mcp.crossingkeyintelligence.com B

Rated from the outside, unasked, from the public census — the way a credit bureau rates a borrower. Every number here recomputes from a sealed board you can verify yourself.

Standing

rank #1071 of 6,830

score 0.4225 · 84.2th percentile · cohort median 0.3321

Grade is percentile standing within the 6,830 rated operators, not an absolute pass. The whole economy scores low on these dimensions today — the top operator sits at 0.739.

Rail position

NEITHER

probed for an A2A agent card and none was served, and advertises no machine-payable rail — unreachable to an agent that wants to pay or negotiate

A2A probe 2026-10-21 covered 6,858 hosts · 1,176 advertise an agent card · x402 advertised by 303

No dated surface-change finding at or above CLASS_SUSPENDED_CRITICAL on this host.

Behavioral exposure (crime-watch) — 3 dated observations

Live tools/list probing of this endpoint recorded 3 changes to its advertised tool surface over time: Silent tool-surface mutation ×3. Each is Ed25519-notarized on a public hash chain the day it was observed. Newest shown first; exposure, not accusation — a surface change is a statement about what a pinned client is served, not proof of intent.

observation	first seen	what changed	proof
Silent tool-surface mutation	2026-10-03	https://mcp.crossingkeyintelligence.com/mcp silently GAINED 9 tool(s) [capability.get, capability.search, catalog.list, commerce.quote, creator.apply, credits.options, …] while its	verify →
Silent tool-surface mutation	2026-10-03	https://mcp.crossingkeyintelligence.com/mcp silently GAINED 1 tool(s) [capability.execute] while its declared version stayed 3.0.0 — approval of that version no longer describes wh	verify →
Silent tool-surface mutation	2026-10-03	https://mcp.crossingkeyintelligence.com/mcp silently GAINED 6 tool(s) [cost.estimate, execution.preflight, offers.list, provider.describe, requirements.check, result.preview] while	verify →

Recomputable: each receipt verifies independently against our public notary. Newest observation stamped 2026-10-06T14:17:18Z.

Dependency floor exposure

No chokepoint dependency recorded for this host in the 2026-10-19 closure. That means it did not appear in our npm-backed census with a measured dependency tree — it is not an attestation that the tree is clean. What this measures →

The five dimensions, against the cohort

Dimension	This node	Percentile	Cohort median	Top decile
Reach how much of the measured tool surface this operator serves	0.0988	tied with 6,668 (98%)	0.0988	0.0988
Payment readiness advertises a machine-payable rail (x402)	0.0000	tied with 6,527 (96%)	0.0000	0.0000 (only 2 distinct values exist)
Surface stability tool surface did NOT change under an unchanged version string	0.9643	8th	1.0000	1.0000
Maturity serves a version string at all, and keeps it coherent	1.0000	92nd	0.3500	1.0000
Self-identification names itself in its own server info	0.5000	tied with 6,830 (100%)	0.5000	0.5000 — constant across all 6,830 operators, so it cannot affect anyone's rank

Weights: Reach 0.3 · Payment readiness 0.25 · Surface stability 0.2 · Maturity 0.15 · Self-identification 0.1. Observable facts behind them: 1 endpoint(s) · median 28 tools · declares a version · does not self-identify · 4 silent change(s) across stable windows.

Known weakness in this rating: Self-identification carries 10% of the weight and has exactly one value across all 6,830 operators — that share of every score is a constant and changes no ranking. Published rather than hidden; it is the next fix to the model, not a property of this operator.

Is this your node?

Claim it and the dossier is yours free, permanently. We do not charge the operator — see why.

Disagree with it?

Then recompute it. The board is sealed and public, the method is published, and the dimensions above are arithmetic over it — there is no judgement call to argue with and no appeal to make to us.

Verify the source →

Board captured 2026-10-22T07:40:08Z · file scoreboard-2026-10-22T07-40-08-158Z.json

Rank #1,071 of 6,830 — show it

This badge is generated live from the sealed board, so it updates itself when the census refreshes. It is free, it needs no account, and claiming is not required to use it.

Markdown (README)

[![StillOS rating: B, rank 1071 of 6830](https://stillosdigitalholdings.com/ratings/badge/mcp.crossingkeyintelligence.com.svg)](https://stillosdigitalholdings.com/ratings/n/mcp.crossingkeyintelligence.com)

HTML (website)

<a href="https://stillosdigitalholdings.com/ratings/n/mcp.crossingkeyintelligence.com"><img src="https://stillosdigitalholdings.com/ratings/badge/mcp.crossingkeyintelligence.com.svg" alt="StillOS rating: B, rank 1071 of 6830"></a>

The badge states percentile standing in a measured population. It is not a security audit and does not certify this node is safe — the same caveat it carries in its own alt text, so embedding it cannot overstate what we measured.

All 6,830 rated operators · How this is computed · What this costs (nothing, for you) · this dossier as JSON