OpenTofu MCP Server
Registry search for OpenTofu providers, modules and docs. Also offered as a hosted endpoint.
By defaultRead-only
With our presetSame: read-only is built in
Flagsegress
Labels are provisional (from the project README/docs, checked Oct 27, 2026). Not yet tested in our sandbox.
Read-only switch
Registry/documentation lookup only.
Preset
Read-onlyThe docs say it only reads, so there's no flag to add. Not sandbox-tested yet.
- Matched by name only. OpenTofu MCP Server is documented as read-only, but a fork or another server with the same name might not be. So Codex still prompts and Claude Code asks.
Claude Code
.mcp.json
{
"mcpServers": {
"opentofu": {
"command": "npx",
"args": ["@opentofu/opentofu-mcp-server"]
}
}
}.claude/settings.json
{
"permissions": {
"ask": ["mcp__opentofu"]
}
}ask makes Claude Code prompt before every tool from this server.
Codex
~/.codex/config.toml
# Generated by redpresets.dev on 2026-11-03. Unverified preset. Also use a read-only credential. Not a security guarantee.
[mcp_servers.opentofu]
command = "npx"
args = ["@opentofu/opentofu-mcp-server"]
default_tools_approval_mode = "prompt"default_tools_approval_mode = "prompt" asks before every tool call.
Cursor
.cursor/mcp.json
{
"mcpServers": {
"opentofu": {
"type": "stdio",
"command": "npx",
"args": ["@opentofu/opentofu-mcp-server"]
}
}
}Don't add write tools to mcpAllowlist in .cursor/permissions.json. Leave them on approval. Cursor's docs say the allowlist is not a security boundary.
Credential layer
Needs no cloud credentials.
Facts
- Vendor
- OpenTofu official
- License
- NOASSERTION
- GitHub stars
- 111
- Latest release
- v1.0.0 (Jun 6, 2025)
- Last push
- Sep 21, 2026
- Status
- Active
- Last checked
- Oct 27, 2026
Sources
- README · checked Oct 27, 2026
Spotted something wrong? Labels change when vendors change. Suggest a correction