MCP server for the delta-v game that provides tools for agents to interact with game sessions, queue matches, observe game state, send actions, and access game resources
10 tools with verb-prefixed names (delta_v_*) and basic descriptions present. However, critical gaps undermine quality: (1) Parameter descriptions are minimal or absent for most tools, e.g., delta_v_send_action's 'action' param is described only as 'Action object to submit' with no schema detail on what fields it accepts; (2) Output schemas are not documented anywhere in the visible code, LLMs cannot infer what fields to expect from responses; (3) Error handling is absent, no guidance on retryability, user-fixable vs. fatal errors, or recovery steps; (4) Tool descriptions lack context on when to use each tool vs. alternatives (e.g., delta_v_observe vs. delta_v_wait_for_turn distinction is unclear); (5) No enum constraints on scenario/action parameters, inviting hallucinated values. Naming follows verb_noun convention well, but parameter and output documentation is severely underdeveloped.
Create a new delta-v game session by queueing for a match with optional scenario and server URL
Get the leaderboard of agents resource
Get the match log resource for a completed game
Get the match replay resource for a completed game
Get the current game rules resource
Retrieve historical events from a game session
Get the current game observation for a session including game state, legal actions, and tactical information
Reconnect to an existing game session after a disconnection
Output schemas completely undocumented. No visible schema definitions for any tool's return type. LLMs cannot plan downstream calls or extract required fields (e.g., sessionId from create_session response).
Parameter 'action' in delta_v_send_action lacks schema. Described only as 'Action object to submit' with no type, required fields, or valid action types. LLM cannot construct valid payloads.
No error handling guidance. Tools provide no recovery hints, retryability classification, or actionable error messages. If a tool fails, LLM has no direction on next steps.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | 2025-06-18+ | v2 |
Submit an action to the game server and optionally wait for the result
Block until it is the agent's turn to act, with optional timeout
Scenario parameter lacks enum constraint. 'scenario' defaults to 'duel' but no list of valid scenarios provided. LLM may hallucinate invalid scenario names.
Tool descriptions lack context on when to use each vs. alternatives. delta_v_observe and delta_v_wait_for_turn both retrieve game state but distinction is unclear. No guidance on which to call when.