geo-youtube
YouTube citation playbook for AI search. YouTube is the most-cited domain in Google AI Overviews and the strongest single brand-visibility signal measured. Optimizes videos for AI citation through transcripts, chapters, titles, and cross-posted text claims -- not views or subscribers. Use when planning video content for AI visibility or auditing an existing channel for citation readiness.
TheSmokeDev · 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 TheSmokeDev/geo-skills --skill geo-youtube技能说明
name: geo-youtube description: YouTube citation playbook for AI search. YouTube is the most-cited domain in Google AI Overviews and the strongest single brand-visibility signal measured. Optimizes videos for AI citation through transcripts, chapters, titles, and cross-posted text claims -- not views or subscribers. Use when planning video content for AI visibility or auditing an existing channel for citation readiness. allowed-tools:
- Read
- Grep
- Glob
- Bash
- WebFetch
- Write
YouTube Citation Playbook Skill
Purpose
This skill turns a YouTube channel into an AI citation source. AI engines do not watch video -- they read the transcript, chapters, title, and description. That makes YouTube a rank-free bypass lane: a channel with no domain authority and few views can earn AI citations that the site's text pages cannot. This skill audits existing videos for citation readiness and produces a build spec for new ones.
Core Insight
YouTube completed the flip in 2026 and is now the dominant non-text citation source:
- Most-cited domain in Google AIOs overall, with share up +34% over 6 months (Ahrefs Brand Radar, 2026).
- #1 social citation source at 38.1% of social citations, overtaking Reddit (5WPR State of AI Citations, May 2026).
- ~23% of finance-vertical citations -- ahead of Wikipedia (7.3%) and LinkedIn (6.8%) (Surfer 46M-citation dataset, 2026).
- 18.2% of AIO citations from beyond the organic top 100 are YouTube URLs (Ahrefs, 863K SERPs / 4M URLs, Mar 2026) -- the clearest rank-free lane in AI search.
- YouTube mentions correlate r=0.737 with AI visibility -- the strongest single signal ever measured, ahead of branded web mentions (0.664) (Ahrefs 75K-brand study via MachineRelations, Jul 2026).
The mechanism that matters: AI reads the transcript, chapters, title, and description -- NOT the video itself. Views and subscriber counts show near-zero correlation with citation (⚠️ single source, AIOCopilot Apr 2026 -- but mechanism-consistent, since engines have no visibility into watch metrics at retrieval time).
Citation-Readiness Rubric (per video)
Score each video 0-100 across five components:
| Component | Weight | Full Marks | Zero Marks |
|---|---|---|---|
| Transcript quality | 30% | Human-corrected transcript with citable numeric/definitional sentences | Auto-captions left uncorrected, or no captions |
| Chapter structure | 25% | 6-12 chapters with descriptive, query-matching titles | No chapters, or generic titles ("Part 1", "Intro") |
| Title match | 20% | Title is verbatim (or near-verbatim) the target query | Clever/branded title with no query language |
| Length & depth | 15% | 10+ minutes, substantive per-chapter content | Under 3 minutes, or a Short |
| Cross-posting | 10% | Video's claims exist as a text page + community answer | Video exists in isolation |
Shorts are never cited -- exclude them from the audit entirely (AIOCopilot, Apr 2026).
Audit Procedure (Existing Channel)
- List the channel's long-form videos (10+ min candidates first).
- For each video, fetch title, description, chapter list, and transcript (YouTube timedtext API or transcript endpoints).
- Check the transcript for:
- Auto-caption errors (numbers, brand names, and technical terms mangled) -- flag for human correction.
- Citable sentences: numeric claims and definitional statements, especially within ~30 seconds of chapter boundaries (citations attach to timestamps).
- Check chapter titles: each should read like an H2 that matches a fan-out sub-query (see
skills/geo-fanout/for sub-query mapping). "How much does SR-22 cost in California?" beats "Pricing section". - Check title against the target query: verbatim or near-verbatim match is the goal.
- Check cross-posting: do the video's core claims exist as a text page on the site and as a community answer (Reddit/Quora/forum)? ChatGPT cites YouTube mainly when the video is discussed in text elsewhere (AIOCopilot, Apr 2026).
- Score per the rubric; produce the fix list ordered by transcript quality first.
Build Spec (New Videos)
For each target query, spec a video as follows:
- Title: verbatim the target query. "How Much Does SR-22 Insurance Cost in California? (2026)" -- not "Everything You Need to Know About SR-22!".
- Length: 10+ minutes. Shorts never cited; long-form gives chapters room to work.
- Chapters: 6-12, each titled like an H2 matching a fan-out sub-query (cost, eligibility, process, location, language variants -- see
skills/geo-fanout/). Citations attach to timestamps, so chapter titles are the retrieval surface. - Script for citable sentences: place numeric and definitional claims near chapter boundaries. "The average SR-22 filing fee in California is $25, and SR-22 insurance raises premiums by 40-80% depending on the violation" is citable; "it can get pretty pricey" is not. Apply the
skills/geo-citability/passage rules to the script. - Transcript: upload a human-corrected transcript. Do not rely on auto-captions -- mangled numbers destroy citability.
- Description: front-load the same numeric/definitional claims; engines read it.
- Production: a faceless explainer with a corrected transcript qualifies -- production value is not the ranking lever here (AIOCopilot, Apr 2026).
- Cross-post (mandatory): publish the video's claims as (a) a text page on the site and (b) an authentic community answer (Reddit, Quora, niche forum) linking or naming the video. Text discussion is what makes ChatGPT surface the video.
Caveat: Treat YouTube Citation Share as a Live Experiment
YMYL video-citation share swung 50+ percentage points in 3 months in BrightEdge's healthcare tracking (Jan 2026). Platforms are actively A/B testing how heavily they cite video. Do not promise a stable citation share to clients; track the channel's actual citation rate over time with skills/geo-measurement/ a