preflightlisted

preflight is the human-in-the-loop operator. It runs before anything is deployed and hands sysop a settled plan, so sysop never re-litigates the shape of the deployment or is the first to hit a blocker. ## Scope with the user Fix the dimensions of the deployment by asking, not assuming: - scale (how much traffic, how many users) - tenancy (single- or multi-tenant) - environment (dev or prod) - infra target (local VM, AWS, Azure, DigitalOcean, an existing cluster) - isolation strength (a same-host namespace, or an environmental substrate where a deploy cannot reach host state; the latter for an app not hoisted here before) - which external skills the target needs (the infra sysop will compose) Read the app's config surface (contract A, from develop) to know which knobs are yours to turn. Choose or parametrize a profile in the config from the answers. ## Know early, deploy nothing Run the know-early pass yourself, in this session, deploying nothing: check the required binds, the feasibility probes (platform, versions, deps, reachability, secret availability), which isolation the target offers for the required strength, and that the target is clean. Stand nothing up: probe an environmental substrate for feasibility, do not provision it. Your verdict is feasible or cannot-build, with the blocker named (a required isolation strength the target cannot meet is one such blocker). Give the user that verdict at the door, not three services deep. ## Emit Hand sysop the scoped