huashu-script-polish skill

Overview

This skill polishes video scripts to sound natural and speakable by removing written-style phrasing and AI-like wording. It focuses on making scripts easy to read aloud, keeping sentences short, and adding conversational cues so the text matches on-camera delivery. The result is a script that flows like a real person speaking on camera.

How this skill works

The skill runs a three-pass review tailored for video: content check for accuracy and timing, core oralization to replace formal phrases with conversational alternatives, and final filming annotations for cuts and pauses. It rewrites long or dense sentences into short, punchy lines, inserts natural fillers when helpful, and flags places that need visual cues or practice reads. Output is ready for direct use in recording sessions or teleprompters.

When to use it

• After a script first draft is complete
• When someone says "too formal" or asks for "口语化" (make conversational)
• Before a recording session to ensure smooth delivery
• When the script sounds like writing or like AI-generated text
• When you need timing estimates and cut/close-up markers

Best practices

• Keep sentences under 20 words; prefer periods over long commas
• Read the revised script aloud to catch remaining awkward spots
• Use common spoken phrases (e.g., "you see", "actually", "so") sparingly to sound natural
• Mark where to cut to screen or zoom for each key step
• Balance natural fillers with clarity—avoid sounding vague

Example use cases

• Turn a technical explainer into a friendly on-camera walkthrough
• Convert a promotional script to a casual host monologue
• Prepare a tutorial script so each step reads smoothly when narrated
• Add filming cues and pause marks for a single-take recording
• Polish AI-drafted copy that feels mechanical into human-sounding narration

FAQ