Mindket

Charlie Hills Social Media Skills

Newversion 2 · updated
Star4593

Overview

Charlie Hills’s 17 free skills for Claude and Codex: voice profile, LinkedIn posts, hooks, Reel scripts, YouTube thumbnails, newsletters and analytics. A short interview comes first, and the other skills then write drafts in your voice. You post them yourself.

Charlie Hills Social Media Skills is an open-source project on GitHub in the Marketing & SEO category, listed by Mindket. It is free. This is version 2, last updated on . Install it from GitHub as its README says. Get Mindket’s setup notes on your Mindket account, as a zip.

At a glance

Runs in
  • Claude Code: a plugin from the maker’s marketplace, two commands
  • Codex: the skill-installer, or a project-local copy under .agents/skills
  • Claude Desktop: one skill folder zipped and uploaded through Customise skills
  • Gemini: only for the image prompts, which the skills write and Gemini turns into pictures
Start with
/plugin marketplace add charlie947/social-media-skills
/plugin install social-media-skills

Platform

CodexCoding agent
ClaudeAI chat
Claude CodeCoding agent
Claude DesktopDesktop app
GeminiAI chat

About

What it is

Charlie Hills Social Media Skills is a free pack of 17 skills for Claude and Codex. They cover LinkedIn posts, Instagram Reels, YouTube thumbnails, newsletters and analytics. Everything starts from a voice profile. The skill voice-builder interviews you, reads 3 to 5 pieces of your writing and saves about-me.md and voice.md in your project. The skills that write in your voice read those two files first.

  1. You answer voice-builder’s questions and paste samples
  2. Your agent saves about-me.md and voice.md in your project
  3. You name a job, such as a post, hooks or a Reel script
  4. The matching skill reads your two files and drafts
  5. You review the draft and post it yourself

Who it suits

Creators, founders and marketers who publish on LinkedIn, Instagram, YouTube or a newsletter and want drafts that sound like them. Most of the skills are written for LinkedIn. Only the Reel link route needs an Apify token and a Google AI key. The other skills run on what you paste, apart from niche research, which needs web search or a browser.

Who uses it, and how
WhoSkills they reach forExample
LinkedIn creatorspost-writer, hook-generator, post-scorerA creator pastes meeting notes, gets a post in their own voice, then scores it against an export of their past posts.
Newsletter writersnewsletter-voice, content-matrix, reels-scriptingA writer pairs one newsletter section with a reference Reel and gets a script of 30 to 45 seconds with a matching caption.
Video creatorsyoutube-thumbnail, pinned-commentA YouTuber gives a working title and a headshot, approves a short brief, then gets a Gemini prompt for the thumbnail.
Marketers and analystsanalytics-dashboard, niche-researchSomeone uploads a LinkedIn Analytics export and gets a dashboard with five data-backed recommendations.

Where to use it

Codex and Claude. The README gives an install route for each.

  • Claude Code: a plugin from the maker’s marketplace, two commands
  • Codex: the skill-installer, or a project-local copy under .agents/skills
  • Claude Desktop: one skill folder zipped and uploaded through Customise skills
  • Gemini: only for the image prompts, which the skills write and Gemini turns into pictures

How to set it up

Get it saves a zip of Mindket’s setup notes in your downloads folder, with every route from Charlie Hills’s README.

What you need first
NeedDetail
Claude Code, Codex or Claude DesktopOne of the three routes the README documents
GitFor the Codex copy route and for updating a copy
Your own writing3 to 5 samples for voice-builder. It can load a starter set instead, which borrows another author’s style
Apify token and Google AI keyOnly for reels-scripting when you give it a Reel link. post-scorer can use Apify too, and it is optional
Gemini accessOnly to make pictures from the image prompts. The prompts need no integration
  1. Add the marketplace in Claude Code. You type the first line inside a session.
  2. Install the plugin. You type the second line, and Claude Code shows its details first.
  3. Run voice-builder first. Say “Build my voice” and your agent starts the interview.
/plugin marketplace add charlie947/social-media-skills
/plugin install social-media-skills

In Codex, you ask for the install in words. The README’s example installs two skills, and any of the 17 folders works the same way.

Use skill-installer to install voice-builder and post-writer from charlie947/social-media-skills, paths skills/voice-builder and skills/post-writer.

In Claude Desktop, you zip one skill folder, references included, and upload it through Customise skills. Run this from the skills folder of a clone. Windows has no zip command of its own, so there you need another zip tool.

zip -r voice-builder.skill voice-builder

How to use it

Ask for a job in plain words. Your agent picks the skill that matches. Most of the first column is the README’s own examples, and the rest are trigger phrases from each SKILL.md.

What you say and what comes back
You saySkillWhat comes back
Build my voicevoice-builderabout-me.md and voice.md in your project, after an interview and 3 to 5 samples
Write me a post about AI agentspost-writerAngles and frameworks to pick from, then a draft of 150 to 300 words in a code block
Write me hooks for this topichook-generatorSix two-line hooks, each line at most 40 characters
Score this draft against my historypost-scorerA scorecard out of 50, or out of 40 when it has neither your voice profile nor your own posts
Make me a carousel from thisgemini-carouselA slide brief for you to approve, then one Gemini prompt per slide
What should I post this weekniche-researchA table of up to 20 stories from the last 7 days, with dates and links
Give me post ideascontent-matrix24 to 40 post ideas from your pillars and 8 formats
Turn this outlier Reel into a scriptreels-scriptingA 30 to 45 second script on your newsletter topic and a matching caption, after a QA check
I need a thumbnail for my videoyoutube-thumbnailA brief to approve, then a Gemini prompt that uses your reference photo and brand colours
Write me a pinned commentpinned-commentFrom your post, an image prompt first, then a pinned comment of four short lines

Example

You have run voice-builder, so about-me.md and voice.md sit in your project, and you want a LinkedIn post about AI agents.

Write me a post about AI agents

Your agent reads your two voice files, asks which angle and framework you want, then gives a draft in a code block.

It adds two or three sentences on why it chose that hook and structure, and you get up to three rounds of changes before it saves the final post.

For newcomers and experts

If you are new. A skill is a text file that tells your agent how to do one job. Run voice-builder first. It starts by asking you questions, then asks for 3 to 5 pieces of your writing. Keep the two files it makes in your own project, not inside an installed skill folder.

For experts. The skills chain through files. voice-builder writes about-me.md and voice.md, newsletter-voice adds newsletter-voice.md on top, and the rest read them. The README lists what each workflow needs and what happens without it.

What each workflow needs
WorkflowNeedsWithout it
Niche researchLive web search or a browserDated sources you supply, or research pending
Post scorerYour post history, with Apify optionalAn editorial review, with no history comparison
Reel link analysisAPIFY_API_TOKEN, GOOGLE_AI_API_KEY, Node.js and two packagesA video you supply skips Apify, and a transcript gives a labelled script
Gemini visual skillsNothing to draft prompts, Gemini image access for picturesPrompts only
AnalyticsYour LinkedIn export, React and Recharts for the live previewThe computed analysis and source, preview pending

reels-scripting names Gemini 2.5 Flash as its model and says not to swap it without your approval. It also asks your approval before a paid provider run.

Good to know

The skills draft and review. They never publish for you.

  • Scope: the README calls the repository a public toolkit, not the maker’s full private LinkedIn system, his Figma pipeline or his accounts
  • Comments: scraping covers original post bodies and aggregate counts only, and never comments or replies
  • Images: the image skills give Gemini prompts, and the README says a picture must be opened and inspected before it counts as reviewed
  • Starter samples: voice-builder’s “use samples” borrows another author’s style until you replace it with your own writing
  • Updates: copied or uploaded skills do not refresh by themselves, so compare the new source with your copy and keep your changes
  • Language: several skills write British English unless your voice.md says otherwise, and they never use em dashes
  • Maker: Charlie Hills, MIT licensed. Mindket lists it and is not Charlie Hills

README and files

README
# Charlie Hills Social Media Skills: setup notes

This kit holds setup notes written by Mindket. It does not contain the skills or their reference files. Those install from Charlie Hills’s repository: https://github.com/charlie947/social-media-skills

License: MIT. Full text: https://github.com/charlie947/social-media-skills/blob/main/LICENSE

Social Media Skills is made by Charlie Hills. Mindket lists it and is not Charlie Hills. His README says a link back to his newsletter is appreciated: https://charliehills.substack.com

The pages these notes came from: the README (https://github.com/charlie947/social-media-skills/blob/main/README.md), each skill’s SKILL.md under https://github.com/charlie947/social-media-skills/tree/main/skills, the plugin file (https://github.com/charlie947/social-media-skills/blob/main/.claude-plugin/marketplace.json) and VERSIONS.md. Facts about where Claude Code loads skills and plugins come from Claude Code’s own docs, named where they are used. Mindket copied every command and ran none of them.

## What it is

Seventeen skills for Claude and Codex that cover LinkedIn posts, Instagram Reels, YouTube thumbnails, newsletters and analytics. They start from a voice profile: voice-builder interviews you, reads 3 to 5 pieces of your writing and saves about-me.md and voice.md in your project. The skills that write in your voice read those two files first.

The README calls this repository a public toolkit, not the complete private LinkedIn AI OS, its Figma production pipeline, or the maintainer’s accounts and configuration.

| Group | Skills |
|---|---|
| Voice foundation | voice-builder, newsletter-voice |
| LinkedIn | profile-optimizer, post-writer, graphic-designer, post-formatter, hook-generator, post-scorer, content-matrix, niche-research, gemini-infographic, gemini-carousel, quote-post |
| Instagram Reels | reels-scripting |
| YouTube | youtube-thumbnail |
| Community | pinned-comment |
| Analytics | analytics-dashboard |

Source: the README’s “Skill Categories”.

## Files in this kit

- setup/claude-code.md: the Claude Code plugin route, two commands
- setup/codex.md: the Codex skill-installer route and the project-local copy
- setup/claude-desktop.md: one skill folder zipped and uploaded
- setup/first-run.md: voice-builder first, what each skill needs from you, and updating

## What you need first

- Claude Code, Codex or Claude Desktop. The README documents an install route for each.
- Git for the Codex copy route.
- Your own writing: 3 to 5 samples for voice-builder. It can load a starter set instead, which borrows another author’s style until you swap in your own.
- An Apify token and a Google AI key, only for reels-scripting when you give it a Reel link. post-scorer can also use Apify, and it is optional.
- Gemini access, only to turn the image prompts into pictures. The prompts need no integration.

## What installs where

- Claude Code plugin: Claude Code keeps the plugin’s files itself, and its skills run under the plugin’s name (setup/claude-code.md).
- Codex: each skill folder under .agents/skills/ in your project (setup/codex.md).
- Claude Desktop: the skills you upload, one zip each (setup/claude-desktop.md).
- Your voice files: about-me.md and voice.md live in your own content project. The README says to keep them there and not inside an installed skill folder.

## Your data

The skills are instructions for your agent. What you paste, such as your writing samples, goes to your agent as in any other chat with it. Two routes use outside services only when you choose them: reels-scripting fetches a Reel through Apify and sends the video to Google’s Gemini 2.5 Flash, and post-scorer can fetch your own posts through Apify. Both run on your own accounts. The README says scraping is limited to original post bodies and aggregate counts, and that comments and replies are never collected.

## Update and remove

The README says copied or uploaded skills do not refresh when the source changes: pull the source with git, compare the installed folder with the new source, keep any files you changed, and reload in a fresh task. The plugin route updates through Claude Code’s own plugin mechanism. setup/claude-code.md says what Claude Code’s docs give for removing a plugin.
Files
README.md  4 KB
setup/claude-code.md  2 KB
setup/claude-desktop.md  1 KB
setup/codex.md  3 KB
setup/first-run.md  4 KB

5 files, 15 KB, downloaded as a zip.

What’s new

  • Version 2 · 2026-11-09

    Updated by Mindket to the current version of this kit.

  • Version 1 · 2026-11-07

Reviews

No ratings yet

Write a review

No reviews yet.

What people say

Posts and comments that name this project on Hacker News, Bluesky, X, Reddit, Facebook and YouTube, found by a weekly search and on Trendshift. Mindket doesn’t write or check them. Open one to read it where it was posted.

Searched

More from Mindket

See all from Mindket

anaksor-ocr

Sokheng · Kit

Reads Khmer and English text from documents, receipts and photos of signs, using a detector for text lines and a recogniser for each line. Suits developers who…

New

Qwen2.5-7B-Instruct as GGUF files in eight quantization sizes, to chat with locally through llama.cpp. For people who want a capable 7.6B model on their own co…

New

ACE-Step 1.5 XL SFT

ACE-Step · Kit

Generates music from text with a 4B diffusion decoder, 50 steps and guidance control, aiming for higher audio quality than the base. For creators with a GPU of…

New

Similar in Marketing & SEO

See all in Marketing & SEO

Huashu Design

alchaincyf · Skill

Turns a one-line request in your AI agent into a finished design: a launch animation, a clickable app prototype, an editable slide deck, an infographic or a se…

New29.4k

Pixel Art XL

nerijs · Kit

A pixel art style add-on for Stable Diffusion XL that turns text prompts into pixel-style images, with an optional LCM LoRA for faster runs. For game makers, d…

New

Licon MSR for LTX-2.5

LiconStudio · Kit

Keeps characters, clothing, objects and backgrounds consistent in generated video by feeding the model up to five reference images. A LoRA for LTX-2.5 in Comfy…

New