==================
/// MCP ///
/// TFM ///
==================
[server:online]
[protocol:ready]tfmcp
by nwiizo
Terraform Model Context Protocol (MCP) CLI/server that lets AI assistants analyse, plan and apply Terraform with enterprise-grade security.
328
20
Open Source01
mcp
Launch tfmcp as a Model Context Protocol (MCP) server
02
analyze
Analyze Terraform configurations and provide validation and best-practice recommendations
03
help
Display help information for tfmcp or one of its subcommands
04
resources/list
Return a list of Terraform resources available in the current project via the MCP API
05
prompts/list
Return a list of available prompts/templates that can be used through the MCP API
06
set_terraform_directory
Change the active Terraform project directory at runtime
Installation
1. Prerequisites
• Rust tool-chain (stable, v1.74 or newer)
• Terraform CLI installed and on your PATH (v1.6+ recommended)
• A Unix-like shell or PowerShell
2. Install the CLI from crates.io
cargo install tfmcp The installer will download, compile and place the binary under `~/.cargo/bin/tfmcp`.
3. Verify installation
tfmcp --help
tfmcp --version4. (Optional) Enable Claude Desktop integration
Set your Claude session token so the tool can ask Claude for plan reviews:
export CLAUDE_SESSION_TOKEN="<your_token>"5. Initialise a Terraform workspace and let tfmcp create its context file
terraform init
tfmcp context init # writes .tfmcp.toml6. Typical workflow
tfmcp read-config . # parse *.tf files
tfmcp plan . # run and store plan JSON
tfmcp apply . # apply with safety checks
tfmcp state push state.tfstate7. Upgrading
cargo install tfmcp --forceDocumentation
License: MIT License
Updated 7/30/2025