// SCAN REPORT

Portals

51 tools. 23 can modify or destroy data without limits.

23 write tools that can modify data. Rate limits recommended.

Last updated: 5 Jul 2026

23 can modify or destroy data

28 read-only

51 tools total

Verified server · catalogue entry verified 05/07/2026 · full schemas captured for 44 of 51 tools

How to control Portals ↓

TOOL MAP

What Portals exposes to your agents

Read (28) Write / Execute (23) Destructive / Financial (0)

Running more than one server? Check your whole stack →

CONTEXT-WINDOW COST

What Portals costs in tokens

13,351 tokens of tool definitions, loaded on every request

6.7% of a 200k context window

2,963 heaviest tool: apply_operations

Full per-tool breakdown → Model it in the token calculator →

MOST DANGEROUS TOOLS

High Risk

The most dangerous Portals tools

Execute · High simulate_key_input This tool executes keyboard input simulation that directly triggers event handlers and game logic flows. Execute · Medium simulate_trigger_zone_input This tool executes code/operations by simulating input events that cause external state changes in the Portals application (triggering task execution, zone activation). Execute · High change_task_state This tool executes operations in a live game environment (via IFrameBridge) that cause state transitions and trigger reactive logic. Execute · High connect_to_game This tool initiates an external connection to a live game client via WebSocket, triggers events, and modifies global state (sets the room ID for subsequent tools).

23 of Portals's 51 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Portals

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

Rate limit write operations

{
  "add_to_room_inventory": {
    "limits": [
      {
        "counter": "add_to_room_inventory_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "analyze_scene": {
    "limits": [
      {
        "counter": "analyze_scene_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 Portals — 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 PORTALS →

Instant setup, no code required.

FULL CATALOGUE

All 51 Portals tools

EXECUTE 9 tools

Execute change_task_state Change the state of a task/quest or activate an effector in the live game via the IFrameBridge. Two modes: (1) Execute connect_to_game Connect to the live game client in the browser via WebSocket bridge. Sends an mcp_connect event and waits for Execute generate_ai_image Generate an AI image from a text prompt using Google Imagen 4 Fast (via Replicate). Kicks off an async job — p Execute generate_ai_texture Generate a seamless PBR texture from a text prompt using Google Imagen 4 Fast (via Replicate). Kicks off an as Execute image_to_3d_model Generate a 3D model from an image using Tripo AI. Provide a URL to an image (jpg, png, webp, etc.). Returns a Execute record_video Record a video of the live 3D scene from a CameraObject. Pass roomItemId to record from an existing CameraObje Execute simulate_key_input Audit Portals OnKeyPressedEvent/OnKeyReleasedEvent trigger mappings and optionally simulate key press/release Execute simulate_trigger_zone_input Audit Portals Trigger-zone pressBtn/keyCode behavior and optionally simulate enter/exit/key frames. Use this f Execute text_to_3d_model Generate a 3D model from a text prompt using Tripo AI. Returns a task ID. Use check_3d_model_task to poll for

WRITE 14 tools

Write add_to_room_inventory Add marketplace items (and packs — auto-expanded to their contained items) to a room's build inventory so they Write apply_operations Apply targeted operations (add, modify, remove items/logic/quests) to a room and upload the result. Operations Write claim_marketplace_item Claim a free marketplace item for the signed-in user, copying it into their inventory. Required before adding Write create_room Create a new room from a template. Returns the new room ID. Write duplicate_room Duplicate an existing room. Creates a copy with all items, settings, tasks, and quests. Returns the new room I Write generate_music Generate an instrumental or vocal music track from a text prompt using ElevenLabs. Returns a URL to the genera Write generate_sound_effect Generate a sound effect from a text description using ElevenLabs. Returns a URL to the generated audio file (M Write place_marketplace_items Claim a batch of marketplace items (skipping any already-owned) and compose the apply_operations payload that Write set_room_data Replace the entire room data. Provide an absolute path to a JSON file. JSON structure: { roomItems, settings, Write set_room_settings Update ONLY a room's scene settings (lighting, skybox, fog, postprocess, movement, avatars, voice chat, etc.) Write update_room_settings Update room settings like name, description, image, visibility, etc. Write upload_glb Upload a .glb file (or all .glb files from a folder) to the Portals CDN. Provide filePath for a single file, o Write upload_image Upload an image (.jpg, .jpeg, .png, .gif) to the Portals CDN. Provide filePath for a single file, or folderPat Write wire_locked_door Compose the apply_operations payload to wire a locked-door mechanic between an existing key item and an existi

READ 28 tools

Read analyze_scene Analyze the spatial composition of a room — zones, relationships, design checks, and LLM reasoning templates. Read authenticate Authenticate with your access key. Only call this if the server is not already authenticated — the server auto Read check_3d_model_task Check the progress of a Tripo AI 3D model generation task. Polls the backend until the task completes or times Read compare_scene Compare two room snapshots to see what changed — added/removed/moved/rescaled objects, zone changes, and stats Read get_context Returns inline specs, syntax references, gotchas, and recipe matches for your task. Recommended before buildin Read get_help Return focused setup/workflow guidance and matching docs:// resources. Read get_pack_items List the items contained in a marketplace pack (when search_marketplace returned an item with isPack=true). Re Read get_room_build_items List EVERY item currently available in this room's in-game build palette — covers items the user added via P1P Read get_room_data Download all room data to a temporary JSON file and return the file path. Use the file path to read the data. Read get_runtime_data Fetch live runtime variables and JS function effector results from the Unity game client. Returns numeric para Read get_user_inventory List assets the signed-in user has already claimed or purchased. Useful before claiming new items (avoid dupli Read inspect_room_data Return a compact, model-friendly summary of Portals room data without dumping the full JSON. Prefer file_path Read list_generated_3d_models List all AI-generated 3D models (GLBs) for the current user. Returns models with their status, GLB URLs, thumb Read list_generated_images List all AI-generated images for the current user. Returns each image with its prompt, URL (empty while still Read list_generated_sounds List AI-generated sounds (speech, sound effects, music) for the current user. Returns sound metadata including Read list_generated_textures List all AI-generated textures for the current user. Returns each texture with its URL (empty while still gene Read list_marketplace_facets List the distinct values currently available across the marketplace catalogue (categories, subcategories, them Read list_voices List available ElevenLabs voices for text_to_speech. Results are paginated (100 per page). Use filters to narr Read lookup Fast knowledge lookup. Returns specs for items, triggers, effects, systems, or recipes directly. Supports batc Read plan_gameplay_mechanic Create a compact, capability-grounded implementation contract for a Portals gameplay mechanic. This does not e Read poll_game_events Poll for events sent from the live game client via the WebSocket bridge. Returns and clears all queued events. Read position_camera Compute optimal camera placement(s) to frame one or more room items. Uses smart framing, collision avoidance, Read query_room Query room items by type, position, triggers, text search, parent, or quest reference. Automatically pulls fre Read render_scene Take a screenshot of the live 3D scene via the game's MCP camera for visual analysis. Pass a roomItemId of an Read resolve_gameplay_capability Classify a Portals gameplay request against the shared harness grounding layer. Returns supported_directly, su Read search_marketplace Search the Portals creator marketplace for assets (3D models, sounds, packs). Returns up to limit items with i Read search_recipes Search bundled recipe manifest entries by keywords, tags, and patterns. Read text_to_speech Generate speech audio from text using ElevenLabs. Returns a URL to the generated audio file (MP3). Optionally

EXPLORE

Related servers

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

FAQ

Questions about Portals

How do I prevent bulk modifications through Portals? +

The Portals server has 14 write tools including add_to_room_inventory, apply_operations, claim_marketplace_item. 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 Portals.

How many tools does the Portals MCP server expose? +

51 tools across 3 categories: Execute, Read, Write. 28 are read-only. 23 can modify, create, or delete data.

How do I enforce a policy on Portals? +

Register the Portals 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 Portals tool call.

Deterministic rules across all 51 Portals tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON PORTALS →

Instant setup, no code required.

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