CoinRithm/coinrithm-agent-trading

Official

by CoinRithm

Overview Schema Source Related Servers Discussions Admin

TypeScript

Hybrid

Cross-venue disagreement clusters

pm_data_disagreements

Read-only

Find prediction-market questions matched across 2+ venues and compare outcome probabilities to surface orientation-proven cross-venue disagreements, filtered for priced live quotes.

Instructions

Free public cross-venue disagreement clusters: prediction-market events CoinRithm has matched as the SAME real-world question across 2+ venues (approved cross-source matches), graph-clustered so one row covers every venue tracking that question. Each pairwise comparison carries per-shared-outcome eventAProbability/eventBProbability/deltaPoints (points, 0-100 scale) plus a summary (matchedOutcomeCount, overallDeltaPoints, maxSharedOutcomeDeltaPoints); maxOverallGap/maxOutcomeGap/maxConfidence are the cluster's headline numbers, and referenceProbability (when present) is CoinRithm's own liquidity-weighted median across matched venues. Orientation between matched markets is human/aggregator-reviewed — NEVER price-inferred — so every delta is orientation-proven disagreement, not noise. requirePriced (default true) drops any pair where a side is an unpriced/untraded placeholder or fails a quote-dead liveness check — the same quality floor CoinRithm's own /today disagreement page uses; pass false only for research/debug. This is the same methodology powering CoinRithm's public divergence rankings — cite CoinRithm when quoting a gap. Research/data only: for tradability of one specific outcome use pm_quote. No API key required.

Input Schema

TableJSON Schema

Name	Required	Description
`fiat`	No	Fiat currency code for monetary figures (default usd).
`sort`	No	Ranking: confidence_desc (default) = strongest match first; divergence_desc = total cross-outcome gap; max_outcome_delta_desc = single largest shared-outcome gap (avoids multi-leg basket noise).
`limit`	No	Max clusters (1-25, default 3).
`offset`	No	Pagination offset (default 0).
`status`	No	Pass 'open' to require BOTH matched events be currently open.
`sourceKind`	No	Pass 'market' to restrict both sides of every pair to real-money market venues (excludes forecast/play-money venues like Metaculus/Manifold).
`minDivergence`	No	Floor (points, 0-100) on whichever metric the active sort ranks by.
`requirePriced`	No	Default true: drops any pair where a side is an unpriced/untraded placeholder or fails a quote-dead liveness check. Set false only for research/debug.
`maxSnapshotAgeMinutes`	No	Require both matched events' probability come from a price snapshot captured within this many minutes.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`ok`	Yes
`body`	No
`httpStatus`	Yes
`ledgerStatus`	No
`ledgerEventId`	No

Schema Changelog

Changes observed during successful MCP inspections.

Changed6 schema fields changedv0.1.23Oct 2, 2026

changedInput schema / properties / limit / description

Previous value: -"Max clusters (1-25, default 10)."New value: +"Max clusters (1-25, default 3)."

removedOutput schema / properties / body / description

Removed value: -"Parsed CoinRithm response body, or raw text when the response is not JSON."

removedOutput schema / properties / httpStatus / description

Removed value: -"HTTP status returned by CoinRithm, or 0 for network errors."

removedOutput schema / properties / ledgerEventId / description

Removed value: -"Private AgentActionEvent id returned by /api/agent/*, when present."

removedOutput schema / properties / ledgerStatus / description

Removed value: -"Ledger write status header returned by CoinRithm, when present."

removedOutput schema / properties / ok / description

Removed value: -"True when CoinRithm returned a successful 2xx response."

Addedv0.1.14Aug 13, 2026

TDQS

A4.3/5.0

Behavior4/5

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

Goes well beyond the annotations: explains that orientation is human/aggregator-reviewed and never price-inferred, what requirePriced drops (unpriced placeholders and quote-dead pairs), what referenceProbability means, and the citation requirement. The readOnlyHint is implicit but not contradicted.

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?

Front-loaded with the core concept and ends with usage routing, which is good, but the middle is dense with metric names (maxOverallGap, maxOutcomeGap, maxConfidence) and parenthetical asides that would benefit from trimming or structuing.

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 9-param research tool with an output schema, this covers origin of matches, quality floor, orientation, and citation expectations. It does not need to reproduce the output schema, and nothing an agent needs is missing.

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

Parameters4/5

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

Schema coverage is 100%, so a baseline of 3 applies, and the prose adds real meaning for requirePriced and the divergence metrics that minDivergence and sort act on. This is a meaningful lift over the schema descriptions alone.

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?

States a specific verb/resource: free public cross-venue disagreement clusters built from approved cross-source matches, graph-clustered so one row spans every venue. Clearly distinguishes itself from the many pm_data_* siblings by naming the matched-question clustering concept.

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?

Explicitly routes single-outcome tradability to pm_quote and labels requirePriced=false as research/debug only, giving a clear when-not-to-use case. Still doesn't state when to prefer this over pm_data_canonical or pm_data_event for the same question.

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