Reels Scripting

Reels Scripting

Codex and Claude runtime

• Use this skill in Codex or Claude with the tools actually available in the current task. AskUserQuestion examples describe the questions, not a required API: use an available question tool within its limits, or ask in chat. Reuse answers and source material already supplied.
• Work in the user-selected project. Read its about-me.md, voice.md and relevant brand files before personalised work. Confirm the intended author if files conflict or contain starter defaults. Ask for missing facts or run voice-builder; never inherit the maintainer's identity, accounts or private files.
• Resolve bundled references/ relative to this skill folder. For an explicitly requested profile refresh, read and update the canonical about-me.md, voice.md or newsletter-voice.md in place, preserving unrelated user facts and rules. Consumers must reread those canonical files. Use a new filename only for new deliverables that would collide with unrelated existing files. Installation alone never starts an interview or writes files. Do not write persistent learnings unless requested.
• Use supplied evidence first. Verify external claims through available search/source tools when needed. If a source or integration is unavailable, name the missing capability and offer supplied text/export input. Never invent facts, first-person experience, metrics or a successful tool run.
• Connect only services needed for the chosen route through the user's existing account. Never print credentials or overwrite connections. Drafting, saving and reviewing do not authorise publishing, sending messages or changing accounts.

CRITICAL: Auto-start on load

When this skill triggers, go straight to Step 1. Do not summarise.

Prerequisites

The full Apify + Gemini video-analysis route needs:

• APIFY_API_TOKEN environment variable (Instagram scraping)
• GOOGLE_AI_API_KEY environment variable (Gemini 2.5 Flash video analysis)
• Node.js 18+ and the apify-client and @google/generative-ai packages

Check only the capabilities needed for the chosen route. For URL extraction and Gemini video analysis, verify the user's existing credentials are present without printing them, then verify the provider and model are available. Missing credentials, packages, a private video or an unavailable model leave that stage pending. Explain what is missing; do not modify account configuration or silently substitute a provider.

A supplied video can skip Apify. A supplied transcript/analysis can skip Steps 3 and 4 and support a labelled transcript-based script in Step 5, but is not a Gemini video-analysis result. Reuse installed packages; check current official SDK and actor documentation before generating a helper script, and get the user's approval before incurring provider costs.

Step 1. Get the reference

Ask:

Paste the reference Reel URL or Notion link. This is the outlier Reel you want to reverse-engineer the format from.

Wait for the URL.

If the user pastes a Notion link, read it through an available authorised Notion connector, locate the Instagram Reel URL on the page, and extract it. If no Reel URL is found on the Notion page, ask the user to paste the Reel URL directly.

Step 2. Get the newsletter topic

Ask:

What's the topic from your newsletter you want to repurpose into this Reel? Paste the relevant newsletter section, or type the core idea in a sentence.

Wait for the topic. Read newsletter-voice.md, voice.md, and about-me.md from the project if they exist, so the script matches the user's voice.

Step 3. Scrape and download the Reel

Use a new outputs/reels/<slug>/ folder in the selected project. Reuse an existing suitable helper if present, otherwise write a small task-specific Node.js script there that:

1. Uses apify-client and the existing apify/instagram-reel-scraper route, after checking its current documented schema, for the single requested Reel. Never scrape comments, replies or comment threads, including via deepScrape or numComments. Verify the provider supports post/video-only collection before calling it; otherwise request a video upload. Do not guess alternate actor inputs.
2. Extracts videoUrl from the returned item.
3. Downloads the video to outputs/reels/<slug>/downloads/{username}_{shortCode}.mp4.
4. Saves raw scrape data to outputs/reels/<slug>/reel_data_{shortCode}.json.

Run the script. Confirm file size and metadata (views, likes, aggregate comment count if available, caption first 200 chars) before continuing.

A single Reel is a reference, not a verified outlier. Call it an outlier only when supplied or explicitly requested comparison data establishes its performance relative to the same creator’s recent posts; cite the sample, window and multiple. Do not invent that baseline.

Step 4. Analyse with Gemini 2.5 Flash

Extend the Node script (or run a second pass) that:

1. Reads the downloaded .mp4 as base64.
2. Calls genAI.getGenerativeModel({ model: 'gemini-2.5-flash' }).
3. Sends the video with this exact prompt:

I'm studying this Reel to write my own script in a similar style for my audience of [AUDIENCE FROM about-me.md].

## Full Transcript
- Transcribe EVERY word with timestamps

## Hook
- Exact first words spoken
- Word count of the hook
- What makes it stop the scroll?

## Language Patterns
- Average sentence length
- You/your vs I/me ratio
- Transitions between points
- Where are the 'just' minimisers?

## Structure
- Total duration
- Section breakdown with timings
- What's the before/after moment?
- What's the CTA?

## One key insight
- The single most important technique to learn from this Reel

Save the analysis to outputs/reels/<slug>/analysis_reference_{shortCode}.md.