feat(agent-readiness): revisit the Payments layer (UCP/ACP/x402 discovery) — scored N/A today; all-or-nothing scoring math

## Finding (ora.ai scan 2026-07-06)

Our Payments layer is **N/A** (6 checks not scored). Prod probes confirm nothing is served:

- `https://worldmonitor.app/.well-known/ucp` → **404**
- `https://www.worldmonitor.app/.well-known/acp.json` → **404**
- `/checkout_sessions` → 200 text/html (SPA catch-all soft-404)

#3313 (ACP) and #3314 (UCP) were closed 2mo ago on the "we sell subscriptions, not carts" decision — so this is a deliberate gap, not a regression.

## What Telnyx (#1, 92/A) does

Banks **12/12 + bonuses** with discovery surfaces and *stub* endpoints — no real agentic checkout backend:

- x402 payment challenge on an API GET (PAYMENT-REQUIRED header)
- `/.well-known/ucp` profile (version 2026-04-17, 3 capabilities)
- ACP `/checkout_sessions`: OPTIONS 204 allows POST, **POST 501** with an ACP-shaped error
- ACP `/agentic_commerce/delegate_payment`: same OPTIONS 204 / POST 501 pattern

## Scoring math — do this all-or-nothing

Scores normalize over scored layers. Today: 80/88 = 90. A **full** Payments pass adds 12/12 to both sides. A **partial** implementation (say 6/12) would *lower* the normalized score. So: either ship the complete discovery set (UCP profile + ACP endpoints + x402) in one PR, or leave the layer N/A.

Legitimacy: we do sell (Pro/API tiers via Dodo) — a real ACP surface could eventually front the actual checkout, not just stubs. Decision needed on stance (ship full stub set now vs wait for real agentic checkout vs keep N/A).

Related: #3313, #3314, #4952.

Start by reviewing the related issues #3313, #3314, and #4952, then inspect the listed UCP, ACP, and x402 probe paths and the existing subscription checkout stance. Done means the project has decided whether to ship the complete discovery set, wait for real agentic checkout, or keep Payments N/A; a shipped implementation must cover UCP, both ACP endpoints, and x402 together.

Written by the indexing model from the issue text.