Short Video Hook Lab
Skill Card
- - Category: Creator Content
- Core problem: How to improve first-3-second retention?
- Best for: Hook iteration before recording
- Expected input: Raw script or opening lines + target audience + product angle
- Expected output: Ranked hook variants with rationale and filming complexity
- Creatop handoff: Push winning hook into Creatop script composer
Overview
Generate high-retention openings that are specific, shootable, and platform-native.
Workflow
1) Capture context
Collect:
- - target audience
- core promise
- desired action (watch/comment/click/buy)
- platform (TikTok / Reels / Shorts)
If key context is missing, infer minimal defaults and mark assumptions.
2) Generate 12 hook variants
Cover 6 families:
- - contrarian
- checklist
- mistake/fix
- proof-first
- personal observation
- urgent timing
Evidence rule for proof-first hooks:
- - Use real metrics only when verified.
- If not verified, avoid precise numbers and use non-fabricated wording.
3) Score and select
Score each hook (1–5):
- - clarity
- curiosity
- specificity
- filmability
- promise-action fit (optional tie-break)
Pick Top 3 + 1 backup.
4) Build micro execution pack
For each selected hook, provide:
- - hook line
- 10-second flow (3 bullets)
- on-screen text
- CTA line
- filming complexity (low/medium/high)
Output rules
- - Keep language conversational and human.
- Avoid clickbait that cannot be fulfilled.
- Prefer concrete wording over vague hype.
- If uncertain, choose clarity over cleverness.
License
Copyright (c) 2026 Razestar.
This skill is provided under CC BY-NC-SA 4.0 for non-commercial use.
You may reuse and adapt it with attribution to Razestar, and share derivatives
under the same license.
Commercial use requires a separate paid commercial license from Razestar.
No trademark rights are granted.
短视频钩子实验室
技能卡片
- - 类别: 创作者内容
- 核心问题: 如何提升前3秒留存率?
- 最佳用途: 录制前的钩子迭代
- 预期输入: 原始脚本或开场白 + 目标受众 + 产品角度
- 预期输出: 带评分依据和拍摄复杂度的排序钩子变体
- Creatop交接: 将获胜钩子推送至Creatop脚本编辑器
概述
生成具体、可拍摄且适配平台的高留存开场。
工作流程
1) 捕获上下文
收集:
- - 目标受众
- 核心承诺
- 期望行为(观看/评论/点击/购买)
- 平台(抖音/Reels/Shorts)
若关键上下文缺失,则推断最小默认值并标记假设。
2) 生成12个钩子变体
覆盖6大类型:
- - 反常识
- 清单式
- 错误/修正
- 证据优先
- 个人观察
- 紧迫时机
证据优先型钩子的规则:
- - 仅在使用已验证的真实数据时采用。
- 若未验证,避免精确数字,使用非虚构措辞。
3) 评分与筛选
为每个钩子评分(1-5分):
- - 清晰度
- 好奇心
- 具体性
- 可拍摄性
- 承诺-行为匹配度(可选,用于打破平局)
选出前3名 + 1个备选。
4) 构建微执行包
为每个选中的钩子提供:
- - 钩子文案
- 10秒流程(3个要点)
- 屏幕文字
- 行动号召文案
- 拍摄复杂度(低/中/高)
输出规则
- - 保持语言口语化和人性化。
- 避免无法兑现的标题党。
- 优先使用具体措辞,而非模糊吹嘘。
- 不确定时,选择清晰而非巧妙。
许可协议
版权所有 © 2026 Razestar。
本技能根据 CC BY-NC-SA 4.0 协议提供,仅限非商业用途。
您可署名Razestar后重复使用和改编本技能,并基于相同协议分享衍生作品。
商业用途需从Razestar另行购买付费商业许可。
不授予任何商标权利。