What problem does it solve? AI-generated video characters often read as wooden, dead-eyed, or emotionally flat because prompts describe displayed emotion instead of playable behavior. This Skill rewrites the performance layer of a video prompt as objective-driven behavior — objectives, obstacles, tactics, beats, subtext, body, and eye life — so characters stay consistent and alive across shots. ## Core Features & Use Cases - Acting master profiles: Produces a reusable 150–220-word character profile with fixed block order covering physique, vocal profile, triggered tics, eye life, named gait, and a mask-and-crack transformation clause. - Per-scene adaptation: Rewrites the master profile into each scene's posture, action, and beat while keeping identity constant, and locks a verbatim voice prompt per character. - Quality control: Ships a 15-symptom atlas of bad acting with prompt-level fixes and a 0–5 performance self-check scale targeting level 4+ on hero shots. - Use Case: When a Seedance prompt produces a character who stares blankly and gestures theatrically, use this Skill to rebuild the CHARACTER ACTING block with an objective, 2–4 visible beat changes, listening reactions, and continuous eye life. ## Quick Start Ask the AI to write an acting master profile and per-scene performance paragraph for your character using the higgsfield-acting rules before inserting it into your Seedance video prompt.
