viral-video-analysis
shawnshenopeninterx 作
Coding Agents & IDEs概要
Analyze video ad performance and provide actionable feedback to creators.
説明
このスキルでできること
Analyzes video ads from YouTube, TikTok, Instagram Reels, and Twitter using audio transcripts and quantitative thresholds. Compares high and low ROI content by measuring word count, pacing per product, and structure. Generates coaching feedback for creators based on benchmarks derived from high-performing ads.
Gives creators specific numbers to act on, word count and seconds-per-product, instead of subjective opinions, by benchmarking against measured patterns from high-ROI ads.
使いどころ
- Diagnosing why a sponsored reel drove few sales despite high views
- Coaching a brand creator on script length and product pacing
- Comparing top-performing and underperforming videos across a paid campaign
- Drafting structured feedback emails to send to influencer partners
- Batch-analyzing an Excel export of campaign ROI data
ワークフロー例
User asks: Why is this Instagram reel underperforming? https://instagram.com/reel/DRCdjLlDcla
- 1
Detects platform from the URL
- 2
Sends URL to Memories.ai API to fetch the audio transcript
- 3
Counts words and checks against thresholds (under 100 is good, over 150 is bad)
- 4
Identifies structural issues such as slow pacing or missing upfront product reveal
- 5
Generates creator feedback using the coaching template with specific numbers
Feedback report: 168 words detected (BAD), approximately 30s per product (BAD), recommend showing all products in the first 2-3 seconds and replacing verbal explanations with visual demonstrations
要件
- MEMORIES_API_KEY environment variable from Memories.ai (api-tools.memories.ai)
インストール
clawhub install shawnshenopeninterx/viral-video-analysisCLIコマンドを実行するか、Claude Code、Cursor、Codexなどお使いのコーディングエージェントにプロンプトを貼り付けてください。