{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/cbdb48ff270a4cef9eecde93a8c410d2\" frameborder=\"0\" width=\"832\" height=\"624\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":624,"width":832,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":624,"thumbnail_width":832,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/cbdb48ff270a4cef9eecde93a8c410d2-8e8ccd6902bc8645.gif","duration":208.38,"title":"Ship30 Plugin Update, Deterministic Scoring Fix","description":"This Loom explains the changes made to the Ship30 plugin to improve deterministic scoring and eliminate JSON handoff issues when integrating into Cascade. The author rebuilt the framework as deterministic JavaScript in a code module, keeping the same scoring criteria including voltage formats, hook archetypes, opener types, and rhythms, and now runs faster and cheaper without JSON issues. They added Ship30 diagnostic outputs to content and the dashboard shows three new columns: score, issue, and alternative hooks. Most pieces are currently scoring about 8 to 16 out of 25, likely because Ship30 strips explicit numbers and named audiences. The author asks whether scoring should remain strict to Ship30 methodology or be calibrated to Brooke’s voice, and says compliance is clean and voice is mostly approved."}