Available Tools

Complete reference for the 3 MCP tools provided by the Terraform MCP Server.

---

🛠️ Tool Summary

Tool	Description
terraform_execute	Secure execution of Terraform commands with validation
terraform_doc_search	Semantic similarity search over Terraform documentation
ingest_terraform_docs	Document ingestion with vector embeddings

---

1. terraform_execute

Secure execution of Terraform commands with enterprise-grade security and validation.

Supported Commands

Command	Description	Auto-Approve
init	Initialize Terraform configuration	No
plan	Preview infrastructure changes	No
validate	Validate configuration syntax	No
apply	Apply infrastructure changes	Yes
destroy	Destroy infrastructure	Yes

Parameters

Parameter	Type	Required	Default	Description
command	string	Yes	-	Terraform command to execute
working_directory	string	Yes	-	Directory containing .tf files
variables	dict	No	None	Terraform variables to pass
aws_region	string	No	None	AWS region for execution
strip_ansi	boolean	No	True	Remove ANSI color codes
timeout	integer	No	300	Execution timeout (1-1800 seconds)

Security Features

Feature	Description
Command Whitelisting	Only allowed commands can be executed
Directory Traversal Protection	Blocks paths containing ..
Pattern Detection	Scans for 100+ dangerous patterns
Variable Limits	Maximum 100 variables per execution
Timeout Enforcement	Automatic process termination
Output Sanitization	10,000 character limit

Usage Examples

Initialize Terraform:

"Initialize Terraform in /tmp/terraform-project"

Plan with Variables:

"Run terraform plan in /tmp/project with environment=production, instance_count=3, and AWS region us-west-2"

Apply Configuration:

"Apply the terraform configuration in /tmp/project with instance_type=t3.micro"

Validate Configuration:

"Validate the terraform configuration in /tmp/project"

Response Structure

{
  "success": true,
  "data": {
    "command": "terraform plan",
    "status": "success",
    "result": {
      "return_code": 0,
      "stdout": "Terraform will perform...",
      "stderr": "",
      "execution_time": 2.45
    },
    "metadata": {
      "terraform_version": "1.5.0",
      "aws_region": "us-west-2",
      "variables_count": 2,
      "security_checks_passed": true
    }
  }
}

---

2. terraform_doc_search

Semantic similarity search over ingested Terraform documentation using vector embeddings.

Parameters

Parameter	Type	Required	Default	Description
query	string	Yes	-	Search query (1-1000 chars)
top_k	integer	No	5	Number of results (1-50)
similarity_threshold	float	No	0.7	Minimum similarity (0.0-1.0)
node_types	list	No	None	Document types to search

Node Types

Type	Description	Index Name
resource	Terraform resource documentation	docchunk_resource_embedding_hnsw
data_source	Terraform data source docs	docchunk_datasource_embedding_hnsw
best_practice	Best practices and guidelines	docchunk_bestpractice_embedding_hnsw

Search Features

Feature	Description
HNSW Index	Fast similarity search
Cosine Similarity	Semantic matching scoring
Multi-Type Search	Search across multiple doc types
Threshold Filtering	Configurable precision/recall
Result Distribution	Even distribution across types

Usage Examples

Search Resources:

"Find AWS S3 bucket configuration examples"

Search Data Sources:

"Search for VPC data source configuration"

Search Best Practices:

"What are the best practices for Terraform state management?"

Multi-Type Search:

"Search for EC2 instance configuration in resources and best practices"

Response Structure

{
  "success": true,
  "data": {
    "query": "AWS EC2 instance configuration",
    "results_count": 3,
    "results": [
      {
        "content": "resource \"aws_instance\" \"example\" {...}",
        "similarity_score": 0.89,
        "node_type": "resource",
        "id": "docchunk_001"
      }
    ],
    "search_parameters": {
      "top_k": 5,
      "similarity_threshold": 0.7
    },
    "service_info": {
      "provider": "openai",
      "model": "text-embedding-ada-002",
      "dimensions": 1536
    }
  }
}

---

3. ingest_terraform_docs

Sophisticated document processing system for ingesting Terraform documentation into the knowledge graph.

Parameters

Parameter	Type	Required	Default	Description
filter_types	list	Yes	-	Document types to ingest
filter_services	list	No	None	AWS services to filter
scan_dirs	list	No	["docs/"]	Directories to scan

Filter Types

Type	Description	Processing
resource	Terraform resource docs	Structured chunking
data_source	Data source documentation	Structured chunking
terraform	Both resources and data sources	Structured chunking
best_practice	Best practices documents	LLM extraction
readme	README files	Standard chunking

Ingestion Features

Feature	Description
Multi-Format Support	HTML, Markdown, PDF
Intelligent Discovery	Index-based and directory scanning
LLM Extraction	AI-powered content structuring
Semantic Chunking	Metadata preservation
Incremental Processing	Skip already ingested docs
Neo4j Storage	Graph database with vector indexes

Usage Examples

Ingest Resources and Data Sources:

"Ingest AWS provider resources and data sources"

Ingest Best Practices:

"Ingest Terraform best practices documentation"

Ingest READMEs:

"Ingest README files from the project"

Filter by Service:

"Ingest only EC2 and S3 resources"

Response Structure

{
  "success": true,
  "data": {
    "ingestion_summary": {
      "total_documents": 150,
      "successful": 148,
      "failed": 2,
      "skipped": 0
    },
    "types_processed": ["resource", "data_source"],
    "chunks_created": 1250,
    "embeddings_generated": 1250
  }
}

---

Document Processing Pipeline

Stage	Description
Discovery	Parse index files, scan directories
Detection	Identify HTML, Markdown, PDF
Extraction	Load and extract content
Chunking	Create semantic chunks with metadata
Embedding	Generate vector embeddings
Storage	Store in Neo4j with vector indexes

---

Best Practices

For terraform_execute

1. Use Specific Directories: Always use valid working directories
2. Set Appropriate Timeouts: Increase for complex operations
3. Review Variables: Ensure no dangerous patterns
4. Check Return Codes: Verify successful execution

For terraform_doc_search

1. Use Specific Queries: More specific yields better results
2. Adjust Thresholds: Higher for precision, lower for recall
3. Filter by Type: Focus on relevant document types
4. Iterate: Refine queries based on results

For ingest_terraform_docs

1. Start Small: Begin with specific services
2. Monitor Progress: Check ingestion logs
3. Use Incremental: Leverage existing ingestion state
4. Verify Quality: Check ingested content quality

---

Next Steps

• 📖 Examples - Usage patterns and workflows
• ⚙️ Configuration - Server configuration