// SCAN REPORT

Cocos

157 tools. 85 can modify or destroy data without limits.

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

Last updated: 11 Jun 2026

85 can modify or destroy data

72 read-only

157 tools total

Community server · catalogue entry checked 11/06/2026

How to control Cocos ↓

TOOL MAP

What Cocos exposes to your agents

Read (72) Write / Execute (74) Destructive / Financial (9)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Cocos tools

Destructive · High batch_delete_assets This tool permanently removes assets without reversible recovery options. Destructive · High clear_all_reference_images This tool performs a bulk destructive operation that removes all reference images without individual confirmation, which cannot be undone through normal undo mechanisms if the operation completes. Destructive · Medium clear_broadcast_log Clearing a log is an irreversible destructive action — once the broadcast messages log is cleared, the historical data cannot be recovered. Destructive · Low clear_console Clearing the console irreversibly removes all existing log/output history from the editor console.

85 of Cocos's 157 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Cocos

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

Deny destructive operations

{
  "batch_delete_assets": {
    "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_component": {
    "limits": [
      {
        "counter": "add_component_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "check_builder_status": {
    "limits": [
      {
        "counter": "check_builder_status_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

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

Create a free account and register Cocos — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON COCOS →

Instant setup, no code required.

FULL CATALOGUE

All 157 Cocos tools

DESTRUCTIVE 9 tools

Destructive batch_delete_assets Delete multiple assets in batch Destructive clear_all_reference_images Clear all reference images Destructive clear_broadcast_log Clear the broadcast messages log Destructive clear_console Clear editor console Destructive delete_asset Delete an asset Destructive delete_node Delete a node from scene Destructive remove_component Remove a component from a node. componentType must be the component\ Destructive remove_reference_image Remove reference image(s) Destructive revert_prefab Revert prefab instance to original

EXECUTE 20 tools

Execute align_view_with_node Apply selected node position and angle to current view Execute build_project Build the project Execute change_gizmo_tool Change Gizmo tool Execute change_view_mode_2d_3d Change 2D/3D view mode Execute execute_component_method Execute method on component Execute execute_scene_script Execute scene script method Execute execute_script Execute JavaScript in scene context Execute focus_camera_on_nodes Focus scene camera on nodes Execute listen_broadcast Start listening for specific broadcast messages Execute open_asset_external Open asset with external program Execute open_preferences_settings Open preferences settings panel Execute refresh_assets Refresh asset database Execute refresh_reference_image Refresh reference image display Execute reimport_asset Reimport an asset Execute run_project Run the project in preview mode Execute scene_snapshot_abort Abort scene snapshot creation Execute soft_reload_scene Soft reload current scene Execute start_preview_server Start preview server Execute stop_listening Stop listening for specific broadcast messages Execute stop_preview_server Stop preview server

WRITE 54 tools

Write add_component Add a component to a specific node. IMPORTANT: You must provide the nodeUuid parameter to specify which node t Write add_reference_image Add reference image(s) to scene Write align_camera_with_view Apply scene camera position and angle to selected node Write attach_script Attach a script component to a node Write batch_import_assets Import multiple assets in batch Write begin_undo_recording Begin recording undo data Write cancel_undo_recording Cancel undo recording Write change_gizmo_coordinate Change coordinate system Write change_gizmo_pivot Change transform pivot point Write close_scene Close current scene Write compress_textures Batch compress texture assets Write copy_asset Copy an asset to another location Write copy_node Copy node for later paste operation Write create_asset Create a new asset file or folder Write create_node Create a new node in the scene. Supports creating empty nodes, nodes with components, or instantiating from as Write create_prefab Create a prefab from a node with all children and components Write create_scene Create a new scene asset Write cut_node Cut node (copy + mark for move) Write duplicate_node Duplicate a node Write duplicate_prefab Duplicate an existing prefab Write end_undo_recording End recording undo data Write import_asset Import an asset file Write import_preferences Import preferences configuration from file Write instantiate_prefab Instantiate a prefab in the scene Write move_array_element Move array element position Write move_asset Move an asset to another location Write move_node Move node to new parent Write open_scene Open a scene by path Write paste_node Paste previously copied nodes Write remove_array_element Remove array element at specific index Write reset_component Reset component to default values Write reset_node_property Reset node property to default value Write reset_node_transform Reset node position, rotation and scale Write reset_preferences Reset preferences to default values Write reset_scene_view Reset scene view to default settings Write restore_prefab Restore prefab instance from asset Write restore_prefab_node Restore prefab node using prefab asset (built-in undo record) Write save_asset Save asset content Write save_asset_meta Save asset meta information Write save_scene Save current scene Write save_scene_as Save scene as new file Write set_component_property Set component property values for UI components or custom script components. Supports setting properties of bu Write set_grid_visible Show/hide grid Write set_icon_gizmo_3d Set IconGizmo to 3D or 2D mode Write set_icon_gizmo_size Set IconGizmo size Write set_node_property Set node property value (prefer using set_node_transform for active/layer/mobility/position/rotation/scale) Write set_node_transform Set node transform properties (position, rotation, scale) with unified interface. Automatically handles 2D/3D Write set_preferences_config Set preferences configuration Write set_reference_image_data Set reference image transform and display properties Write set_reference_image_opacity Set reference image opacity Write set_reference_image_position Set reference image position Write set_reference_image_scale Set reference image scale Write switch_reference_image Switch to specific reference image Write update_prefab Update an existing prefab

READ 72 tools

Read check_builder_status Check if builder worker is ready Read check_server_connectivity Check server connectivity and network status Read detect_node_type Detect if a node is 2D or 3D based on its components and properties Read export_asset_manifest Export asset manifest/inventory Read export_preferences Export current preferences configuration Read find_asset_by_name Find assets by name (supports partial matching and multiple results) Read find_node_by_name Find first node by exact name Read find_nodes Find nodes by name pattern Read generate_available_url Generate an available URL based on input URL Read get_active_listeners Get list of active broadcast listeners Read get_all_nodes Get all nodes in the scene with their UUIDs Read get_all_preferences Get all available preferences categories Read get_asset_dependencies Get asset dependency tree Read get_asset_details Get detailed asset information including spriteFrame sub-assets Read get_asset_info Get asset information Read get_assets Get assets by type Read get_available_components Get list of available component types Read get_broadcast_log Get recent broadcast messages log Read get_build_settings Get build settings - shows current limitations Read get_component_info Get specific component information Read get_components Get all components of a node Read get_console_logs Get editor console logs Read get_current_scene Get current scene information Read get_editor_info Get editor and environment information Read get_log_file_info Get information about the project log file Read get_network_interfaces Get available network interfaces Read get_node_info Get node information by UUID Read get_node_tree Get detailed node tree for debugging Read get_performance_stats Get performance statistics Read get_prefab_info Get detailed prefab information Read get_prefab_list Get all prefabs in the project Read get_project_info Get project information Read get_project_logs Get project logs from temp/logs/project.log file Read get_project_settings Get project settings Read get_scene_hierarchy Get the complete hierarchy of current scene Read get_scene_list Get all scenes in the project Read get_scene_view_status Get comprehensive scene view status Read get_server_status Get comprehensive server status information Read get_unused_assets Find unused assets in project Read list_reference_images List all available reference images Read load_prefab Load a prefab by path Read open_build_panel Open the build panel in the editor Read query_asset_db_ready Check if asset database is ready Read query_asset_path Get asset disk path Read query_asset_url Get asset URL from UUID Read query_asset_uuid Get asset UUID from URL Read query_component_has_script Check if component has script Read query_current_reference_image Query current reference image data Read query_gizmo_coordinate Get current coordinate system Read query_gizmo_pivot Get current Gizmo pivot point Read query_gizmo_tool_name Get current Gizmo tool name Read query_gizmo_view_mode Query view mode (view/select) Read query_grid_visible Query grid visibility status Read query_icon_gizmo_3d Query IconGizmo mode Read query_icon_gizmo_size Query IconGizmo size Read query_nodes_by_asset_uuid Find nodes that use specific asset UUID Read query_preferences_config Query preferences configuration Read query_reference_image_config Query reference image configuration Read query_scene_classes Query all registered classes Read query_scene_components Query available scene components Read query_scene_dirty Check if scene has unsaved changes Read query_scene_ready Check if scene is ready Read query_server_ip_list Query server IP list Read query_server_port Query editor server current port Read query_sorted_server_ip_list Get sorted server IP list Read query_view_mode_2d_3d Get current view mode Read scene_snapshot Create scene state snapshot Read search_project_logs Search for specific patterns or errors in project logs Read validate_asset_references Validate asset references and find broken links Read validate_json_params Validate and fix JSON parameters before sending to other tools Read validate_prefab Validate a prefab file format Read validate_scene Validate current scene for issues

OTHER 2 tools

Other format_mcp_request Format a complete MCP request with proper JSON escaping Other safe_string_value Create a safe string value that won\

EXPLORE

Related servers

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

FAQ

Questions about Cocos

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

Yes. The Cocos server exposes 9 destructive tools including batch_delete_assets, clear_all_reference_images, clear_broadcast_log. 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 Cocos? +

The Cocos server has 54 write tools including add_component, add_reference_image, align_camera_with_view. 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 Cocos.

How many tools does the Cocos MCP server expose? +

157 tools across 4 categories: Destructive, Execute, Read, Write. 72 are read-only. 85 can modify, create, or delete data.

How do I enforce a policy on Cocos? +

Register the Cocos 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 Cocos tool call.

Deterministic rules across all 157 Cocos tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON COCOS →

Instant setup, no code required.

157 Cocos tools catalogued and risk-classified — across an index of 46,500+ MCP servers.

// WHERE THIS COMES FROM

These policies come from Cocos's registry record.

The record behind this page: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.

Teams ship this data inside their own products. See what a licence covers →