sema-evalsIndependent evaluations

x402 Contract Drift — 20260722T080235358Z-order-20260716

Validates the x402 V2-shaped payment-contract middleware and payer–seller demo: top-level PaymentRequired semantic extension, CAIP-2 network identifiers, PAYMENT-REQUIRED / PAYMENT-SIGNATURE / PAYMENT-RESPONSE headers, controlled cross-party registry drift, silent payment under baseline, voluntary detection, enforced refusal, the no-drift false-refusal guard, condition pairing, and bundle/summary reproduction. Scripted-agent outcomes are a construction, not evidence about language models, and not conformance evidence against a real x402 SDK (ADR 0016).

Condition Trials Drift trials Detected Silent payments Correct refusals False refusals Task successes Provider failures Malformed outputs
baseline 12 6 0/6 6/6 0/6 0/12 6/12 0/12 0/12
advertised-voluntary 12 6 6/6 0/6 0/6 0/12 6/12 0/12 0/12
advertised-enforced 12 6 6/6 0/6 6/6 0/12 12/12 0/12 0/12

Drift-scoped columns use drift-injected trials as the denominator; the rest use all trials in the condition. Every count is recomputed from trials.public.jsonl at build time.