Will tonight's scheduled run actually be allowed to save its work? This service probes the unattended execution surface itself — with your task's exact payload class — before the run, and returns a signed PARITY / NO-GO verdict with the failing gate named.
The unattended surface is a different runtime from the interactive one. This probe runs a minimal deterministic fixture — read, write one small fixture, read back, revert — through the unattended/scheduled path itself, never interactively, and classifies the result from observable signals only into a six-gate taxonomy: HOST-SAFETY-REFUSAL, SESSION-EXPIRED, APPROVER-ABSENT, PROVIDER-REJECTED, HANDOFF-MISSING, or CORRUPT-SUCCESS. The verdict is signed and bound to the surface fingerprint, stored probe-side, and expires when that fingerprint changes, not on a clock. PARITY is a GO; anything else is a NO-GO that routes the work product to draft-and-queue instead of burning the run.
Operator pass ($9.99/mo) and Fleet pass ($19.99/mo) are GATED — not advertised or sold until the field-settling gate clears (real-platform trigger fidelity + per-triggered-run platform cost measured on file). No pass is sold in the meantime.
Stated plainly, because the lab proved the boundaries: verdicts do not transfer across payload classes (in the lab, a small payload was PARITY and a large one NO-GO on the same surface); the lab proof is on surface doubles, and real-platform trigger fidelity — whether an on-demand triggered run carries the same policy context as a naturally-fired one — is the first field measurement, recorded in the launch checklist, not assumed; flip frequency on real platforms, willingness to pay, and the addressable fleet set are unmeasured.
Support: support@textforagents.com