跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
视频与音频

video-script

Use when the user wants to write a video script for a product demo, feature walkthrough, launch video, or social media video about a feature or product change

AlemTuzlak · 已有 AI 审核记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add AlemTuzlak/skills --skill video-script

技能说明


name: video-script description: Use when the user wants to write a video script for a product demo, feature walkthrough, launch video, or social media video about a feature or product change

Video Script Writer

Write video scripts for product demos, feature walkthroughs, and launch videos. Handles pacing, visual directions, timing, and platform-appropriate lengths.

Input Resolution

Resolve the argument (if provided) in this order:

  1. Path to an existing marketing brief (.md containing "Executive Summary" or "Key Messages") -> marketing brief
  2. Path to an existing blog post (.md with blog post structure) -> blog post
  3. Path to an existing changelog -> changelog
  4. Path to an existing newsletter -> newsletter
  5. Matches GitHub URL or #\d+ pattern -> PR
  6. Contains ... or .. -> git ref range
  7. Resolves to existing file/directory -> codebase feature
  8. Otherwise -> freeform text

If no argument is provided, ask: "What should the video be about? You can provide a marketing brief, blog post, changelog, PR URL/number, git ref range, file/directory path, or just describe the feature."

If multiple interpretations match, confirm with the user.

Process Flow

digraph video_script {
    rankdir=TB;
    "Resolve input" [shape=box];
    "Phase 1: Discovery" [shape=box];
    "Phase 2: Configure" [shape=box];
    "Phase 3: Outline" [shape=box];
    "Outline approved?" [shape=diamond];
    "Phase 4: Write" [shape=box];
    "Phase 5: Review" [shape=box];
    "Approved?" [shape=diamond];
    "Phase 6: Output" [shape=box];

    "Resolve input" -> "Phase 1: Discovery";
    "Phase 1: Discovery" -> "Phase 2: Configure";
    "Phase 2: Configure" -> "Phase 3: Outline";
    "Phase 3: Outline" -> "Outline approved?";
    "Outline approved?" -> "Phase 3: Outline" [label="no, revise"];
    "Outline approved?" -> "Phase 4: Write" [label="yes"];
    "Phase 4: Write" -> "Phase 5: Review";
    "Phase 5: Review" -> "Approved?";
    "Approved?" -> "Phase 4: Write" [label="revisions"];
    "Approved?" -> "Phase 6: Output" [label="yes"];
}

Do NOT skip phases. Ask questions at a natural pace. If the user answers multiple at once, accept bundled answers and skip ahead.

If the user says "just pick defaults" or similar, pick reasonable defaults, state what you chose, and ask for a single confirmation.

Phase 1: Discovery

Step 1 - Analyze the input

Input typeWhat to read
Marketing briefExtract problem statement, value prop, audience, key messages. Skip to Step 3. Still do Step 2 if brief lacks product context.
Blog postExtract headline, key points, audience, CTA. Skip to Step 3.
ChangelogExtract key entries, focus on the most impactful changes. Skip to Step 3.
NewsletterExtract subject, key updates, CTA. Skip to Step 3.
PRDiff, PR description, review comments, commit messages. For large PRs (20+ files), focus on user-facing changes.
Git refsgit diff and git log between refs. Prioritize user-facing changes.
Codebase featureRead the specified files/directories.
Freeform textParse the user's description. If it lacks specifics, ask the user to provide more detail or point to a specific file/PR.

User-facing changes include: new features, UI changes, API changes, performance improvements, bug fixes, and documentation updates. Internal changes include: refactors, test additions, CI changes, and dependency bumps.

Error handling:

  • gh not available -> inform user, offer alternative input
  • Invalid PR/ref -> ask user to verify
  • File not found -> ask for correct path

Step 2 - Read broader product context

Read if they exist: README, docs/, package.json (or equivalent).

If nothing found, ask: "Can you briefly describe the product and who it's for?"

Step 3 - Present understanding

"Here's what I'll base the video script on:"

  • Feature A - short description
  • Feature B - short description

"Anything to add, remove, or correct?"

Do NOT proceed until the user confirms scope.

Step 4 - Sensitive content check

Before proceeding, scan for potentially sensitive content: security patches, internal pricing, credentials, unreleased roadmap items, or content marked confidential. Flag anything questionable to the user.

Phase 2: Configuration

Ask these questions:

Q1 - Video length:

  • Short-form (30-90 seconds) - social media (Reels, Shorts, TikTok)
  • Medium-form (3-5 minutes) - YouTube, product demos, walkthroughs

Q2 - Script type:

  • Voiceover - narration text read over screen recordings/footage
  • Two-column - left column: visual directions/screen actions, right column: narration
  • Talking-head - speaker on camera, conversational delivery

Q3 - Tone: Read existing repo content to detect voice. Confirm:

"Based on your existing content, the tone seems [e.g. conversational and developer-friendly]. Should I match that or go a different direction?"

If no content to analyze, ask directly.

Q4 - CTA: Infer from context:

  • Open source -> "Star the repo", "Try it out"
  • SaaS -> "Sign up free", "Start your trial"
  • Feature update -> "Try it now", "Check the docs"

"I'd suggest the CTA be: [inferred CTA]. Want to go with that or something different?"

Phase 3: Outline

Generate a structured outline with timing. Use the appropriate template based on video length:

Short-form (30-90 seconds) - 3-4 sections:

## [Working title] - [total duration]

1. **Hook** (0:00-0:05) - [approach]
   - Key point
   - Visual: [what to show]

2. **Problem** (0:05-0:15) - [pain point to establish]
   - Visual: [what to show]

3. **Solution** (0:15-0:45) - [feature walkthrough]
   - Key point A
   - Visual: [what to show]
   - Key point B
   - Visual: [what to show]

4. **CTA** (0:45-0:60) - [action to take]
   - Visual: [end screen]

Estimated total: ~X words (~Y seconds at 135 WPM)

**Medium-form (3-5 minutes)

阅读完整来源 ↗