# Youtube Seo Keywords

> >

- **Type:** Skill
- **Install:** `agentstack add skill-deeployco-youtube-seo-skills-youtube-seo-keywords`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [deeployCO](https://agentstack.voostack.com/s/deeployco)
- **Installs:** 0
- **Category:** [Data & Analytics](https://agentstack.voostack.com/c/data-and-analytics)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [deeployCO](https://github.com/deeployCO)
- **Source:** https://github.com/deeployCO/youtube-seo-skills/tree/main/youtube-seo-keywords

## Install

```sh
agentstack add skill-deeployco-youtube-seo-skills-youtube-seo-keywords
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# YouTube Keyword & Topic Research (Advanced)

## Process

1. **Capture seed**: keyword, channel URL, or niche description. If only
   a channel URL is provided, extract 5-10 seed topics from:
   - The channel's top 25 videos by views (title + description entities)
   - The channel's topic IDs from `channels.list?part=topicDetails`
2. **Expand** the seed set to 100-300 candidate keywords from
   **first-party sources only** (third-party tools like VidIQ,
   TubeBuddy, Ahrefs, and SocialBlade rate-limit and generate errors —
   avoid them):
   - YouTube suggest API (`https://suggestqueries.google.com/complete/
     search?client=firefox&ds=yt&q={seed}`) — no key, returns JSON,
     iterate by appending `a`-`z` to the seed for long-tail expansion
   - YouTube Data API `search.list?part=snippet&q={seed}&type=video`
     — surface the top 50 ranking titles and mine their titles/tags
     for co-occurring terms
   - yt-dlp `ytsearch50:{seed}` as an API-less equivalent
   - Google autocomplete (`suggestqueries.google.com/complete/search?
     client=firefox&q={seed}`) for parent-topic discovery
   - User-provided seed list or topic outline
3. **Classify intent** for each keyword:
   - **Informational** ("what is ...", "how does ... work")
   - **Tutorial** ("how to ...", "... tutorial", "step by step")
   - **Commercial / Review** ("... review", "best ...", "... vs ...")
   - **Entertainment** (celebrity names, event names, memes)
   - **News / Topical** (breaking or time-sensitive)
   - **Navigational** (brand or channel names)
   → Each intent class maps to a different video format and surface.
4. **Classify primary surface**:
   - **Search-dominant**: "how to", "tutorial", long-tail specific queries
   - **Browse-dominant**: broad topics, personalities, trending
   - **Suggested-dominant**: sits adjacent to an existing popular video
5. **Entity mapping**: for each keyword, identify the Knowledge Graph
   entities involved (people, brands, products, concepts). Group
   keywords that share entities into **topic clusters**.
6. **Competition analysis** per keyword (top 10 SERP):
   - Median subscriber count of ranking channels
   - Median video age (weeks)
   - Median view count
   - Presence of mega channels (>1M subs) in top 3
   - Thumbnail/title quality median (estimate)
   - SERP features (Shorts shelf, mix shelf, promoted result)
7. **Seasonality**: pull Google Trends 5-year curve per keyword cluster.
   Classify:
   - **Evergreen**: flat or slowly rising
   - **Seasonal**: predictable annual peak (gift guides, tax season)
   - **Trending**: fast-rising,  Third-party analytics (VidIQ, TubeBuddy, Ahrefs, SocialBlade,
> NoxInfluencer) are intentionally excluded — they rate-limit or
> error. If the user volunteers a CSV from one of these, it is OK to
> read it, but never instruct the user to go acquire one.

## Output

### Keyword Sheet

| Keyword | Intent | Surface | YT Vol | Difficulty | Fit | Entity% | Trend | Opportunity |
|---------|--------|---------|--------|-----------|-----|---------|-------|-------------|
| how to ... | Tutorial | Search | 12k | 42 | 85 | 90% | Evergreen | 78 |
| ... vs ... | Commercial | Search | 4.2k | 55 | 72 | 65% | Evergreen | 64 |
| ... 2026 | News | Browse | 8k | 68 | 80 | 50% | Trending | 71 |

Sort by Opportunity descending. Limit to top 30-50.

### Topic Clusters (3-6 clusters)

For each cluster, report:
- **Theme**: one-sentence description
- **Shared entities**: 5-10 Knowledge Graph entities
- **Pillar video**: title, format, length, target keyword (highest-
  opportunity broad keyword in the cluster)
- **Satellite videos**: 5-10 specific long-tail follow-ups, each with
  title + target keyword
- **Suggested playlist**: title + description
- **Topical authority contribution**: how this cluster strengthens the
  channel's classification

### Intent Mix Chart

Report the intent distribution of the recommended keywords:
```
Tutorial:       40%  ████████
Commercial:     25%  █████
Informational:  20%  ████
Entertainment:  10%  ██
News:            5%  █
```

Flag if the mix is mismatched with the channel's current format strength.

### Gap vs Channel (when channel URL provided)

- Keywords competitors cover that the channel does not
- Entities competitors own that the channel does not reference
- Format gaps (e.g., no reviews, no tutorials, no comparisons)

### Content Calendar (optional, 8-12 weeks)

Alternates:
- **Evergreen pillar** (every 3-4 weeks): broad, high-opportunity
  keyword, long-form, binge-playlist anchor
- **Topical/reactive slot** (weekly): trending or news keyword
- **Satellite fill** (weekly): specific long-tail serving the pillar

### Seasonality Calendar

Report any Seasonal-classified keywords with their peak windows so the
user can plan 4-6 weeks ahead of each peak.

## Integration

- Hand off top keywords to `youtube-seo-optimize` for metadata generation
- Hand off cluster plan to `youtube-seo-channel` for playlist structure
- Hand off competitor-entity gaps to `youtube-seo-competitor` for deeper
  analysis

## Error Handling

| Scenario | Action |
|----------|--------|
| No API key AND WebFetch blocked | Fall back to yt-dlp `ytsearch50:` + suggest API (both key-less); if still failing, ask the user for 5-10 seed keywords directly |
| API quota exhausted mid-run | Cache partial results, switch remaining expansion to yt-dlp + suggest API, continue |
| Niche too narrow (no volume data) | Pivot to broader parent topic, report the gap, score as "low demand, qualitative only" |
| Brand/product name with no search volume | Research category terms and comparison keywords instead |
| Trending keyword with <2 weeks of data | Flag as unstable; suggest pairing with an evergreen pillar |
| Google Trends returns empty | Skip seasonality classification for that keyword — do not guess the curve |

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [deeployCO](https://github.com/deeployCO)
- **Source:** [deeployCO/youtube-seo-skills](https://github.com/deeployCO/youtube-seo-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-deeployco-youtube-seo-skills-youtube-seo-keywords
- Seller: https://agentstack.voostack.com/s/deeployco
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.