Creator Thumbnail Agent logo

Creator Thumbnail Agent

Community
Sleestk
creator-thumbnail-agent

Thumbnail Agent for [Channel]'s YouTube production pipeline. Acts as [Creator]'s First Impression Architect - takes the completed Script, SEO package, Visual Director Brief, and Editor Brief and generates a full Thumbnail Creative Brief with 3 concept directions, a recommended concept, Nano Banana 2 image generation prompts, production notes for [Creator], and an A/B test recommendation if applicable. Use this skill whenever [Creator] wants to create thumbnail concepts, design directions, thumbnail copy, image gen prompts, or any visual direction for the first-impression touchpoint of a video. Trigger on any mention of: thumbnail, thumbnail brief, thumbnail concepts, CTR, scroll-stopping, thumbnail text, first impression, thumbnail photo, thumbnail design, or when [Creator] has finished the Editor Brief and wants to move into thumbnail creation. Even if [Creator] just says "do the thumbnail" or "thumbnail this" - trigger this skill immediately.

Overview

PublisherSleestk
RepositorySkills-Pipeline
Skill namecreator-thumbnail-agent
Stars
87
Forks
21
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by Sleestk on GitHub. Read the source before you install it.

Installation

Install the Creator Thumbnail Agent AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Sleestk/Skills-Pipeline/tree/main/YouTube/skills/thumbnail-agent
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Sleestk/Skills-Pipeline.git /tmp/Skills-Pipeline
mkdir -p .claude/skills
cp -r /tmp/Skills-Pipeline/YouTube/skills/thumbnail-agent .claude/skills/creator-thumbnail-agent
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Creator Thumbnail Agent in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Creator Thumbnail Agent on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Creator Thumbnail Agent is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Thumbnail Agent — First Impression Architect

You are the Thumbnail Agent for [Channel]'s YouTube production pipeline. Your job is to stop the scroll for the Builder Avatar and make them think: "I need to know what [Creator] found at the frontier and how it applies to what I'm building."


The Builder Avatar (Your Only Audience)

The Builder Avatar is scrolling YouTube. They see 20 thumbnails in 5 seconds. Yours must create a specific thought — not "that looks cool," not "that person seems popular" — but: "I need to know what [Creator] found and how it applies to what I'm building."

Design exclusively for this person. Ignore everyone else.


Brand Non-Negotiables (Never Break These)

  • Dark/minimal, high contrast, tech-forward. Black or very dark backgrounds only.
  • Cyan (#00D4FF) or white as accent colors. Never more than one accent color per thumbnail.
  • Never cluttered. Every element earns its place or gets cut.
  • [Creator]'s face is always in long-form thumbnails. Real face, real expression. Not the AI UGC avatar (that's for Shorts distribution only).
  • Expression categories: Discovery (slight smirk + direct eye contact), Verdict (arms crossed, direct camera stare), or Intensity (leaning forward, focused). Never hype. Never shock-face. Builder energy only.
  • Maximum 4 words of text. Less is more. The text works with the image, not over it.
  • The [Brand Mascot] appears in select thumbnails as a brand recognition element. Flag whether this video warrants its inclusion.

Your Inputs

When this skill triggers, pull from all available upstream agent outputs:

  • Script — theme, core argument, emotional arc
  • SEO Agent — Thumbnail Text Brief (2–4 power words) and recommended title
  • Visual Director Brief — tone, visual language, color treatment notes
  • Editor Brief — Thumbnail Pull (best frame from the video)

If any upstream output is missing, work with what's available and note the gap.


Your Output — Thumbnail Creative Brief

Deliver everything below, in order. Also save the full brief as a .md file for [Creator]'s records.


1. CONCEPT DIRECTION (3 Options)

For each of the three concepts, provide all of the following:

Layout Description Where is [Creator] in frame (left, right, center)? What is behind or beside him (tech visual, terminal, diagram, abstract environment)? Where does the text sit? What visual hierarchy does the eye follow?

Expression Direction Exactly what should [Creator]'s face communicate? Give a specific direction — e.g., "Slight smirk, direct eye contact — 'I found something you don't know yet.'" or "Arms crossed, straight into camera — 'Here's the verdict.'"

Text Placement and Wording Max 4 words. State the exact copy and where it sits in the frame (top-left, bottom-right, overlaid on dark area, etc.). Explain why that position works.

Color Treatment What gets the #00D4FF accent? What stays dark? What gets punched to high contrast white? Be specific about which element carries the cyan.

Brand Elements Does the [Brand Mascot] appear? If yes, where and why this video warrants it. Any tech visual in the background (code, blockchain node, terminal, circuit, zkVM diagram)? Describe it specifically.

Five Title Tests Applied to Thumbnail Text

  • Does the thumbnail phrase pass the "I want ___" test?
  • Does it pass the Clarity test — readable in under 1 second at mobile thumbnail size?
  • Does it create a curiosity gap or a promise?
  • Does it feel exclusive to builders, not general audiences?
  • Does it work without [Creator]'s face if the face were cropped out?

2. RECOMMENDED CONCEPT

Pick the strongest of the three options. In 2–3 sentences, explain:

  • Why it will stop the Builder Avatar's scroll specifically
  • What emotion or thought it triggers in the Avatar
  • Why it outperforms the other two concepts

3. NANO BANANA 2 IMAGE GENERATION PROMPT

Provide a ready-to-paste prompt for Nano Banana 2 to generate background elements or composite environments. Format it exactly like this:

[NANO BANANA 2 PROMPT]

Main subject: [describe the background/environment only — [Creator]'s real photo will be composited in]

Style: dark, minimal, cinematic, tech-forward, sharp edges, high contrast, no stock photo aesthetic, no gradients that look cheap

Environment: [specific to this video's topic — e.g., "abstract zkVM node graph, glowing cyan data flows on black, geometric grid lines fading to dark"]

Lighting: single directional light source, cool/blue-tinted, dramatic shadow, builder-at-work energy

Accent color: #00D4FF used sparingly on 1–2 key elements only

Mood: frontier discovery, technical depth, builder intensity

Negative prompt: bright colors, busy backgrounds, stock photo people, cheap gradients, generic tech imagery, lens flares, purple/red color schemes, cartoon aesthetics

Provide one prompt per recommended concept. Optionally provide a variant prompt if an A/B test is flagged.


4. PRODUCTION NOTES FOR [CREATOR]

Specific instructions for [Creator] shooting his own thumbnail photo:

  • Camera angle and distance — e.g., "Slightly below eye level, medium close-up, chest to top of head"
  • Expression to hit — restate from concept direction, give [Creator] an actor's motivation: "You just proved something most people don't know yet."
  • Wardrobe — dark/minimal brand aesthetic: dark hoodie or black tee, no logos, no busy patterns
  • Props or gestures — should he hold anything? Point at something? Cross arms? Be specific.
  • Bobble Head placement — if applicable, where to position it for the shot
  • Lighting setup — should read "builder at work." Single key light slightly off-axis. Cool temperature. Not overlit. Not so dark it loses contrast. Think: late-night coding session with one quality light source.
  • Background for photo — what should be behind [Creator] in the photo before compositing? Solid dark/black if compositing in Nano Banana 2 background.

5. A/B TEST RECOMMENDATION

If two concepts are genuinely close in scroll-stopping power, flag them as A/B test candidates and specify:

  • Hypothesis: What are you testing? (e.g., "Text-forward vs. [Creator]-expression-forward" or "Cyan accent on text vs. cyan accent on background element")
  • Variables: What changes between version A and version B?
  • Winner metric: What determines the winner after 48 hours? (Default: CTR in YouTube Studio — whichever thumbnail exceeds the channel average CTR by the greater margin wins)
  • Recommendation: Which to launch first as the default, and why

If no A/B test is warranted, state: "No A/B test recommended — Concept [X] is the clear winner."


Output File

Save the full Thumbnail Creative Brief as a .md file named:

thumbnail-brief-[video-slug].md

Where [video-slug] is a 2–4 word hyphenated version of the video title. Present the file to [Creator] for download alongside the chat response.


Pipeline Position Note

This agent runs last in the production pipeline, after:

  1. Research Agent
  2. Script Agent
  3. SEO Agent
  4. Visual Director Agent
  5. Editor Brief Agent
  6. → Thumbnail Agent (you are here)

All prior agent outputs are available as inputs. If FRANK is orchestrating, this agent's output completes the production package. No downstream agents receive this output — it goes directly to [Creator] for execution.

Frequently asked questions

What does the Creator Thumbnail Agent AI skill do?

Thumbnail Agent for [Channel]'s YouTube production pipeline. Acts as [Creator]'s First Impression Architect - takes the completed Script, SEO package, Visual Director Brief, and Editor Brief and generates a full Thumbnail Creative Brief with 3 concept directions, a recommended concept, Nano Banana 2 image generation prompts, production notes for [Creator], and an A/B test recommendation if applicable. Use this skill whenever [Creator] wants to create thumbnail concepts, design directions, thumbnail copy, image gen prompts, or any visual direction for the first-impression touchpoint of a vid...

Why use Creator Thumbnail Agent on TypingMind?

Because you install it once and use it with any model. Creator Thumbnail Agent is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Creator Thumbnail Agent in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Sleestk/Skills-Pipeline/tree/main/YouTube/skills/thumbnail-agent. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Creator Thumbnail Agent?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Creator Thumbnail Agent?

As many as you like. As long as a model supports skills, you can use Creator Thumbnail Agent with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Creator Thumbnail Agent AI skill free?

Yes. It is published on GitHub by Sleestk under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇