This week's UCP activity reveals a coherent strategy: building the trust infrastructure that enables enterprise adoption of agentic commerce. Three significant developments — eligibility verification, attestation extensions, and formalized totals contracts — address different facets of the same problem: how do businesses trust AI agents enough to transact with them?
When a human shops online, businesses have multiple trust signals: browser cookies, login history, shipping address consistency, payment method verification. When an AI agent shops on behalf of a human, most of these signals vanish. UCP must recreate — and improve upon — these trust mechanisms at the protocol level.
PR #250 establishes the contract for eligibility verification — how agents prove customer qualifications (student, employee, member) without revealing identity.
PR #264 adds cryptographic attestation for eligibility claims — enabling third parties to vouch for customer status with verifiable signatures.
PR #261 formalizes how order totals are calculated and communicated — ensuring agents and businesses agree on pricing before checkout.
The eligibility system enables restricted discounts (student pricing, employee discounts, membership perks) to work in agentic commerce without exposing personal data:
{
"eligibility": {
"claims": [{
"type": "student",
"issuer": "university.edu",
"verified_at": "2026-03-01T00:00:00Z"
}],
"attestations": [{
"claim_hash": "sha256:...",
"attestor": "sheerid.com",
"signature": "..."
}]
}
}
The key insight: attestation is separate from claims. A university issues the claim ("this person is a student"), and a verification service attests to its validity without needing the underlying identity data.
Cart totals seem straightforward until you consider edge cases:
PR #261 establishes a polymorphic totals contract — a flexible structure that can represent all these scenarios while giving AI agents deterministic expectations about what they'll pay.
Taken together, recent UCP developments form a coherent trust stack:
Each layer addresses a specific trust gap. Together, they provide the security infrastructure that enterprises require before allowing AI agents to transact at scale.
Businesses implementing UCP should prepare for:
The specification is converging toward a stable trust model. Early implementers should track these PRs closely — the patterns established now will define agentic commerce for years.