Hcloud

76 tools. 53 can modify or destroy data without limits.

11 destructive tools with no built-in limits. Policy required.

Last updated:

53 can modify or destroy data
23 read-only
76 tools total

Community server · catalogue entry verified 03/07/2026

How to control Hcloud ↓

What Hcloud exposes to your agents

Read (23) Write / Execute (42) Destructive / Financial (11)
Critical Risk

The most dangerous Hcloud tools

53 of Hcloud's 76 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

How to control Hcloud

PolicyLayer is an MCP gateway — it sits between your AI agents and Hcloud, and nothing reaches the server without passing your rules. These are the rules we recommend:

Deny destructive operations
{
  "delete_firewall": {
    "deny_if": [
      {
        "conditions": [],
        "on_deny": "Blocked by default. Requires approval."
      }
    ]
  }
}

Destructive tools should never be available to autonomous agents without human approval.

Rate limit write operations
{
  "add_load_balancer_service": {
    "limits": [
      {
        "counter": "add_load_balancer_service_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations
{
  "get_action": {
    "limits": [
      {
        "counter": "get_action_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

  1. Create a free account and register Hcloud — nothing to install.
  2. Add these rules — paste them, or build them visually. Tune the limits to your setup.
  3. Point your MCP client (Claude, Cursor, anything) at your gateway URL.
ENFORCE POLICY ON HCLOUD →

Instant setup, no code required.

All 76 Hcloud tools

WRITE 37 tools
Write add_load_balancer_service Add a service to a load balancer Write add_load_balancer_target Add a target to a load balancer Write add_network_subnet Add a subnet to a network (e.g. for private networking 10.0.0.0/16 in a zone) Write apply_firewall_to_resources Apply a firewall to servers or a label selector. Pass apply_to: [{ type: Write assign_floating_ip Assign a floating IP to a server Write assign_primary_ip Assign a Primary IP to a server (server must be powered off) Write attach_iso Attach an ISO to a server Write attach_server_to_network Attach a server to a network Write attach_volume Attach a volume to a server Write change_server_type Change the type of a server (scales resources). Server must be powered off. Write create_firewall Create a new firewall. Rules: direction Write create_floating_ip Create a floating IP Write create_load_balancer Create a new load balancer Write create_network Create a new network Write create_placement_group Create a new placement group Write create_primary_ip Create a Primary IP (datacenter-scoped; server must be off to assign). Provide datacenter OR assignee_id (serv Write create_server Create a new server Write create_ssh_key Create a new SSH key Write create_volume Create a new volume Write detach_iso Detach an ISO from a server Write detach_server_from_network Detach a server from a network Write detach_volume Detach a volume Write remove_firewall_from_resources Remove a firewall from servers. Pass remove_from: [{ type: Write remove_load_balancer_target Remove a target from a load balancer Write set_firewall_rules Set rules on an existing firewall (overwrites all rules). Same rule format as create_firewall. Write unassign_floating_ip Unassign a floating IP Write unassign_primary_ip Unassign a Primary IP from its server (server must be powered off) Write update_firewall Update a firewall (name, labels) Write update_floating_ip Update a floating IP (description, labels) Write update_load_balancer Update a load balancer (name, labels) Write update_load_balancer_service Update a load balancer service (listen_port identifies the service) Write update_network Update a network (name, labels) Write update_placement_group Update a placement group (name, labels) Write update_primary_ip Update a Primary IP (name, labels) Write update_server Update a server (name, labels) Write update_ssh_key Update an SSH key (name, labels) Write update_volume Update a volume (name, labels)

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

Questions about Hcloud

Can an AI agent delete data through the Hcloud MCP server? +

Yes. The Hcloud server exposes 11 destructive tools including delete_firewall, delete_floating_ip, delete_load_balancer. These permanently remove resources with no undo. PolicyLayer blocks destructive tools by default so they never reach the upstream server.

How do I prevent bulk modifications through Hcloud? +

The Hcloud server has 37 write tools including add_load_balancer_service, add_load_balancer_target, add_network_subnet. Set a rate limit in your policy -- for example, 10 calls per hour prevents an agent from making more than 10 modifications per hour. PolicyLayer enforces this at the gateway, before calls reach Hcloud.

How many tools does the Hcloud MCP server expose? +

76 tools across 3 categories: Destructive, Read, Write. 23 are read-only. 53 can modify, create, or delete data.

How do I enforce a policy on Hcloud? +

Register the Hcloud MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.

Enforce policy on every Hcloud tool call.

Deterministic rules across all 76 Hcloud tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

Instant setup, no code required.

76 Hcloud tools catalogued and risk-classified — across an index of 43,000+ MCP servers.

// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.