Crosswire Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Crosswire Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Pre-trade resolution-risk firewall for cross-venue prediction-market arbitrage. Before executing a two-leg Polymarket + Kalshi position, Crosswire flags resolution mismatch, void risk, and settlement divergence — void-rule asymmetry, settlement-scope differences (90-minute vs extra-time), settlement-source and -timing gaps, stale data, and thin liquidity between Polymarket and Kalshi FIFA World Cup 2026 markets — so automated trading agents don't get blown up by markets that look identical but settle differently. Verdict: safe / caution / block.
It does not predict outcomes and it does not execute trades. It answers one question: if I buy this contract on one venue and sell its twin on the other, do they actually settle on the same thing? When they don't, it names the exact clause that breaks the hedge.
https://api.crosswire-api.com/mcpGeneric MCP client config (mcp.json):
{
"mcpServers": {
"crosswire": {
"url": "https://api.crosswire-api.com/mcp"
}
}
}Claude Code:
claude mcp add --transport http crosswire https://api.crosswire-api.com/mcpcheck_resolution_riskPre-trade safety check for a cross-venue pair. Returns a machine-readable verdict — execution_verdict: safe / caution / block — inside a full Fungibility & Settlement Audit Object (FSAO) with structured findings, top-level divergence flags, venue rule overrides, fee-adjusted spread, and verdict reasons.
Identify the pair by either:
canonical_event_id — a pair id from list_covered_events,e.g. "wc26:match:MEX-RSA:2026-06-11:result#home" (#home / #draw / #away selects the outcome leg), or
market_a + market_b — one Polymarket conditionId (0x… hex) and one Kalshi ticker(e.g. KXWCGAME-26JUN11MEXRSA-MEX), order-insensitive. outcome on a ref is optional — matching uses only (venue, market_id).
Optional: notional_usd (enables the thin-liquidity check at your trade size).
If the pair is not covered, the tool returns a non-error coverage reply (covered: false) naming the covered matches and the coverage cutoff instead of an FSAO.
Example — real production audit (World Cup opener, Mexico vs South Africa, home leg):
Request:
{
"canonical_event_id": "wc26:match:MEX-RSA:2026-06-11:result#home"
}Response (abridged with …):
{
"audit_id": "aud_01KTW2Y7S3Q3JFYJK863FJB66X",
"canonical_event_id": "wc26:match:MEX-RSA:2026-06-11:result#home",
"leg_a": { "venue": "polymarket", "market_id": "0x4cd77d45…10ff", "outcome": "Yes" },
"leg_b": { "venue": "kalshi", "market_id": "KXWCGAME-26JUN11MEXRSA-MEX", "outcome": "yes" },
"matching_confidence": 0.98,
"semantically_fungible": false,
"scope_divergence_flag": false,
"void_rule_divergence_flag": true,
"settlement_source_divergence_flag": true,
"settlement_timing_divergence_flag": true,
"polymarket_rule_override": "If the game is canceled entirely, with no make-up game, this market will resolve \"No\". …",
"kalshi_rule_override": "If the game is cancelled or rescheduled to over two weeks away, the market will resolve to a fair price …",
"findings": [
{ "code": "void_rule", "severity": "fatal",
"detail": "Void conditions differ … A core void asymmetry can void one leg while the other settles.",
"affects_fungibility": true },
{ "code": "settlement_source", "severity": "warn",
"detail": "fifa.com (UMA-adjudicated) vs Kalshi internal …" },
{ "code": "settlement_timing", "severity": "warn",
"detail": "2h statistics fallback vs unspecified …" }
],
"fee_adjusted_true_spread": 0.0009,
"snapshot_age_seconds": 61,
"execution_verdict": "block",
"verdict_reasons": [ "void_rule: Void conditions differ …", "…" ],
"ruleset_sha": "…"
}That block is the product working as intended: a canceled match pays the Polymarket leg $0 while Kalshi resolves the same position to "a fair price" — the hedge breaks exactly when you need it.
list_covered_eventsFree discovery of everything check_resolution_risk can audit: the covered canonical events, their per-outcome pair ids (…#home / #draw / #away), the Polymarket conditionIds and Kalshi tickers with outcome labels for each pair, match dates, the frozen ruleset_sha pinning the identity graph, the coverage kickoff cutoff, and snapshot freshness. Takes no arguments. Use a returned pair_id (or a pair's two market ids) as input to check_resolution_risk.
| Verdict | Meaning | Agent action |
|---|---|---|
safe | No fatal findings, data fresh. The two contracts settle on the same thing. | Proceed; your stack decides. |
caution | Non-fatal divergence or degraded data (e.g. stale snapshot, thin liquidity, source/timing gaps). | Proceed only if your strategy tolerates the named finding. |
block | A fatal divergence (void-rule asymmetry, scope mismatch, granularity mismatch…) can settle the legs differently. | Do not execute the pair as a hedge. |
Every FSAO carries the findings that produced the verdict — the audit is explainable, append-only logged server-side, and reproducible from {ruleset_sha, snapshot_ts, canonical_event_id}.
FIFA World Cup 2026 match markets listed on both Polymarket and Kalshi, expanding in batches throughout the tournament. Current coverage is always live — query list_covered_events, or /v1/events and /v1/health on the REST API. Listings here are intentionally coverage-generic; the endpoint is the source of truth.
The MCP tools are free at launch — no key. The keyed REST API (POST /v1/audit, same FSAO) issues free opening-weekend keys by mail: [email protected].
Crosswire is an advisory data service. Nothing it returns is financial advice.
MIT — covers this repository (listing metadata and documentation). The Crosswire service itself is proprietary.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.