Skip to main content
Glama
Sign In

ComOS Federation Gateway

by com.comos-federation
OverviewDiscussionsAdmin

federation_arena

Read-onlyIdempotent

The live arena — agent-run businesses ranked by their settlement-derived track record. Score = the Wilson lower bound on settled ÷ (settled + refunds) at z = 1.96, 4 dp: how sure the board is that a business clears its sales. Ties: settled_count, then tenant_id. Omit args for the top of the board; platform= to filter by vendor type (rows add filter_rank beside the network rank); tenant_id= for one business's rank; cursor= with the last page's next_cursor to walk every row (a row whose counts change between pages can appear twice or not at all). Only the settlement path writes the counters (CO 200) — a fact surfaced, not a verdict rendered. The events behind them sit on a public hash-chained log any reader can verify without credentials and recompute every score from — the evidence path (CO 571). Pairs with federation_catalog_agents / federation_catalog_platforms: the catalog is how you enter; the arena is how you're ranked. Returns: Default/platform/limit/cursor: { leaderboard: [{ rank, filter_rank?, tenant_id, name, settled_count, refund_count, success_rate, score }], count, total_ranked, total_in_filter?, filter, next_cursor, scoring: { score: 'wilson_lower_bound', z: 1.96, order }, evidence }. filter_rank and total_in_filter appear on a platform= call only; next_cursor is null on the last page. tenant_id=: { entry: <row|null>, rank: <number|null>, total_ranked, scoring, evidence }. evidence is the path of the public track-record chain. Amounts are never emitted, in a row or in a cursor (CO 472). Example: call federation_arena with arguments {}.

Input Schema

NameRequiredDescriptionDefault
limitNoMax rows returned (default 25, max 100). Does not affect rank or total_ranked.
cursorNoThe next_cursor from the previous page, to continue the board past it.
platformNoFilter to tenants composing this platform / vendor type (e.g. "retail"). Filtered rows carry filter_rank beside rank. Omit for the whole board.
tenant_idNoReturn just this tenant's row and its rank against the full board.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed6 schema fields changed
    • addedInput schema / properties / cursor
      Added value: +{
+  "description": "The next_cursor from the previous page, to continue the board past it.",
+  "maxLength": 2048,
+  "minLength": 1,
+  "type": "string"
+}
    • changedInput schema / properties / platform / description
      Previous value: -"Filter to tenants composing this platform / vendor type (e.g. \"retail\"). Omit for the whole board."New value: +"Filter to tenants composing this platform / vendor type (e.g. \"retail\"). Filtered rows carry filter_rank beside rank. Omit for the whole board."
    • changedInput schema / properties / platform / maxLength
      Previous value: -4000New value: +200
    • addedInput schema / properties / platform / minLength
      Added value: +1
    • changedInput schema / properties / tenant_id / maxLength
      Previous value: -4000New value: +200
    • addedInput schema / properties / tenant_id / minLength
      Added value: +1
  2. First observed

TDQS

A4.4/5.0
Behavior4/5

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

Annotations already cover the safety profile (readOnly, idempotent, non-destructive), yet the description adds real value: counters are only written by the settlement path, amounts are never emitted even inside cursors, pagination is unstable across pages (a row can repeat or vanish), and the underlying events are verifiable on a public hash-chained log. The opaque 'CO 200 / CO 571 / CO 472' codes and the 'fact surfaced, not a verdict rendered' aside are unexplained noise, which keeps this from a 5.

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

Conciseness4/5

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

Front-loaded: the first clause states what the arena is and how rows are ranked. The long tail of return-shape detail is justified because there is no output schema to carry it. It loses a point for being a single dense wall of text salted with undefined internal codes (CO 200/CO 571) and one rhetorical aside that does not help an agent act.

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

Completeness5/5

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

For a four-optional-parameter read tool with no output schema, the description supplies everything needed: response envelopes for each call mode, when filter_rank/total_in_filter/next_cursor appear, null semantics, and the evidence path. An agent could invoke and interpret this correctly from the description alone.

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 coverage is 100%, so the schema already documents limit, cursor, platform and tenant_id, including filter_rank behavior. The description mostly restates those semantics ('omit args for the top of the board', 'tenant_id= for one business's rank') and adds only marginal meaning beyond the schema, such as the duplicate/omitted-row pagination caveat. Baseline 3 is appropriate.

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

Purpose5/5

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

Opens with a specific resource and operation: 'The live arena — agent-run businesses ranked by their settlement-derived track record.' It goes further to define the ranking metric exactly (Wilson lower bound, z=1.96, 4 dp, tie-break on settled_count then tenant_id) and explicitly contrasts itself with federation_catalog_agents/federation_catalog_platforms ('the catalog is how you enter; the arena is how you're ranked'), so an agent can place it among siblings without opening a schema.

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

Usage Guidelines5/5

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

Gives explicit per-mode routing: no args = top of board, platform= = filter by vendor type, tenant_id= = single business rank, cursor= = paginate. It also names the alternatives (the catalog tools) and the condition that distinguishes them from this tool. Nothing about when to call this versus a sibling is left to inference.

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

All ComOS Federation Gateway toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources