Skip to main content
Glama

Start Match

horizon_start_match
DestructiveIdempotent

Executes start match under the exact original revision. Use when: explicit asynchronous multiplayer actions after reading horizon://docs/matches.

Instructions

Executes start match under the exact original revision. Use when: explicit asynchronous multiplayer actions after reading horizon://docs/matches. Not for: automatic polling, page traversal, timeout progression, invented winners or automatic rewards. Reconcile with horizon_get_matches and horizon_get_match. Consent and game consensus do not validate honest gameplay. Requires: userId and the player's session: sign in with horizon_signin_email or horizon_signin_anonymous and pass its accessToken as sessionToken. Side effects: One HTTP attempt. Start requires the creator and every joined UUID exactly once. Turn requires the current unexpired actor. Timeout advances only one turn. Active leave forfeits without selecting a winner. Lobby cancel is creator-only. Full Inbox retains a durable notice without undoing a committed move. Returns: {match,pollAfterSeconds,serverTime}; departing lobby actors receive metadata only and lose point-read access. Errors: 404 hides private/missing/expired matches. 409 REVISION_CONFLICT or STATE_CONFLICT: explicitly reread and reconcile the original intent, never replace the revision or automatically resubmit. RESULT_CONFLICT preserves the original report. RUN_EXPIRED/RUN_ALREADY_USED never allow replacement ranked runs. 422 ACCOUNT_MATCH_LIMIT_REACHED refuses create only, when the account's open matches (lobby, running, reporting) across all its projects are at the plan limit; join, turns, leave, report and cancel are not blocked, so wait for matches to finish instead of retrying. 501 FEATURE_NOT_SUPPORTED is explicit for simpleServer ranked mode. 401 means the session is missing or expired: sign in again and retry with the new accessToken. 403 for a session means the token belongs to another user: pass the userId that signed in. Needs HORIZON_API_KEY. Every failure returns an error result (isError) with the HTTP status and body: a 401 that horizon_test_connection also returns means HORIZON_API_KEY is wrong, 429 means the account's per-minute rate limit was reached (wait a minute, then retry).

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "userId": "<userId>",
      "sessionToken": "<sessionToken>",
      "matchId": "<matchId>",
      "expectedRevision": "<expectedRevision>",
      "turnOrder": "<turnOrder>"
    },
    "name": "horizon_start_match"
  }
}

Input Schema

NameRequiredDescriptionDefault
userIdYesActive player matching the session and project
matchIdYesOriginal server match UUID in the selected project
turnOrderYesEvery joined participant in the explicit turn order
sessionTokenYesPlayer accessToken sent only as Bearer
expectedRevisionYesOriginal reviewed revision; never substitute a newer revision after an uncertain write

Schema Changelog

Changes observed during successful MCP inspections.

  1. Addedv1.9.1

TDQS

A3.8/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare destructiveHint=true, idempotentHint=true and openWorldHint=true, so the safety profile is covered; the description nevertheless adds substantial context: one HTTP attempt per call, creator-plus-full-roster requirement, session token derivation, a 404/409/422/501 error taxonomy, and 429 rate-limit behavior. The deduction is for scope drift, since several disclosed 'side effects' (turn, timeout, leave, cancel) belong to other tools and are noise here rather than transparency about this call.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

It is well front-loaded and uses labelled sections (Use when / Not for / Requires / Side effects / Returns / Errors), which aids scanning. But it is very long for a five-parameter tool and spends several sentences on behaviour that this call does not perform, burying the actionable requirements (creator + full roster, 409 handling, 422 semantics) in dense run-on prose.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description compensates by naming the return shape {match, pollAfterSeconds, serverTime} and by spelling out auth, error and conflict handling for a destructive, revision-guarded mutation. Almost everything needed to call it correctly is present; only the ambiguity about which operations this endpoint truly performs keeps it short of a 5.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already defines all five parameters, making 3 the baseline. The description does add real meaning for two of them: sessionToken must be the accessToken from horizon_signin_email or horizon_signin_anonymous, and expectedRevision must never be replaced with a newer revision after an uncertain write. turnOrder's 'every joined UUID exactly once' is implied more than explained, so the lift is moderate rather than decisive.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening sentence names a specific verb and resource ('Executes start match') and pins it to the 'exact original revision', which is enough to separate it from horizon_get_match or horizon_create_match. However, the body then enumerates semantics for turn, timeout, leave and cancel, which read like sibling operations and blur what this tool actually does. An agent gets the gist but may briefly wonder whether this is a multi-action endpoint.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

'Use when: explicit asynchronous multiplayer actions after reading horizon://docs/matches' sets a context, and the 'Not for' list concretely excludes automatic polling, page traversal, timeout progression and invented winners/rewards. It also routes reconciliation to horizon_get_matches and horizon_get_match. What it lacks is an explicit pointer to the sibling that handles turns, timeouts, leaving or cancelling, which the body discusses anyway.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools