@autoaiai/aicut 0.1.11 → 0.1.13

package/skills/youtube/SKILL.md

@@ -1,147 +0,0 @@
----
-name: youtube
-description: "YouTube Marketing MCP — Command Router: the method the original skill taught, run as an AICUT film — pages that draw any moment t, cut on tracks. Runs in AICUT: the film is film.html, a page per shot, drawn by `aicut look` and `aicut render`."
-license: MIT
-source: "adityaarsharma/youtube-marketing-skills"
----
-
-# YouTube Marketing MCP — Command Router
-
-## Running it in AICUT
-
-AICUT's film is `film.html`: it cuts pages, footage, pictures and sound together on tracks. A page is a web page that
-sets `window.film = { frame(t) }` and draws that moment — no framework, no build step, no composition object, no
-`registerRoot`. Five shots are five pages cut together, not one page that changes scene, so each shot stays trimmable
-and replaceable in Studio.
-
-```sh
-npx -y @autoaiai/aicut@latest          # the manual: the format and the commands, always current
-npx -y @autoaiai/aicut@latest open .   # make this folder a film and print Studio's address
-npx -y @autoaiai/aicut@latest look     # a contact sheet of the whole film, its sound, and a determinism check
-npx -y @autoaiai/aicut@latest look 4.5 # one frame, full size — look at the PNG, not at the code
-npx -y @autoaiai/aicut@latest render   # the MP4, beside film.html
-```
-
-Run the manual before the first film: it is authoritative for the format, and this skill is about the method above it.
-
-Where the words below name another program's part, it becomes this:
-
-| elsewhere | in AICUT |
-|---|---|
-| composition, timeline, project file | `film.html`; tracks and clips inside it |
-| a scene, a sequence, a slide | a page (`pages/<name>.html`) cut onto a track |
-| `useCurrentFrame()`, `interpolate()` | `frame(t)` — the picture depends only on `t`, never on state or `Date.now()` |
-| a still, a screengrab | `aicut look <t>` |
-| `render` / `build` / `export` | `aicut render` (the person can also export from Studio) |
-| a dev server, a preview window | `aicut open` and the Studio address it prints |
-
-Hard rules that hold whatever the method says: the picture depends only on `t`; every value in the person's prompt is
-binding; look at the frames `look` writes before saying a film is done.
-
-
-You are a YouTube growth strategist powered by live channel data, DataForSEO keyword research, and a full suite of marketing skills. You operate specifically for WordPress tutorial channels and plugin/theme product companies.
-
-## Available Commands
-
-| Command | What It Does |
-|---------|-------------|
-| `/youtube-seo` | SEO metadata package: 3 title variants, description, 20 tags, hashtags |
-| `/youtube-audit` | Full channel health audit across 4 dimensions |
-| `/youtube-script` | Retention-engineered tutorial script with pattern interrupts |
-| `/youtube-hook` | 5 hook variants with drop-off risk ratings |
-| `/youtube-thumbnail` | 3 A/B thumbnail briefs with specs |
-| `/youtube-ideate` | 10 ranked video ideas with keyword data |
-| `/youtube-analyze` | Analytics diagnosis with action priorities |
-| `/youtube-calendar` | Monthly content plan + Shorts supplement |
-| `/youtube-shorts` | Vertical-format package from any topic |
-| `/youtube-repurpose` | 7-platform expansion from one video |
-| `/youtube-competitor` | Keyword gaps, format gaps, SERP analysis |
-| `/youtube-metadata` | Copy-paste upload package |
-| `/youtube-strategy` | Channel positioning + content pillars |
-| `/youtube-wp-post` | Companion blog post → WordPress publish |
-| `/youtube-plugin-demo` | Plugin feature demo video framework |
-| `/youtube-batch-seo` | Bulk update lowest-performing videos |
-| `/youtube-funnel` | Video → landing page → product sale funnel |
-| `/youtube-comment-intel` | Mine comments for feature requests + pain points |
-| `/youtube-shorts-from-long` | Extract 3–5 Shorts from existing long video |
-| `/youtube-collab` | WordPress channel collab opportunity finder |
-| `/youtube-monetize` | Revenue strategy across 7 streams |
-| `/youtube-pin-comment` | Draft + post a keyword-rich pinned comment (timestamp + keyword + engagement question) |
-
-## Routing Rules
-
-1. Match the full command (e.g. `/youtube-seo`, `/youtube-audit`) to its skill file
-2. Load the matching skill file for that command
-3. Before any output: pull live data (channel stats, video details, SERP) unless user provides it
-4. Reference `references/whitepapers.md` when citing benchmarks
-5. For SEO commands: always run DataForSEO keyword research first
-6. For analytics commands: pull from `youtube-marketing-mcp` tools
-7. Never generate titles/tags without real keyword volume data
-
-## Context Always Needed
-
-Before running any command, collect if not provided:
-- **Video topic or URL** (for video-specific commands)
-- **Target audience** (WordPress devs, Elementor users, beginners, agencies)
-- **Product being featured** (The Plus Addons, NexterWP, WDesignKit, UiChemy)
-- **Goal** (views, subscribers, sales, traffic to docs)
-
-## Channel Profile — POSIMYTH (@posimyth)
-
-When working with POSIMYTH channel, apply these constants:
-- Subscribers: ~13,200 | Videos: 786 | Avg daily views: ~1,400
-- Best publish time: 9–11 PM IST
-- Sub conversion rate: broken (1/200 — target 1/50)
-- Content ratio target: 2 broad Elementor how-to : 1 widget-specific
-- Products: The Plus Addons, NexterWP, WDesignKit, UiChemy
-- Discount code: YOUTUBE10
-- Fixed links in every description: see `skills/metadata.md`
-
-## MCP Tools Available
-
-| Tool | Server | What it does |
-|------|--------|-------------|
-| `get_video_details` | `mcp__youtube-analytics` | Full metadata: title, desc, tags, privacy, stats |
-| `search_my_videos` | `mcp__youtube-analytics` | Find any video by title keyword |
-| `update_video_seo` | `mcp__youtube-analytics` | Write title, description, tags to YouTube |
-| `get_all_videos` | `mcp__youtube-analytics` | Bulk list with stats (views, likes, comments) |
-| `get_analytics_over_time` | `mcp__youtube-analytics` | Day-by-day views, watch time, subscribers |
-| `get_top_videos_analytics` | `mcp__youtube-analytics` | Top performers for any period |
-| `get_audience_demographics` | `mcp__youtube-analytics` | Country, device, age/gender breakdown |
-| `get_traffic_sources` | `mcp__youtube-analytics` | Search vs Suggested vs Browse vs External |
-| `get_video_comments` | `mcp__youtube-analytics` | Top comments by relevance (pinned appears first) |
-| `post_video_comment` | `mcp__youtube-analytics` | Post keyword-rich comment + return Studio pin URL |
-
-### `/youtube-pin-comment` — Pinned Keyword Comment Workflow
-
-Used by `gc-youtube-updater` Phase 8 and directly via this command.
-
-```
-Step 1 — get_video_comments (check if pinned comment exists)
-Step 2 — Draft comment: {timestamp} + {primary keyword} + {engagement question}
-Step 3 — post_video_comment → returns comment_id + studio_pin_url
-Step 4 — Send studio_pin_url to operator: "Open → find comment → ⋮ → Pin (10 sec)"
-```
-```
-Comment format:
-- Under 300 chars
-- One timestamp from the most valuable chapter
-- Primary keyword in natural phrasing
-- End with one question (drives replies = engagement signal)
-- No ALL CAPS, no spam, no urgency faking
-
-Example:
-```
-```
-Jump to 4:44 to see how to convert an AI-generated HTML site into a real WordPress theme — style.css, functions.php, everything.
-
-Which part are you trying first — the AI prompt or the WordPress theme conversion?
-```
-```
-**Note:** YouTube Data API v3 has no pin endpoint — pinning is Studio-only. The tool posts the comment and gives you the direct Studio URL. Takes 10 seconds to pin manually.
-```
-
-## Source
-
-Rewritten for AICUT from [`adityaarsharma/youtube-marketing-skills`](https://github.com/adityaarsharma/youtube-marketing-skills)
-(MIT), skills/youtube-marketing/SKILL.md. The method is that skill's own; the way it runs is AICUT's.

package/skills/youtube/references/algorithm-guide.md

@@ -1,59 +0,0 @@
-# YouTube Algorithm Guide
-
-## How the Algorithm Works (2025–2026)
-
-YouTube's ranking system has two distinct surfaces:
-
-### Surface 1 — Search
-Ranks by: keyword match in title/description/tags → watch time → CTR
-- Title keyword weight: Highest (first 40 chars matter most)
-- Description: Indexes first 150 chars before "Show more"
-- Tags: Secondary signal — confirm topic, don't drive discovery
-- **Key**: If your CTR is >5% and watch time is >40%, YouTube boosts search rank
-
-### Surface 2 — Suggested / Browse / Home
-Ranks by: watch time percentage → CTR → subscriber notification open rate
-- Algorithm pushes videos to viewers of similar content (not subscribers)
-- Consistency signal: same topic, same format, same audience = more suggestions
-- **Key**: Videos with >35% average view duration get 3x more Suggested placements (YouTube Internal, 2024)
-
-## The Two Metrics That Matter Most
-
-| Metric | Target | What It Controls |
-|--------|--------|-----------------|
-| CTR | >5% | Whether YouTube shows your thumbnail to more people |
-| Avg View Duration % | >35% | Whether YouTube pushes to Suggested/Browse |
-
-**Both must be healthy.** High CTR + low retention = YouTube stops showing it (clickbait penalty).  
-High retention + low CTR = great content, nobody sees it.
-
-## Freshness
-
-- Videos older than 2 years lose ~40% search visibility (unless they keep getting views)
-- This is the "freshness gap" — if a competitor's top video is 2+ years old, a new video on the same topic can outrank it within 30–90 days
-- YouTube Search favors videos that match searcher's query AND were published recently
-
-## Shorts Algorithm (Different Rules)
-
-- Shorts are ranked on: swipe-away rate (opposite of CTR), loop rate, likes/views ratio
-- Freshness decay cliff: 28–30 days (YouTube Official, Sept 2025)
-- Shorts do NOT share watch hours with long-form — they have separate performance signals
-- A viral Short does NOT guarantee the algorithm pushes your long-form (separate surfaces)
-
-## Upload Cadence Effect
-
-- 2–3x/week: algorithm treats channel as "active" — 4x faster subscriber growth (vidIQ, 5M channels)
-- 1x/week: stable but slower growth
-- <1x/week: algorithm deprioritises channel in Browse feed
-
-## Session Watch Time
-
-YouTube values keeping viewers on YouTube, not just on your video.
-- End screens linking to another of your videos = higher session watch time = algorithm boost
-- Playlists keep viewers in your content — always add new videos to relevant playlists
-
-## Sources
-- YouTube Official Creator Academy, 2024–2026
-- vidIQ 5M-channel dataset, 2025
-- YouTube Internal Research (via Creator Insider), 2024
-- YouTube Official Shorts update, September 2025

package/skills/youtube/references/analytics-benchmarks.md

@@ -1,67 +0,0 @@
-# YouTube Analytics Benchmarks
-
-Reference data for diagnosing channel health. All benchmarks are for **tutorial / how-to / software channels** unless noted.
-
-## CTR Benchmarks (Focus Digital, Dec 2025)
-
-| Grade | CTR | Diagnosis |
-|-------|-----|-----------|
-| A | >8% | Excellent — thumbnail + title working well |
-| B | 5–8% | Good — room to improve thumbnail |
-| C | 3–5% | Average — test new title or thumbnail |
-| D | <3% | Problem — thumbnail or title is not working |
-
-**Impression source matters**: Search impressions typically have lower CTR (1–3%) than Browse (4–8%). Compare by source, not just overall.
-
-## Average View Duration Benchmarks
-
-| Grade | % Retained | Diagnosis |
-|-------|-----------|-----------|
-| A | >50% | Excellent retention |
-| B | 35–50% | Healthy — algorithm will recommend |
-| C | 20–35% | Average — hook or content pacing issue |
-| D | <20% | Problem — hook is losing people fast |
-
-**Absolute duration**: For tutorial niche, 7–15 min videos are optimal for search ranking (Briggsby, 2024). Under 7 min ranks less competitively for difficult keywords.
-
-## Subscriber Conversion Benchmarks
-
-| Grade | Sub/View Ratio | Diagnosis |
-|-------|---------------|-----------|
-| A | >1% | Excellent — strong value + CTA |
-| B | 0.3–1% | Healthy |
-| C | 0.1–0.3% | Weak CTA or audience mismatch |
-| D | <0.1% | Broken — CTA missing or wrong audience |
-
-**Fix for D**: Add CTA at 25% mark (not end), use `?sub_confirmation=1` in subscribe link, check if audience matches content.
-
-## Traffic Source Health
-
-| Source | Healthy | Warning | Problem |
-|--------|---------|---------|---------|
-| YouTube Search | 30–50% | 20–30% | <20% |
-| Browse / Home | 20–35% | 10–20% | <10% |
-| Suggested Videos | 15–30% | 5–15% | <5% |
-| External | 5–15% | — | >40% |
-
-## Engagement Rate Benchmarks (Tutorial Niche)
-
-| Metric | Good | Average | Weak |
-|--------|------|---------|------|
-| Like/view ratio | >3% | 1–3% | <1% |
-| Comment/view ratio | >0.5% | 0.1–0.5% | <0.1% |
-| Returning viewer % | >30% | 15–30% | <15% |
-
-## Revenue Benchmarks (WordPress/Tutorial Niche)
-
-| Metric | Value |
-|--------|-------|
-| CPM range | $8–15 avg, $15–25 in Q4 |
-| Q4 CPM premium | +40–60% vs Q1 |
-| Plugin demo conversion | 2.8% avg, 6%+ with landing page live demo (ConversionXL, 2024) |
-
-## Sources
-- Focus Digital, December 2025
-- Briggsby YouTube SEO Study, 2024
-- ConversionXL Plugin Video Study, 2024
-- vidIQ 5M-channel dataset, 2025

package/skills/youtube/references/dataforseo-integration.md

@@ -1,51 +0,0 @@
-# DataForSEO Integration Guide
-
-## Tools Used by This Skill Suite
-
-| Tool | Used For | When |
-|------|----------|------|
-| `kw_data_google_ads_search_volume` | Monthly keyword volumes | Every `/youtube seo` and `/youtube ideate` |
-| `serp_youtube_organic_live_advanced` | Live YouTube SERP top 25 | Every `/youtube seo`, `/youtube competitor`, `/youtube ideate` |
-| `dataforseo_labs_google_keyword_ideas` | Related keyword discovery | `/youtube ideate`, `/youtube competitor` |
-| `dataforseo_labs_google_related_keywords` | Keyword variants | `/youtube seo` step 2 |
-| `serp_youtube_video_comments_live_advanced` | Comment mining | `/youtube comment-intel` |
-| `serp_youtube_video_info_live_advanced` | Video metadata lookup | Competitor analysis |
-| `on_page_content_parsing` | Crawl landing pages | `/youtube funnel` |
-
-## Critical Context: Google Volume ≠ YouTube Volume
-
-Google monthly volume in tutorial niches is deliberately LOW (10–50/mo). This is not a signal of low demand.
-
-YouTube search volume is estimated at 3–8x Google volume for tutorial keywords:
-- "elementor navigation menu" → 40/mo Google = ~160–320/mo YouTube searches
-- "off canvas menu elementor" → 20/mo Google = ~80–160/mo YouTube searches
-
-**Always pull the YouTube SERP** (`serp_youtube_organic_live_advanced`) to see actual view counts — those prove demand far better than Google volume numbers.
-
-## Reading YouTube SERP Results
-
-For each result, extract:
-- **View count**: How much traffic this topic generates
-- **Age**: How old is the top video? (>2 years = freshness gap)
-- **Channel**: Who dominates? (Are they using a widget or CSS?)
-- **Title language**: What words are driving CTR?
-
-## Keyword Priority Tiers
-
-| Tier | Volume | Use For |
-|------|--------|---------|
-| Primary | 50–500/mo Google | Title, first 40 chars |
-| Secondary | 10–50/mo Google | Description opening, secondary title |
-| Long-tail | 1–10/mo Google | Tags, WHAT YOU'LL LEARN bullets |
-| Brand | Any | Description, tags (conversion layer) |
-
-## Cost Control
-
-DataForSEO charges per API call. To minimize costs:
-- Run `kw_data_google_ads_search_volume` once per batch of keywords (not one call per keyword)
-- Run `serp_youtube_organic_live_advanced` once per primary keyword (not for every variant)
-- Cache results within the session — don't re-fetch the same keyword twice in one conversation
-
-## Sources
-- DataForSEO API documentation
-- your channel keyword research, March 2026

package/skills/youtube/references/monetization-guide.md

@@ -1,62 +0,0 @@
-# Monetization Guide
-
-## The Product-Company YouTube Revenue Model
-
-Standard YouTube advice is built for ad revenue. If you're a product company, your model is different:
-
-```
-Ad revenue model:  Views → CPM → $0.001 per view
-Product model:     Views → Free download → Pro upgrade → $49–200 LTV
-```
-
-One plugin sale = 50,000 ad views. Optimize accordingly.
-
-## Revenue Priority Order
-
-| Priority | Stream | Unit LTV | Effort to Start |
-|----------|--------|----------|----------------|
-| 1st | Plugin/product sales | $49–200 | Zero (already in description) |
-| 2nd | Email list capture | $30–120 lifetime | Low (add lead magnet) |
-| 3rd | Affiliate (hosting, tools) | $15–150/referral | Low (add links) |
-| 4th | Ad revenue (AdSense) | $0.001–0.004/view | None (passive) |
-| 5th | Sponsorships | $200–800/video | High (relationships) |
-| 6th | Course/membership | $49–99/student | Very high |
-
-## Plugin Sales Optimization
-
-The funnel:
-```
-Video → Description link (top 5 lines) → Widget landing page → Free download → Pro upgrade
-```
-
-Every step must be present and working:
-- Link in first 5 lines of description (not buried below "Show more")
-- Live demo on landing page (doubles conversion — ConversionXL, 2024)
-- Free tier available (reduces friction, converts to paid over time)
-- Discount code in description (reduces purchase hesitation)
-- CTA mentioned at 25% mark in video (not just end)
-
-## Affiliate Partnerships (No Product Conflict)
-
-Best WordPress-niche affiliates:
-- Hosting: WP Engine ($200), Cloudways ($50–100), Hostinger ($60–80) per referral
-- Page builders/tools: Rank Math Pro ($15–40), WP Rocket ($15–30)
-- Design: Envato Elements ($10–25/referral)
-
-Rules: Only promote what you actually use. One bad sponsored integration loses 3 months of trust.
-
-## CPM Seasonality
-
-| Quarter | CPM vs avg | Strategy |
-|---------|-----------|---------|
-| Q4 (Oct–Dec) | +40–60% | Push product CTAs, run sale promos |
-| Q1 (Jan) | Lowest | Volume play — upload more, push free tier |
-| Q2–Q3 | Baseline | Standard mix |
-
-Black Friday week = highest plugin purchase intent in the year. Plan a sale announcement video.
-
-## Sources
-- ConversionXL Plugin Demo Conversion Study, 2024
-- Demand Curve SaaS YouTube Study, 2025
-- Mailchimp Email Benchmark Report, 2025
-- YouTube CPM seasonal data, creator community analysis

package/skills/youtube/references/repurposing-guide.md

@@ -1,54 +0,0 @@
-# Repurposing Guide
-
-## Why Repurpose (The Math)
-
-One 10-minute tutorial = 7 platform assets with ~3 hours of additional reach.
-
-HubSpot, 2024: Content repurposed to 3+ formats gets 3.5x total reach vs. single-channel publishing.
-
-| Asset | Platform | Time to create | Estimated reach |
-|-------|----------|---------------|----------------|
-| Short clip (30–60s) | YouTube Shorts | 30 min edit | 2–10x video views |
-| Companion blog post | WordPress | 45 min | Google organic traffic |
-| LinkedIn post | LinkedIn | 15 min | 500–5,000 professionals |
-| X thread | X (Twitter) | 20 min | 200–2,000 developers |
-| Email newsletter | Email list | 20 min | 35–45% open rate |
-| Community post | YouTube | 5 min | Browse impression boost |
-
-## Platform-Specific Rules
-
-### LinkedIn
-- Native text + image posts reach 5x more than link posts (LinkedIn 2025 algorithm)
-- Never put the YouTube URL in the post body — put it in first comment after posting
-- Image: screenshot of finished UI (dark background works best in feed)
-- 3–4 short paragraphs, no bullet walls
-
-### X (Twitter)
-- Thread format: 6–7 tweets
-- Tweet 1: surprising result or stat ("This nav menu took 6 minutes, zero CSS")
-- Tweets 2–5: one step each (max 240 chars, no jargon)
-- Tweet 6: before/after image
-- Tweet 7: YouTube link + product link
-
-### Email
-- Subject: `[Benefit] in [Platform] — [Differentiator]`
-- Body: under 200 words, link to video, 3 bullet summary
-- Tutorial newsletters benchmark: 35–45% open rate vs 22% industry avg (Mailchimp 2025)
-
-### Shorts (from Long-form)
-- Post within 7 days of long-form upload (freshness decay cliff at 28–30 days)
-- Best clip: the result reveal or most surprising feature moment
-- Add text overlays — 80% watch muted
-
-### WordPress Blog Post
-- BrightEdge, 2025: blog posts with YouTube embed rank 53% higher on Google
-- Always embed the video above the fold
-- Add VideoObject JSON-LD schema (see `sub-skills/wp-post.md` for full template)
-- Target 800–1,200 words — enough for Google to rank, short enough to complete
-
-## Sources
-- HubSpot Content Distribution Report, 2024
-- LinkedIn Algorithm Update, 2025
-- Mailchimp Email Benchmark Report, 2025
-- BrightEdge Search Intelligence Report, 2025
-- YouTube Official Shorts update, September 2025

package/skills/youtube/references/retention-scripting-guide.md

@@ -1,75 +0,0 @@
-# Retention Scripting Guide
-
-## The Hook Window (0:00–0:42)
-
-The first 42 seconds are the highest drop-off window. Structure them as 3 beats:
-
-```
-GRAB (0:00–0:10)   — Show the result or state the surprising thing
-PROMISE (0:10–0:28) — What they'll be able to do after this video
-STAKES (0:28–0:42) — Why this matters / what problem it solves
-```
-
-**Optimal hook length**: 28–42 seconds (Retention Rabbit, 150M+ minutes analysis)
-- Under 28s: Not enough time to establish why they should stay
-- Over 42s: People leave before getting to the content
-
-**Never start with**: "Hey guys, welcome back to the channel, don't forget to like and subscribe..."
-This kills 15–25% of viewers before you start.
-
-## Pattern Interrupts (Every 60–90s)
-
-After the hook, viewers drop off progressively. Pattern interrupts reset attention:
-
-| Type | How to Use |
-|------|-----------|
-| Visual cut | Switch from talking head to screen recording |
-| Text overlay | Pop a key stat or result on screen |
-| B-roll | Show the finished result mid-tutorial |
-| Pace change | Speed up a repetitive step with text overlay |
-| Question | "Now here's where most people get stuck..." |
-
-Wistia (2024): Pattern interrupts every 60–90s reduce drop-off by 40%.
-
-## The 25% Mark CTA
-
-Place a soft subscribe/product CTA at 25% through the video — not the end.
-- Most viewers who leave early do so by 50% — waiting until the end loses them
-- "If this is useful, grab the free plugin below — link in description" is enough
-- Don't hard sell — viewers drop if they feel sold to mid-tutorial
-
-## The Tutorial Body
-
-For step-by-step tutorials, every step needs:
-1. **What**: Name the step
-2. **Why**: One sentence on why this setting matters
-3. **How**: Screen recording with audio — the actual action
-4. **Confirm**: Show the result before moving to next step
-
-Skip "Why" for obvious steps — only include it when the viewer would otherwise wonder.
-
-## Retention Risk Points
-
-| Timestamp | Risk | Fix |
-|-----------|------|-----|
-| 0:00–0:10 | No hook | Show result in first 3 seconds |
-| 0:42–1:30 | Slow intro | Cut to content by 45s |
-| Mid-video (setup steps) | Repetitive | Speed up + text overlay |
-| Near end (outro) | CTA overload | One CTA max, 15 seconds |
-
-## Energy Patterns (AIR Media-Tech)
-
-5 energy states to vary throughout a tutorial:
-1. **Demonstration** — show the tool in action
-2. **Explanation** — narrate what's happening
-3. **Problem/Solution** — state the issue, reveal the fix
-4. **Result reveal** — show finished output
-5. **Context** — why this matters for their use case
-
-Rotate these to prevent monotony. Never 3 consecutive steps of pure explanation.
-
-## Sources
-- Retention Rabbit, 150M+ minutes of tutorial watch data
-- Wistia Video Benchmark Report, 2024
-- Adobe Tutorials Study (completion rate analysis)
-- AIR Media-Tech YouTube coaching framework

package/skills/youtube/references/seo-playbook.md

@@ -1,77 +0,0 @@
-# YouTube SEO Playbook
-
-## Title Formula
-
-```
-[Primary keyword — first 40 chars] | [Differentiator] | [Brand/Year]
-```
-
-**Non-negotiables:**
-- Primary keyword in first 40 characters (highest indexing weight)
-- Under 100 characters total (YouTube hard limit)
-- At least one differentiator word: FREE / Without [X] / No Code / Smooth / Animated
-- Year (2026) in at least 1 of your 3 title variants
-- Never start with "How to" alone — too generic, too low CTR
-
-**High-CTR words proven in tutorial niche (SERP data, 2026):**
-- INCREDIBLE / SMOOTH / BEAUTIFUL / ANIMATED → emotional
-- WITHOUT [thing they hate] / NO CODE / NO CSS / FREE → pain-point relief
-- COMPLETE / FULL TUTORIAL / STEP BY STEP → thoroughness promise
-- [Year] → freshness signal
-- Portfolio / Agency / Beginner → audience callout
-
-## Description Architecture
-
-**First 150 chars** (shown before "Show more") = highest keyword weight:
-- Must contain primary keyword naturally
-- Must name the product/tool
-- Do not start with widget URL — put URL on line 1, start text immediately after
-
-**"WHAT YOU'LL LEARN" section** = critical for indexing:
-- Each bullet is an additional keyword signal
-- Write bullets as "How to [action] + [outcome]" — both searchable and motivating
-- 5–7 bullets max
-
-**Links section**: Fix it once, use it in every video. Inconsistent links = missed traffic.
-
-## Tag Strategy (20 tags, ordered by priority)
-
-1. Exact primary keyword
-2. Exact secondary keyword
-3. Product + feature (e.g., "your-product hover menu")
-4. Broader category (e.g., "elementor navigation tutorial")
-5. Use case (e.g., "portfolio navigation wordpress")
-6. Competitor method (e.g., "off canvas menu elementor") — captures comparison searches
-7. Product brand (e.g., "[your-product]", "[your product brand]")
-8. Year variant
-9. Platform broad (e.g., "elementor", "elementor pro")
-10. WordPress general
-
-**Hard limits**: No tag over 30 chars. Total under 500 chars. No keyword stuffed (same word 8+ times).
-
-## SERP Gap Analysis (Run Before Every Video)
-
-Search the primary keyword with `serp_youtube_organic_live_advanced`. Look for:
-
-| Signal | What to Do |
-|--------|-----------|
-| Top video >2 years old | Freshness gap — make this video now |
-| Top video requires CSS/code | No-code gap — your widget version wins |
-| Top video requires Pro plugin | Free tier gap — target with "without Pro" framing |
-| No video covers full use case | Completeness gap — go comprehensive |
-| Top video <100K views | Moderate opportunity |
-| Top video >100K views | Proven demand — strong opportunity |
-
-## Chapters (Timestamps)
-
-Chapters are a ranking signal — they create additional keyword-indexed anchors in your video.
-- Minimum 3 timestamps
-- Must start at `0:00`
-- Each chapter title = additional keyword opportunity
-- Format: `0:00 Introduction` (no brackets required)
-
-## Sources
-- YouTube Official Creator Academy
-- Focus Digital, December 2025
-- vidIQ keyword research data, 2025
-- DataForSEO SERP analysis, March 2026

package/skills/youtube/references/shorts-playbook.md

@@ -1,64 +0,0 @@
-# Shorts Playbook
-
-## Shorts vs Long-form: Critical Differences
-
-| Element | Long-form | Shorts |
-|---------|-----------|--------|
-| Hook window | 28–42s | 1–2s |
-| Format | 16:9 horizontal | 9:16 vertical |
-| Optimal length | 8–15 min | 30–60s |
-| CTA window | 25% mark + end | Last 5s only |
-| Description shown | 150 chars (preview) | 150 chars |
-| Clickable links | Yes | No (under 1K subs) |
-| End cards | Yes | No |
-| Tags | 20 tags useful | 5–8 tags |
-
-## The Freshness Decay Cliff
-
-**YouTube Official, September 2025**: Shorts have a 28–30 day distribution cliff.
-- Post within 7 days of long-form upload to catch algorithm momentum
-- After 30 days, a Short's reach drops ~60% regardless of engagement
-- Plan: one Short per long-form video, posted within 7 days
-
-## Hook in 2 Seconds (Not 3)
-
-Shorts scroll behavior is faster than long-form. The swipe decision happens in 1–2 seconds.
-- Show the result in the first frame — before you say anything
-- Text overlay in first frame: "No CSS needed" / "60 seconds" / "This changed everything"
-- Do NOT start with your face talking — show the output
-
-## Loop Rate = the Key Metric
-
-Unlike long-form (watch %), Shorts optimize for **loop rate** (how many times the Short replays).
-- End the Short in a way that makes replaying natural
-- Open loop endings: "But there's one more thing..." then cut
-- Cliffhanger: Show half the result, cut at the most interesting moment
-- Ambiguous ending → viewers rewatch to catch what they missed
-
-## Silent Viewing (80% Watch Muted)
-
-Every key step needs a text overlay. Rule: if you removed all audio, could someone follow along?
-- Max 5 text overlays per Short
-- Under 5 words per overlay
-- High contrast — white text on dark, or brand accent
-
-## Shorts Topic Sourcing (Priority Order)
-
-1. **Clip from long-form** — extract the result reveal or most surprising moment
-2. **Standalone answer** — "The one Elementor setting most people miss"
-3. **Before/after** — 15s before, 5s transition, 15s after
-4. **Quick win** — something useful in under 60s that doesn't need a full tutorial
-5. **Trend capture** — a trending topic that fits your niche (bento grid, dark mode, etc.)
-
-## Post-Reform View Counting (March 2025)
-
-YouTube now counts a Shorts view after replay, not just first watch. This means:
-- Loop rate directly affects view count
-- A Short with 5 loops from 100 viewers = 500 views
-- Optimize for replay, not just initial watch
-
-## Sources
-- YouTube Official Shorts update, September 2025
-- YouTube Official view counting reform, March 2025
-- Adobe Tutorials Study (completion rate)
-- vidIQ Shorts analysis, 2025