What problem does it solve? Short-form video creators struggle to objectively evaluate whether their hooks, scripts, and captions will hold viewer attention before publishing. This Skill provides a structured adversarial critique based on measurable retention proxies instead of vague intuition or unfounded virality claims. ## Core Features & Use Cases - Multi-Dimension Evaluation: Scores content across hook latency, specificity, open-loop strength, information density, stakes escalation, payoff timing, identity relevance, visual novelty, caption readability, and CTA specificity. - Structured YAML Output: Returns strengths, blockers, high-leverage changes, evidence gaps, and numeric scores for hook, completion proxy, shareability, and clarity, plus a pass/revise/fail verdict. - Epistemic Guardrails: Explicitly avoids inferring platform algorithm internals or inventing trend evidence, keeping scores as internal diagnostics rather than predictions of views. - Use Case: Before publishing a TikTok script, submit the hook, caption, and storyboard to receive a scored critique identifying which specific changes would most improve retention. ## Quick Start Ask the AI to critique your short-form video hook, script, and caption using the virality critic and return the structured YAML verdict.
