world-model-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| update_entityB | Create or update an entity in the spatial world model. Allows specifying position (3D coordinates), orientation (pitch/yaw/roll), bounding box volume, custom properties, tags, and confidence score. |
| query_entitiesA | Find entities by keyword query (FTS5 search), type, region, spatial proximity, tags, or status. Alternatively, provide entity_id for single-entity location and historical trajectory lookup. |
| set_relationB | Record or update a spatial relationship between two entities (e.g. on, inside, next_to, above, below, near, contains, occluded_by, connected_to, facing, holding, part_of). |
| get_spatial_mapB | Return a structured spatial layout, topological graph, 3D asset export (gltf/obj), or high-level environment summary of the known world. |
| simulate_movementB | Simulate physical movement and test for AABB collisions, or calculate waypoint navigation paths between entities and coordinates. |
| ingest_observationA | Merge structured vision perception detections into the world model (re-identifying existing objects and boosting confidence), or reconcile observed state against the expected frustum view. |
| get_expected_viewA | Calculate what entities should be visible from an observer's pose and field of view frustum cone. |
| link_to_goalB | Associate entities or spatial regions with state-memory task DAGs (link/unlink), or extract goal-relevant spatial context slices. |
| record_outcomeA | Update the world model after an action executes (moving an entity, modifying properties, destroying or creating objects). |
| manage_spatial_specC | Manage Spatial Spec-Driven Development (Spatial SDD) baseline contracts and live compliance verification against physical constraints (min clearance, containment, occupancy). |
| create_evidence_packA | Package entity positions, observation reconciliations, and snapshot states into an immutable, SHA-256 hashed cryptographic evidence pack for compliance and state-memory task verification. |
| use_spatial_blackboardA | Multi-agent shared spatial blackboard for publishing ephemeral intentions, waypoints, and claiming exclusive spatial access to prevent collisions. |
| manage_snapshotB | Unified spatial snapshot and time-travel management: save checkpoints, restore states, diff two snapshots, list history, undo mutations, or inspect world state at historical timestamps. |
| generate_game_inputsB | Generate Playwright MCP automation inputs (WASD, mouse-look, clicks) or project/unproject 3D entity coordinates and screen pixels. |
| wait_for_spatial_stateA | Poll and wait until an entity reaches a specific spatial state (exists, becomes active, confidence exceeds threshold, or enters region). |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| explore-surroundings | Generate spatial context and environment awareness prompt for agent's location |
| plan-navigation | Generate navigation and movement plan between entities |
| diagnose-spatial-anomalies | Analyze world model for physical overlaps, orphan relations, and permanence decay anomalies |
| navigate-game-world | Step-by-step perception-action loop recipe for autonomous game navigation with Playwright and Three.js/2D canvas |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| world-health | Server health status, version, and timestamp |
TDQS
Scored across 15 tools
Every tool targets a distinct aspect of spatial world modeling: entity CRUD, querying, relations, map generation, simulation, observation ingestion, view prediction, goal linking, outcome recording, spatial specs, evidence packs, blackboard coordination, snapshot management, game input generation, and state waiting. There is no ambiguity between operations like update_entity vs record_outcome vs ingest_observation—each has a clear and separate purpose.
All tools follow a consistent verb_noun snake_case pattern (e.g., update_entity, query_entities, set_relation, get_spatial_map, simulate_movement). Verb choices vary appropriately with their actions, but the style and structure are uniform across all 15 tools, making the set predictable and easy to navigate.
15 tools is well-scoped for a rich domain like spatial world modeling. Each tool covers a distinct feature area without redundancy, and the count remains within the ideal range for a comprehensive MCP server. The number feels justified given the breadth of functionality (state management, simulation, spatial reasoning, evidence, collaboration, and automation).
The tool surface covers the full lifecycle: entity creation/update (update_entity), querying (query_entities), deletion and post-action updates (record_outcome), relationships (set_relation), spatial mapping (get_spatial_map), simulation (simulate_movement), observation ingestion (ingest_observation), expected view (get_expected_view), goal integration (link_to_goal), compliance (manage_spatial_spec), evidence (create_evidence_pack), multi-agent coordination (use_spatial_blackboard), snapshot/history (manage_snapshot), game input generation (generate_game_inputs), and waiting for state changes (wait_for_spatial_state). No obvious gaps prevent an agent from performing critical workflows.