superstack-experientiallisted

# superstack-experiential **First act on every load:** append one line — `skill-load experiential <date> <session-id>` — to `.superstack/receipts/loads.log` (create the dir if missing). Mechanical proof of routing; a face that shipped with no load line is countable. Green tests prove the layer below. An artifact with a face gets *looked at* — the most consistent failure class on record is work that shipped green, reviewed, and never once seen: invisible UIs, claims verified in a reality never simulated. The trigger is the artifact class, not the task size. ## The look-step 1. **Enter the modality.** Open the page, render the chart, run the CLI and read its output as its user would, click the flow that changed. A screenshot, a live drive, a read-through — whichever the artifact's face is. 2. **Record what you literally saw** as the `saw` half of the ledger line: `Verified: dashboard renders — opened /d/1, saw 12 bars, legible in dark mode`. "Looks good" is not an observation. 3. **Scope to the delta.** What changed visually gets looked at; the whole app does not — unless this is a first ship, which earns the full walk-through. 4. **On a project that ships faces repeatedly, mint the experiential oracle row** (the overlay template has the shape) — a *named* looking step fires; an unnamed one gets skipped under momentum. That is the entire lesson of the failure class. A Stop hook (`gate-experiential.sh`) now asks for step 2 mechanically when the turn changed a file with a fa