// SCAN REPORT

Tailscale

89 tools. 47 can modify or destroy data without limits.

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

Last updated: 6 May 2026

47 can modify or destroy data

42 read-only

89 tools total

// TOOL MAP

Read (42) Write / Execute (37) Destructive / Financial (10)

// WHAT CAN GO WRONG

Destructive tools (tailscale_delete_device, tailscale_delete_device_invite, tailscale_delete_device_posture_attribute) permanently delete resources. There is no undo. An agent calling these in a retry loop causes irreversible damage.

Write operations (tailscale_accept_device_invite, tailscale_approve_user, tailscale_batch_update_posture_attributes) modify state. Without rate limits, an agent can make hundreds of changes in seconds — faster than any human can review or revert.

// RECOMMENDED RULES

Deny destructive operations

tailscale_delete_device:
  rules:
    - action: deny

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

Rate limit write operations

tailscale_accept_device_invite:
  rules:
    - rate_limit: 30/hour

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

tailscale_authorize_device:
  rules:
    - rate_limit: 60/minute

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

// ALL 89 TAILSCALE TOOLS

DESTRUCTIVE 10 tools

Destructive tailscale_delete_device Destructive tailscale_delete_device_invite Destructive tailscale_delete_device_posture_attribute Destructive tailscale_delete_key Destructive tailscale_delete_log_stream_config Destructive tailscale_delete_posture_integration Destructive tailscale_delete_service Destructive tailscale_delete_user Destructive tailscale_delete_user_invite Destructive tailscale_delete_webhook

WRITE 37 tools

Write tailscale_accept_device_invite Write tailscale_approve_user Write tailscale_batch_update_posture_attributes Write tailscale_create_aws_external_id Write tailscale_create_device_invite Write tailscale_create_key Write tailscale_create_posture_integration Write tailscale_create_user_invite Write tailscale_create_webhook Write tailscale_rename_device Write tailscale_resend_device_invite Write tailscale_resend_user_invite Write tailscale_restore_user Write tailscale_rotate_webhook_secret Write tailscale_set_contacts Write tailscale_set_device_ip Write tailscale_set_device_posture_attribute Write tailscale_set_device_routes Write tailscale_set_device_tags Write tailscale_set_devices_authorized Write tailscale_set_dns_configuration Write tailscale_set_dns_preferences Write tailscale_set_log_stream_config Write tailscale_set_nameservers Write tailscale_set_search_paths Write tailscale_set_service_device_approval Write tailscale_set_split_dns Write tailscale_suspend_user Write tailscale_update_acl Write tailscale_update_device_key Write tailscale_update_key Write tailscale_update_posture_integration Write tailscale_update_service Write tailscale_update_split_dns Write tailscale_update_tailnet_settings Write tailscale_update_user_role Write tailscale_update_webhook

READ 42 tools

// FAQ

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

Yes. The Tailscale server exposes 10 destructive tools including tailscale_delete_device, tailscale_delete_device_invite, tailscale_delete_device_posture_attribute. These permanently remove resources with no undo. Intercept blocks destructive tools by default so they never reach the upstream server.

How do I prevent bulk modifications through Tailscale? +

The Tailscale server has 37 write tools including tailscale_accept_device_invite, tailscale_approve_user, tailscale_batch_update_posture_attributes. Set rate limits in your policy file -- for example, rate_limit: 10/hour prevents an agent from making more than 10 modifications per hour. Intercept enforces this at the transport layer.

How many tools does the Tailscale MCP server expose? +

89 tools across 3 categories: Destructive, Read, Write. 42 are read-only. 47 can modify, create, or delete data.

How do I add Intercept to my Tailscale setup? +

One line change. Instead of running the Tailscale server directly, prefix it with Intercept: intercept -c tailscale.yaml -- npx -y @@yawlabs/tailscale-mcp. Download a pre-built policy from policylayer.com/policies/tailscale and adjust the limits to match your use case.

Tailscale

Other MCP servers with similar tools.

Let agents act without letting them run wild.