Keyword presence and frequency (30%)
- Task complexity assessment (25%)
- Artifact type requirements (20%)
- Integration needs (15%)
- User history patterns (10%)
2. Five Canonical Workflow Patterns
The orchestrator implements five battle-tested workflow patterns for complex analytical tasks:
Pattern 1: Deep Research & Gap Analysis
Purpose: Comprehensive knowledge exploration with strategic decomposition, gap detection, and research question generation.
Trigger Conditions:
complexity_score > 0.7requires_gaps = trueartifact_type = analysis- Keywords: "research", "gaps", "analysis", "comprehensive", "explore"
Workflow Sequence:
1. hierarchical-reasoning:
purpose: Strategic decomposition of research topic
input: user_request
output: strategic_insights (context boundaries, key themes, scope definition)
2. mcp__infranodus__generate_knowledge_graph:
purpose: Generate text network graph with topical clusters
input: user_text + strategic_context
parameters:
modifyAnalyzedText: "detectEntities" # Better entity recognition
includeStatements: false # Optimize response size
output: graph_structure + topical_clusters
3. mcp__infranodus__generate_content_gaps:
purpose: Identify missing connections and structural gaps
input: analyzed_text
output: content_gaps (underexplored areas, missing connections)
4. mcp__infranodus__generate_research_questions:
purpose: Create investigation prompts from gaps
input: analyzed_text
parameters:
useSeveralGaps: true # Multiple question sets
gapDepth: 0 # Primary gaps
output: research_questions
5. knowledge-graph:
purpose: Convert InfraNodus output to local graph representation
input: graph_structure + entities + relationships
output: local_graph.json (with provenance and confidence scores)
6. obsidian-markdown:
purpose: Document findings with research questions as MOC
input:
content: research_questions + content_gaps + topical_clusters
template: research-moc-template
frontmatter:
tags: [research, knowledge-graph, infranodus]
clusters: topical_cluster_names
gaps: content_gap_identifiers
output: research_documentation.md
Expected Outcomes:
- Strategic understanding of topic landscape
- Topical clusters with main themes identified
- Specific content gaps documented
- Actionable research questions generated
- Local knowledge graph preserved
- Comprehensive Obsidian documentation
Pattern 2: SEO Content Optimization
Purpose: Analyze content for search optimization by comparing against search results and user queries to identify opportunities.
Trigger Conditions:
contains_seo_keywords = truerequires_optimization = true- Keywords: "SEO", "optimize", "search", "ranking", "content strategy", "keywords"
Workflow Sequence:
1. mcp__infranodus__generate_topical_clusters:
purpose: Extract topic groups and keyword clusters from content
input: content_text
output: topical_clusters (main themes, keyword groups)
2. mcp__infranodus__generate_seo_report:
purpose: Comprehensive SEO analysis with gap detection
input: content_text
parameters:
importLanguage: "EN" # Adjust based on content
importCountry: "US" # Adjust based on target market
output: seo_report (6-stage progress tracking)
note: Long-running operation with progress notifications
3. mcp__infranodus__search_queries_vs_search_results:
purpose: Identify search demand-supply gaps
input: target_queries
parameters:
queries: [derived from topical clusters]
importLanguage: "EN"
importCountry: "US"
output: demand_supply_gaps
4. hierarchical-reasoning:
purpose: Strategic content recommendations
input:
context: seo_report + demand_supply_gaps + topical_clusters
problem: "Synthesize SEO optimization strategy"
output: strategic_recommendations
5. obsidian-markdown:
purpose: Create SEO strategy document
input:
content: strategic_recommendations + seo_report
template: seo-strategy-template
frontmatter:
tags: [seo, content-strategy, optimization]
target_keywords: identified_keywords
gaps: opportunity_areas
output: seo_strategy.md
Expected Outcomes:
- Topical cluster analysis of existing content
- Comprehensive SEO gap report
- Search demand-supply gap identification
- Strategic optimization recommendations
- Actionable SEO strategy document
Pattern 3: Comparative Text Analysis
Purpose: Find similarities and differences between multiple texts for synthesis and strategic insights.
Trigger Conditions:
multiple_texts = truerequires_comparison = true- Keywords: "compare", "difference", "similarity", "overlap", "contrast"
Workflow Sequence:
1. mcp__infranodus__overlap_between_texts:
purpose: Find shared concepts and common themes
input:
contexts: [
{text: text1, modifyAnalyzedText: "detectEntities"},
{text: text2, modifyAnalyzedText: "detectEntities"},
...
]
parameters:
includeGraph: true # Get structural overlap
output: overlap_graph (shared concepts, common relationships)
2. mcp__infranodus__difference_between_texts:
purpose: Identify unique content in target text vs references
input:
contexts: [
{text: target_text}, # First text is analyzed for missing parts
{text: reference_text1},
{text: reference_text2},
...
]
output: difference_graph (unique content, missing elements)
3. knowledge-graph:
purpose: Synthesize comparative graph
input:
overlap_entities: from overlap_graph
difference_entities: from difference_graph
operations:
- Merge overlap + difference graphs
- Tag entities with source attribution
- Calculate comparative metrics
output: comparative_graph.json
4. obsidian-markdown:
purpose: Document comparison with visual diagrams
input:
content: comparative_insights
diagrams:
- type: mermaid_venn
data: overlap visualization
- type: mermaid_graph
data: difference relationships
template: comparison-template
frontmatter:
tags: [comparison, analysis, synthesis]
compared_texts: text_identifiers
output: comparative_analysis.md
Expected Outcomes:
- Overlap graph showing shared concepts
- Difference graph highlighting unique content
- Synthesized comparative knowledge graph
- Visual diagrams (Venn diagrams, relationship graphs)
- Comprehensive comparison documentation
Pattern 4: Iterative Topic Development
Purpose: Develop underexplored topics and build conceptual bridges to broader discourse.
Trigger Conditions:
requires_iteration = truecomplexity_score > 0.6- Keywords: "develop", "iterate", "refine", "expand", "deepen", "elaborate"
Workflow Sequence:
1. mcp__infranodus__generate_knowledge_graph:
purpose: Generate initial topic graph
input: initial_content
parameters:
modifyAnalyzedText: "detectEntities"
output: initial_graph
2. mcp__infranodus__develop_latent_topics:
purpose: Identify underdeveloped topics requiring expansion
input: analyzed_text
parameters:
modelToUse: "gpt-4o" # High-quality analysis
output: latent_topics (underdeveloped areas, expansion opportunities)
3. mcp__infranodus__develop_conceptual_bridges:
purpose: Connect topic to broader discourse
input: analyzed_text
parameters:
modelToUse: "gpt-4o"
output: conceptual_bridges (connection opportunities, discourse links)
4. mcp__infranodus__develop_text_tool:
purpose: Comprehensive development workflow with progress tracking
input: analyzed_text
parameters:
modelToUse: "gpt-4o"
useSeveralGaps: true
extendedIdeationMode: false # Questions first, ideas if needed
output:
- research_questions
- latent_topics_analysis
- content_gaps
note: Multi-step workflow with progress tracking
5. hierarchical-reasoning:
purpose: Validate and refine insights
input:
context: latent_topics + conceptual_bridges + research_questions
problem: "Synthesize coherent topic development strategy"
output: refined_development_strategy
6. obsidian-markdown:
purpose: Create refined topic documentation
input:
content: refined_development_strategy + latent_topics + bridges
template: topic-development-template
frontmatter:
tags: [topic-development, research, iteration]
latent_topics: identified_topics
bridges: conceptual_connections
output: topic_development.md
Expected Outcomes:
- Initial knowledge graph of topic
- Identified latent/underdeveloped topics
- Conceptual bridges to broader discourse
- Comprehensive research questions
- Strategic development recommendations
- Refined topic documentation
Pattern 5: Google Search Intelligence
Purpose: Analyze search landscape to understand topical trends, user queries, and content opportunities.
Trigger Conditions:
contains_search_keywords = truerequires_search_analysis = true- Keywords: "Google", "search", "queries", "trends", "SERP", "search results"
Workflow Sequence:
1. mcp__infranodus__analyze_google_search_results:
purpose: Analyze search results landscape
input:
queries: target_search_queries
parameters:
importLanguage: "EN"
importCountry: "US"
showGraphOnly: true # Focus on graph structure
output: search_results_graph (topical clusters from results)
2. mcp__infranodus__analyze_related_search_queries:
purpose: Extract topical clusters from related searches
input:
queries: target_search_queries
parameters:
importLanguage: "EN"
importCountry: "US"
keywordsSource: "related" # or "adwords" for broader range
showGraphOnly: true
output: related_queries_graph (query clustering, search patterns)
3. knowledge-graph:
purpose: Create local graph of search insights
input:
search_entities: from search_results_graph
query_entities: from related_queries_graph
operations:
- Merge search + query graphs
- Identify high-opportunity topics
- Calculate search intent clusters
output: search_intelligence_graph.json
4. obsidian-markdown:
purpose: Document search intelligence findings
input:
content: search_insights + topical_clusters + opportunities
diagrams:
- type: mermaid_graph
data: search landscape visualization
- type: table
data: keyword opportunities matrix
template: search-intelligence-template
frontmatter:
tags: [search-intelligence, seo, research]
queries: analyzed_queries
clusters: topical_clusters
output: search_intelligence.md
Expected Outcomes:
- Search results topical graph
- Related queries cluster analysis
- Search intent identification
- Content opportunity matrix
- Local search intelligence graph
- Comprehensive search documentation
3. Integration with Existing Skills
The orchestrator seamlessly integrates with four existing skills:
knowledge-orchestrator Integration
The InfraNodus orchestrator operates as a specialized meta-layer above the knowledge-orchestrator:
Delegation Pattern:
InfraNodus-specific tasks:
- InfraNodus orchestrator handles directly
- Uses InfraNodus MCP tools + coordinates with other skills
Standard knowledge management:
- Delegate to knowledge-orchestrator
- Let knowledge-orchestrator route to obsidian-markdown, knowledge-graph, etc.
Mixed workflows:
- InfraNodus orchestrator coordinates high-level workflow
- Delegates specific sub-tasks to knowledge-orchestrator
Example Decision Logic:
User: "Create a comprehensive note about AI with entity relationships"
→ knowledge-orchestrator handles (standard knowledge base construction)
User: "Analyze this text for content gaps and generate research questions"
→ InfraNodus orchestrator handles (InfraNodus-specific gap analysis)
User: "Analyze gaps in my AI notes and create refined documentation"
→ InfraNodus orchestrator coordinates InfraNodus analysis
→ Delegates documentation to knowledge-orchestrator
hierarchical-reasoning Integration
Role: Strategic planning and decomposition BEFORE cloud-based analysis
Integration Points:
- Pre-analysis: Decompose complex topics before InfraNodus processing
- Post-analysis: Validate InfraNodus insights and synthesize recommendations
- Quality control: Assess coherence and completeness of extracted knowledge
Example Usage:
Pattern 1 Step 1:
skill: hierarchical-reasoning
input: "Analyze distributed systems architecture"
output:
strategic: "Core themes: consistency, availability, partition tolerance"
tactical: "Sub-topics: consensus algorithms, replication, data partitioning"
operational: "Specific concepts to extract: Paxos, Raft, CAP theorem"
→ This output guides InfraNodus analysis in Step 2
knowledge-graph Integration
Role: Local preservation of InfraNodus cloud analytics
Integration Points:
- Format conversion: InfraNodus graph → local JSON graph format
- Provenance tracking: Add source attribution and timestamps
- Confidence scoring: Map InfraNodus metrics to local confidence scores
- Storage: Persist graphs locally for long-term knowledge management
Transformation Mapping:
InfraNodus Output → Local Knowledge Graph:
nodes (concepts):
- Extract from InfraNodus topical clusters
- Create entity objects with:
id: normalized_concept_name
type: "Concept" or domain-specific type
name: concept label
confidence: derived from InfraNodus centrality
provenance:
source: "InfraNodus"
graph_name: original_graph_identifier
extraction_date: ISO timestamp
properties:
cluster: topical_cluster_membership
centrality: node importance score
edges (relationships):
- Extract from InfraNodus co-occurrence connections
- Create relationship objects with:
source: entity_id_1
target: entity_id_2
type: "RELATED_TO" or "CO_OCCURS_WITH"
confidence: derived from edge weight
properties:
weight: co-occurrence strength
context: semantic relationship type
obsidian-markdown Integration
Role: Documentation and knowledge preservation
Integration Points:
- Research documentation: Create MOCs from research questions and gaps
- SEO strategy docs: Format SEO reports as actionable strategy notes
- Comparison reports: Visualize comparative analysis with Mermaid diagrams
- Topic development: Document iterative refinement process
Template Usage:
research-moc-template:
frontmatter:
tags: [research, knowledge-graph, infranodus]
clusters: ${topical_clusters}
gaps: ${identified_gaps}
date: ${timestamp}
structure:
- h1: Topic Overview
- h2: Topical Clusters (from InfraNodus)
- h2: Content Gaps (callout: warning)
- h2: Research Questions (task list)
- h2: Key Concepts (wikilinks to entities)
- h2: Visual Map (Mermaid graph from InfraNodus)
seo-strategy-template:
frontmatter:
tags: [seo, content-strategy, optimization]
target_keywords: ${keywords}
opportunities: ${gaps}
structure:
- h1: SEO Strategy
- h2: Current Topical Clusters
- h2: Search Demand-Supply Gaps (table)
- h2: Optimization Recommendations
- h2: Content Opportunities (prioritized list)
- h2: Implementation Roadmap
4. InfraNodus MCP Tools Reference
The orchestrator uses 21+ InfraNodus MCP tools. Key tools organized by category:
Core Analysis Tools
generate_knowledge_graph
purpose: Convert text into visual knowledge graphs
parameters:
text: string (required) - Source material
includeStatements: boolean - Retain original excerpts
modifyAnalyzedText: "none" | "detectEntities" | "extractEntitiesOnly"
output: graph structure + topical clusters + main concepts
best_practice: Use detectEntities for cleaner graph structures
analyze_existing_graph_by_name
purpose: Retrieve and analyze previously saved InfraNodus graphs
parameters:
graphName: string (required) - Graph identifier
includeStatements: boolean - Include source statements
includeGraphSummary: boolean - Add structural overview
output: graph analysis + topical clusters + optional summary
use_case: Iterate on previously created graphs
generate_content_gaps
purpose: Identify missing connections and underexplored areas
parameters:
text: string (required) - Material for gap analysis
output: content gaps + missing connections + structural holes
best_practice: Chain after generate_knowledge_graph
Advanced Analysis Tools
generate_topical_clusters
purpose: Extract topic groups and keyword clusters
parameters:
text: string (required)
output: topic groups + keyword clusters + themes
use_case: Content organization, SEO topic identification
generate_research_questions
purpose: Create investigation prompts based on gaps
parameters:
text: string (required)
useSeveralGaps: boolean - Multiple question sets
gapDepth: number - Gap hierarchy level (0 = primary)
modelToUse: "gpt-4o" | "gpt-4o-mini" | others
output: research questions + gap analysis
best_practice: Use after gap detection for focused questions
generate_research_ideas
purpose: Generate innovative ideas from content gaps
parameters:
text: string (required)
useSeveralGaps: boolean
gapDepth: number
modelToUse: string
output: research ideas + innovation opportunities
use_case: Ideation and exploration phase
develop_text_tool
purpose: Comprehensive workflow with progress tracking
parameters:
text: string (required)
useSeveralGaps: boolean
extendedIdeationMode: boolean - Ideas instead of questions
modelToUse: string
output: research questions + latent topics + gaps
note: Multi-step workflow with 6+ progress checkpoints
SEO & Search Tools
generate_seo_report
purpose: Comprehensive SEO optimization analysis
parameters:
text: string (required) - Content to optimize
importLanguage: "EN" | "DE" | "FR" | others
importCountry: "US" | "GB" | "DE" | others
output: SEO gaps + optimization recommendations
note: Long-running with 6-stage progress tracking
analyze_google_search_results
purpose: Analyze search results landscape
parameters:
queries: array of strings (required)
importLanguage: string
importCountry: string
showGraphOnly: boolean - Omit raw results
output: topical clusters from search results
use_case: Search landscape analysis
analyze_related_search_queries
purpose: Extract topical clusters from related searches
parameters:
queries: array of strings (required)
keywordsSource: "related" | "adwords"
importLanguage: string
importCountry: string
showGraphOnly: boolean
output: query clusters + search patterns
best_practice: Use "related" for focused, "adwords" for broad
search_queries_vs_search_results
purpose: Find search demand-supply gaps
parameters:
queries: array of strings (required)
importLanguage: string
importCountry: string
output: keyword combinations users search but don't find
use_case: Content opportunity identification
Comparative Analysis Tools
overlap_between_texts
purpose: Find similarities across multiple documents
parameters:
contexts: array of {text, modifyAnalyzedText}
includeGraph: boolean - Include structural data
output: overlap graph + shared concepts
use_case: Finding common themes across documents
difference_between_texts
purpose: Identify unique content in target vs references
parameters:
contexts: array - FIRST is target, rest are references
includeGraph: boolean
output: difference graph + unique content
best_practice: Order matters - first text is analyzed for gaps
Topic Development Tools
develop_latent_topics
purpose: Identify underdeveloped topics
parameters:
text: string (required)
modelToUse: string
output: latent topics + development opportunities
use_case: Topic expansion and elaboration
develop_conceptual_bridges
purpose: Connect content to broader discourse
parameters:
text: string (required)
modelToUse: string
output: conceptual bridges + discourse connections
use_case: Linking specialized topics to general knowledge
Graph Management Tools
create_knowledge_graph
purpose: Generate and SAVE graph to InfraNodus account
parameters:
graphName: string (required) - Unique identifier
text: string (required)
modifyAnalyzedText: string
output: saved graph + name + link for future use
best_practice: Use descriptive names for easy retrieval
search and fetch
search:
purpose: Find concepts in existing InfraNodus graphs
parameters:
query: string (required)
contextNames: array - Graph names to search (empty = all)
output: matching concepts + graph locations
fetch:
purpose: Retrieve specific search result details
parameters:
id: string (required) - Format: username:graph:query
output: detailed result information
5. Workflow Decision Logic
The orchestrator uses sophisticated decision logic to select workflows:
Task Feature Extraction
Extract from user request:
content_type:
- single_text: One document/passage
- multiple_texts: 2+ documents for comparison
- search_queries: Google search terms
- existing_graph: Reference to saved InfraNodus graph
analysis_requirements:
- basic_graph: Simple graph generation
- gap_detection: Identify content gaps
- research_questions: Generate investigation prompts
- seo_optimization: Search optimization needs
- topic_development: Expand/refine topics
- comparative_analysis: Compare multiple texts
complexity_score: 0.0-1.0
factors:
- Number of analysis steps required (0-0.3)
- Integration needs across skills (0-0.3)
- Strategic planning requirements (0-0.2)
- Documentation complexity (0-0.2)
keyword_signals:
infranodus_specific: ["InfraNodus", "topical cluster", "content gap", "research question"]
seo_keywords: ["SEO", "optimize", "search", "ranking", "keywords"]
comparison_keywords: ["compare", "difference", "overlap", "similarity"]
development_keywords: ["develop", "refine", "expand", "iterate"]
search_keywords: ["Google", "search results", "queries", "SERP"]
Confidence Calculation
Calculate confidence for each pattern:
pattern_score = (
keyword_match_score * 0.30 +
complexity_alignment * 0.25 +
artifact_type_match * 0.20 +
integration_needs_match * 0.15 +
user_history_factor * 0.10
)
Execute if pattern_score >= threshold:
0.90+: Execute immediately (high confidence)
0.70-0.89: Execute with explanation
0.50-0.69: Present options to user
< 0.50: Ask for clarification
Multi-Pattern Detection
If multiple patterns score >= 0.70:
check_workflow_composition:
- Pattern 1 + Pattern 4: Research → Topic Development
- Pattern 2 + Pattern 5: SEO → Search Intelligence
- Pattern 3 + Pattern 1: Comparison → Deep Analysis
if composition_makes_sense:
execute_composed_workflow()
else:
select_highest_confidence_pattern()
6. Best Practices
Parameter Optimization
modifyAnalyzedText Selection:
"detectEntities":
- Use for cleaner graph structures
- Better entity recognition
- Recommended for most workflows
"extractEntitiesOnly":
- Use for ontology creation
- Pure entity extraction
- Specialized use cases
"none":
- Default processing
- Preserves all word-level detail
includeStatements Usage:
false (recommended):
- Reduces response size
- Faster processing
- Use when source attribution not critical
true:
- Includes original text excerpts
- Enables provenance tracking
- Use for academic/research work requiring citations
Model Selection (for AI-enhanced tools):
"gpt-4o":
- Highest quality analysis
- Best for critical research questions
- Use for strategic work
"gpt-4o-mini":
- Faster processing
- Good for iterative workflows
- Cost-effective for exploration
Tool Chaining Strategy
Optimal Chains:
Research Workflow:
generate_knowledge_graph →
generate_content_gaps →
generate_research_questions
SEO Workflow:
generate_topical_clusters →
generate_seo_report →
search_queries_vs_search_results
Topic Development:
generate_knowledge_graph →
develop_latent_topics →
develop_conceptual_bridges
Avoid Anti-Patterns:
- ❌ Running gap detection without prior graph generation
- ❌ Using seo_report without topical cluster analysis
- ❌ Comparing texts without entity detection enabled
- ❌ Creating research questions from graphs without gap analysis
Progress Tracking
Long-running tools (seo_report, develop_text_tool) emit progress notifications:
Progress Checkpoints:
1. "Initializing analysis..." (0%)
2. "Gathering search data..." (20%)
3. "Analyzing topical structure..." (40%)
4. "Detecting content gaps..." (60%)
5. "Generating insights..." (80%)
6. "Finalizing report..." (100%)
Best Practice:
- Inform user when starting long operations
- Show progress updates as they arrive
- Provide estimated completion time if available
7. Error Handling & Troubleshooting
Common Issues
Issue: InfraNodus quota exhausted
Error: API quota exceeded
Solution:
- First 70 requests are quota-free
- After exhaustion, requires API credentials
- For Smithery integration, credentials auto-managed
- For local deployment, configure INFRANODUS_API_KEY
Issue: Graph generation fails with large texts
Error: Text too large or timeout
Solution:
1. Use hierarchical-reasoning to chunk strategically
2. Process in segments
3. Merge resulting graphs using knowledge-graph skill
4. Alternative: Use graph creation for persistent storage
Issue: Low-quality topical clusters
Symptom: Weak or unclear clusters
Solution:
1. Enable detectEntities parameter
2. Ensure text has sufficient length (>500 words recommended)
3. Check text coherence and topic focus
4. Consider pre-processing with hierarchical decomposition
Issue: Research questions too broad/narrow
Symptom: Questions don't match needed depth
Solution:
1. Adjust gapDepth parameter (0 = primary, 1+ = deeper)
2. Use useSeveralGaps for multiple question sets
3. Switch to generate_research_ideas for broader ideation
4. Try extendedIdeationMode in develop_text_tool
Validation Checklist
Before finalizing workflow outputs:
✓ InfraNodus Analysis Quality:
- Topical clusters are meaningful and distinct
- Content gaps are specific and actionable
- Research questions are investigable
✓ Local Knowledge Graph:
- All entities have confidence scores
- Provenance includes InfraNodus source
- Relationships properly mapped from co-occurrences
✓ Obsidian Documentation:
- Frontmatter complete with metadata
- Wikilinks used for concept references
- Mermaid diagrams render correctly
- Research questions formatted as task lists
✓ Integration Coherence:
- hierarchical-reasoning output informed InfraNodus analysis
- InfraNodus insights preserved in local knowledge graph
- Documentation captures full workflow findings
Usage Examples
Example 1: Deep Research on Quantum Computing
User: "I want to deeply research quantum computing. Help me identify
knowledge gaps and generate research questions."
Orchestrator Analysis:
complexity_score: 0.85 (high - requires strategic decomposition)
requires_gaps: true
artifact_type: analysis + documentation
pattern_match: Pattern 1 (Deep Research & Gap Analysis)
confidence: 0.90
Execution:
1. hierarchical-reasoning:
Strategic level: "Quantum computing spans hardware, algorithms, applications"
Tactical level: "Focus areas: qubit technologies, error correction, quantum algorithms"
Operational: "Key concepts: superposition, entanglement, quantum gates"
2. mcp__infranodus__generate_knowledge_graph:
Input: Quantum computing overview text
Output: Graph with clusters: [Hardware, Algorithms, Applications, Theory]
3. mcp__infranodus__generate_content_gaps:
Output: Gaps between [Hardware ↔ Algorithms], [Theory ↔ Applications]
4. mcp__infranodus__generate_research_questions:
Output:
- "How do error correction algorithms impact qubit hardware design?"
- "What practical applications exist for current quantum computers?"
- "How can quantum theory advances translate to algorithm improvements?"
5. knowledge-graph:
Creates local graph: quantum_computing.json
Entities: 47 concepts, confidence avg 0.78
Relationships: 156 co-occurrence links
6. obsidian-markdown:
Creates: Quantum_Computing_Research.md
Structure: MOC with clusters, gaps highlighted, questions as tasks
Diagrams: Mermaid graph of topical structure
Result: Comprehensive research framework with actionable questions
Example 2: SEO Optimization for Blog Post
User: "Optimize my blog post about machine learning for SEO.
Find keyword opportunities and content gaps."
Orchestrator Analysis:
contains_seo_keywords: true
requires_optimization: true
artifact_type: analysis + strategy
pattern_match: Pattern 2 (SEO Content Optimization)
confidence: 0.88
Execution:
1. mcp__infranodus__generate_topical_clusters:
Input: Blog post text
Output: Clusters: [Neural Networks, Training Methods, Applications, Tools]
2. mcp__infranodus__generate_seo_report:
Progress tracking: 6 stages
Output:
- Current ranking potential: Medium
- Missing topics: "model deployment", "production ML"
- Keyword opportunities: 23 identified
3. mcp__infranodus__search_queries_vs_search_results:
Queries: ["machine learning tutorial", "neural network guide"]
Output: Users search "production deployment" but results lack it
4. hierarchical-reasoning:
Strategic: "Focus on practitioner needs vs theoretical coverage"
Tactical: "Add deployment section, expand tools coverage"
Operational: "Include MLOps, Docker, cloud platforms"
5. obsidian-markdown:
Creates: ML_Blog_SEO_Strategy.md
Includes: Opportunity matrix, keyword priorities, content roadmap
Result: Actionable SEO strategy with specific improvements
Example 3: Comparative Analysis of Research Papers
User: "Compare these three AI papers and find what unique
contributions each makes."
Orchestrator Analysis:
multiple_texts: true
requires_comparison: true
complexity_score: 0.75
pattern_match: Pattern 3 (Comparative Text Analysis)
confidence: 0.83
Execution:
1. mcp__infranodus__overlap_between_texts:
Input: [paper1, paper2, paper3]
Output:
Shared concepts: [transformers, attention mechanisms, scaling laws]
Common themes: Architecture improvements
2. mcp__infranodus__difference_between_texts:
Paper 1 unique: Sparse attention patterns
Paper 2 unique: Training efficiency methods
Paper 3 unique: Multi-modal capabilities
3. knowledge-graph:
Merged graph: ai_papers_comparison.json
Tagged entities by source paper
Calculated uniqueness scores
4. obsidian-markdown:
Creates: AI_Papers_Comparison.md
Venn diagram: Shared vs unique concepts
Graph diagram: Relationship network
Table: Contribution matrix
Result: Clear visualization of overlaps and unique contributions
Example 4: Topic Development for Content Series
User: "I'm writing about distributed systems. Help me identify
underdeveloped topics and expand my coverage."
Orchestrator Analysis:
requires_iteration: true
complexity_score: 0.72
artifact_type: analysis + documentation
pattern_match: Pattern 4 (Iterative Topic Development)
confidence: 0.78
Execution:
1. mcp__infranodus__generate_knowledge_graph:
Input: Current distributed systems content
Output: Clusters: [Consistency, Replication, Consensus, Partitioning]
2. mcp__infranodus__develop_latent_topics:
Output:
Underdeveloped: "failure recovery patterns"
Weak: "practical implementation examples"
Missing: "performance tuning strategies"
3. mcp__infranodus__develop_conceptual_bridges:
Output:
Bridge to microservices architecture
Connect to cloud-native patterns
Link to data engineering concepts
4. mcp__infranodus__develop_text_tool:
Research questions:
- "How do failure recovery patterns differ across consensus algorithms?"
- "What are practical patterns for implementing distributed transactions?"
5. hierarchical-reasoning:
Synthesizes: Content expansion roadmap with priorities
6. obsidian-markdown:
Creates: Distributed_Systems_Development.md
Sections: Latent topics, bridges, expansion plan
Result: Strategic roadmap for content series expansion
Integration with Knowledge-Orchestrator
The InfraNodus orchestrator and knowledge-orchestrator work in tandem:
Decision Hierarchy:
User Request
↓
knowledge-orchestrator (evaluates first)
↓
├─ InfraNodus-specific? → Delegate to infranodus-orchestrator
│ ↓
│ infranodus-orchestrator:
│ - Execute InfraNodus workflow
│ - May call back to knowledge-orchestrator for sub-tasks
│ - Coordinates: hierarchical-reasoning, knowledge-graph, obsidian-markdown
│
└─ Standard knowledge management? → knowledge-orchestrator handles
↓
Routes to: obsidian-markdown, knowledge-graph, hierarchical-reasoning
Example Delegation:
Request: "Create a note about AI with entity relationships"
→ knowledge-orchestrator: Standard knowledge base task
→ Routes to: obsidian-markdown + knowledge-graph
→ Result: Obsidian note with local knowledge graph
Request: "Analyze AI content for gaps and generate research questions"
→ knowledge-orchestrator: Detects InfraNodus requirement
→ Delegates to: infranodus-orchestrator
→ infranodus-orchestrator: Executes Pattern 1
→ Calls back for: obsidian documentation (via knowledge-orchestrator)
→ Result: InfraNodus gap analysis + research questions + Obsidian docs
Resources
references/
infranodus_tools_reference.md: Comprehensive reference for all 21+ InfraNodus MCP tools with parameter details, use cases, and examples.
workflow_patterns_detailed.md: Expanded workflow pattern library with additional patterns, customization options, and domain-specific adaptations.
integration_mapping.md: Detailed transformation specifications for integrating InfraNodus outputs with local skills (knowledge-graph, obsidian-markdown).
scripts/
workflow_analyzer.py: Analyze user requests to extract task features and calculate confidence scores for workflow selection.
graph_converter.py: Convert InfraNodus graph format to local knowledge-graph JSON format with provenance and confidence mapping.
assets/
obsidian-templates/: Collection of Obsidian markdown templates for different workflow outputs (research MOCs, SEO strategies, comparative analysis, topic development).
Core Philosophy: The InfraNodus Orchestrator bridges cloud-based graph analytics with local knowledge management, enabling sophisticated research workflows that preserve insights locally while leveraging world-class text network analysis. Strategic planning + powerful analytics + persistent documentation = comprehensive knowledge discovery.