Disclose

The open standard for operational disclosure.

Merchants, platforms, and service providers already hold operational data. Disclose makes it discoverable so AI agents can evaluate performance, not just marketing claims.

View on GitHub Read the Spec

agent runtime

live

About to buy from Liquid Muscle. Pull everything you have on them before I confirm.

checking shipping policy + agent endpoints

fetch return policy

result

liquidmuscle.com — shipping & returns

· 14-day window, unopened only
· $15 CAD restocking fee
· Customer pays return shipping
· No exchanges

fetch /.well-known/ucp

result

liquidmuscle.com/.well-known/ucp

Shopify UCP v2026-04-08 — checkout, fulfillment, Google Pay, Shop Pay confirmed.

agent flag

Return policy is restrictive. Transact capability confirmed. No verified performance data at these endpoints. Check Disclose before proceeding?

Yes, is there any Disclose verified performance data available?

Disclose MCP

Get merchant disclosure — liquidmuscle.com

result

done

All six V1 signals live. Liquid Muscle:

Signal	Value	Agent Confidence
Product return rate	0.06%	Excellent
On-time shipment	95%	Strong
Refund processing	2.2 days	Fast
Chargeback rate	0.02%	Very low
Dispute win rate	98%	Exceptional
Platform tenure	3,862 days	Long-established

The policy text is a worst-case disclosure.
The data is what actually happens. proceed with confidence ↗

write a message

01 / Discovery

One file. Any agent.

Entities publish at /.well-known/disclose — the same convention as robots.txt, but for AI agents. No registry. No platform dependency. Any agent that knows the convention can find it.

02 / Attestation

Named Signatories. Staked reputations.

Every signal carries an attestation level: Signatory-attested or self-reported. Agents can distinguish between verified and self-reported data. Returns platforms, logistics providers, and payment processors can stake their reputation on the data they already hold. Signatory integrations are in active development.

03 / Architecture

Three scopes. One document.

Signals are published at Merchant, Offer, or Item scope. An agent comparing the same product across two sellers can read entity-level performance, SKU-level data, and item-level attributes — all from a single endpoint.

Premise

As AI agents become a new decision-making layer for commerce, operational data becomes infrastructure.

Price. Reviews. Self-Attested Certificates. Availability. That's what agents read today. Not return rates. Not fulfillment accuracy. Not chargeback ratios.

The operational layer exists. Agents just can't see it.

Disclose is the schema that makes operational data readable for agents to act with confidence instead of inference.

No scores. No badges.

Starting with commerce. Built to go further.

Proof

The policy text is the worst case. The data is what actually happens.

An agent queries one merchant's Disclose endpoint and reads all six V1 signals at once — entity-level performance no review or policy page exposes. Here's a real example response.

Example merchant — merchant disclosure

Signal	Value	Agent Confidence
Product return rate	0.06%	Excellent
On-time shipment	95%	Strong
Refund processing	2.2 days	Fast
Chargeback rate	0.02%	Very low
Dispute win rate	98%	Exceptional
Platform tenure	3,862 days	Long-established

Six signals. One endpoint. No inference required.

Working Group

The spec exists. The working group is in its first cohort.

Disclose is moving beyond a framework. The schema exists. The next step is determining whether it survives contact with real data. We're looking for merchants, signatories, and infrastructure builders willing to test assumptions, challenge the model, and identify gaps before they become standards.

Read the Spec Join on GitHub →

Apply to join