Atinamos Verification
Core Principle You can pay to be tested. You cannot pay to be trusted.
Explore verification

First-class zero-spend evidence before purchase

Milestone occurred: 11 September 2026
Public milestone record published: 11 September 2026

What happened

Atinamos Verification added a reusable evidence class for independent zero-spend contract/discovery observations. The first published use case was TAT Risk.

During the observation, Atinamos inspected public discovery, OpenAPI, pricing, capabilities, health, readiness and MCP surfaces without using a wallet, creating a payment payload or invoking the paid endpoint.

The public x402 contract was consistent across the observed surfaces, the explicitly free/read-only MCP path succeeded, and GET /ready returned HTTP 500 INTERNAL_ERROR twice while GET /health remained HTTP 200.

Total spend was 0 USDC.

Why it mattered

Until this point most first-class Verification evidence centred on paid execution. The zero-spend class extends the evidence chain in front of purchase.

A buyer can now distinguish:

what the service publicly claims/exposes before spending
→ what happens if payment is actually attempted
→ whether settlement/fulfilment occur
→ whether correctness can be independently tested

Those layers remain separate rather than being collapsed into one verdict.

Evidence boundary

The TAT record is explicit:

discovery              OBSERVED
contract_consistency   OBSERVED_CONSISTENT
mcp                    OBSERVED
free_read_only_tool    OBSERVED_SUCCESSFUL
readiness_endpoint     NEGATIVE_OBSERVATION
payment                NOT_EVALUATED
settlement             NOT_EVALUATED
fulfilment             NOT_EVALUATED
correctness             NOT_EVALUATED
quality                 NOT_EVALUATED

The readiness discrepancy is not represented as “service down” because health remained HTTP 200 in the same observation window.

What it proved

This milestone supports that Atinamos can publish independent pre-spend evidence about:

  • discovery surfaces;
  • machine-contract consistency;
  • MCP availability and explicitly free/read-only tools;
  • readiness/health discrepancies;
  • exact boundaries around what was not tested.

The new class is exposed through the buyer evidence API without inflating paid-test, successful-fulfilment or correctness counts.

What it did not prove

The TAT observation did not evaluate payment, settlement, paid fulfilment, correctness, quality or provider-wide reliability. It is not seller-performance evidence for the paid route.

Primary record

Study 6 — TAT Risk Zero-Spend Contract/Discovery Observation contains the request IDs, observed build, contract details and the retained readiness finding.

Capability transition

contract/discovery evidence before spending
→ paid settlement/fulfilment evidence when authorised
→ independent correctness evidence where a defensible method exists

This is the current latest milestone. New service observations will continue to appear in the evidence corpus, but the milestone timeline changes only when Atinamos proves a materially new capability.

← Previous milestone: unattended generic Assurance · View the full timeline