LinkedIn Content Generation

June 1, 2026 · View on GitHub

A production-ready Claude Code skill that turns your AI assistant into a personal LinkedIn Content Strategist. Generate SEO-optimised posts, carousels, newsletters, and content calendars — with a reinforcement learning memory that gets smarter every time you use it.

✨ What This Skill Does

Once installed, type a slash command in Claude Code and get professional LinkedIn content in seconds:

Command	Output
`/generate-post`	A ready-to-publish LinkedIn post with hook, body, CTA, and hashtags
`/generate-carousel`	Numbered slide-by-slide carousel content + LinkedIn caption
`/generate-newsletter`	A full long-form newsletter edition in Markdown
`/generate-calendar`	A 30-day content calendar table with hooks and CTAs
`/feedback`	Saves what worked to memory for smarter future content
`/show-memory`	Displays your current preferences and learning history
`/clear-memory`	Resets memory back to defaults

Reinforcement Learning built-in: Every time you tell the skill what worked, it updates memory.md. Every future generation reads that memory and adapts to your personal voice, tone, and style — automatically.

📁 Folder Structure

linkedin-content-skill/
│
├── README.md                        ← You are here
│
├── scripts/                         ← Python scripts (visible & editable)
│   ├── generate_post.py             ← Post prompt builder
│   ├── generate_carousel.py         ← Carousel prompt builder
│   ├── generate_newsletter.py       ← Newsletter prompt builder
│   ├── generate_calendar.py         ← Calendar prompt builder
│   ├── memory_manager.py            ← Read/write/clear memory.md
│   ├── utils.py                     ← Shared LinkedIn SEO prompt engine
│   └── memory.md                    ← Your RL database (edit this!)
│
└── .claude/                         ← Hidden folder (Claude Code reads this)
    └── skills/
        ├── generate-post/SKILL.md
        ├── generate-carousel/SKILL.md
        ├── generate-newsletter/SKILL.md
        ├── generate-calendar/SKILL.md
        ├── feedback/SKILL.md
        ├── show-memory/SKILL.md
        └── clear-memory/SKILL.md

Note: The .claude/ folder is hidden on macOS by default. Press ⌘ Cmd + Shift + . in Finder to reveal hidden files.

🚀 Installation — Step by Step

Prerequisites

Claude Code installed (npm install -g @anthropic-ai/claude-code)
Python 3.8 or higher (python3 --version to check)
A Claude Pro or Team account

Option A — Install as a Project-Level Skill

Use this if you want the skill available only inside a specific project folder.

Step 1: Download this repository

# Clone via git
git clone https://github.com/yourusername/linkedin-content-skill.git

# OR download the ZIP from GitHub and unzip it

Step 2: Copy the skill into your project

# Navigate to your existing project (or create a new folder)
mkdir ~/my-project && cd ~/my-project

# Copy the .claude folder and scripts folder into your project root
cp -r /path/to/linkedin-content-skill/.claude .
cp -r /path/to/linkedin-content-skill/scripts .

Your project should now look like:

my-project/
├── .claude/
│   └── skills/
│       ├── generate-post/SKILL.md
│       └── ... (all 7 skills)
└── scripts/
    ├── generate_post.py
    └── ... (all scripts + memory.md)

Step 3: Launch Claude Code

cd ~/my-project
claude

Step 4: Verify installation

Type / in the Claude Code prompt. You should see all 7 LinkedIn commands in the autocomplete list.

Option B — Install as a Global User-Level Skill

Use this if you want the skill available in every project you open with Claude Code.

Step 1: Download this repository

git clone https://github.com/yourusername/linkedin-content-skill.git

Step 2: Copy skills to your global Claude directory

# Create the global skills directory if it doesn't exist
mkdir -p ~/.claude/skills

# Copy all skill definitions globally
cp -r /path/to/linkedin-content-skill/.claude/skills/* ~/.claude/skills/

# Create a permanent home for the scripts (recommended: ~/.claude/linkedin-scripts)
mkdir -p ~/.claude/linkedin-scripts
cp -r /path/to/linkedin-content-skill/scripts/* ~/.claude/linkedin-scripts/

Step 3: Update script paths in each SKILL.md

Because the scripts now live at ~/.claude/linkedin-scripts/, update the bash commands in each .claude/skills/*/SKILL.md file. Replace:

# Old path
python3 scripts/generate_post.py

# New path (use the full absolute path)
python3 ~/.claude/linkedin-scripts/generate_post.py

Repeat for all 7 SKILL.md files.

Step 4: Launch Claude Code from any folder

claude

Type / — your LinkedIn commands are now available globally in every project.

Personalising the Skill

Step 1: Set Your Niche and Tone (Do This First)

Open scripts/memory.md in any text editor and update the top section:

## 🧠 Core Identity & Tone
- **Primary Niche:** AI & Technology        ← Change this to your niche
- **Tone:** Professional and story-driven   ← Change this to your style
- **Voice:** First-person, confident        ← Describe your writing voice

This is the most important personalisation step. Every script reads this file before generating.

Step 2: Start Using and Giving Feedback

The skill learns from you over time. After generating content you love:

/feedback The contrarian hook about AI replacing developers got 500+ reactions. Save this.

Claude Code will automatically run memory_manager.py and store the learning. Future content will reflect it.

🛠️ How to Modify & Extend the Skill

To change how a command works — edit the SKILL.md file

Each command's behaviour is defined in its SKILL.md:

What you want to change	File to edit
How Claude interprets `/generate-post`	`.claude/skills/generate-post/SKILL.md`
How Claude interprets `/generate-carousel`	`.claude/skills/generate-carousel/SKILL.md`
How Claude interprets `/generate-newsletter`	`.claude/skills/generate-newsletter/SKILL.md`
How Claude interprets `/generate-calendar`	`.claude/skills/generate-calendar/SKILL.md`
How Claude saves feedback	`.claude/skills/feedback/SKILL.md`

Example: To make /generate-post always ask for the niche before running, open .claude/skills/generate-post/SKILL.md and add to Step 1:

If `niche` is not specified, ALWAYS ask the user before proceeding.

To change the content quality and SEO rules — edit `utils.py`

scripts/utils.py is the brain of the skill. It contains all the LinkedIn SEO rules injected into every prompt:

Hook formulas — Edit LINKEDIN_SEO_RULES to add new hook patterns
Structure rules — Change the post structure (Hook → Body → CTA)
Hashtag limits — Adjust the hashtag count or strategy
Tone rules — Add new writing style guidance

To add new tone or style options — edit the generator scripts

For example, to add a "humorous" tone to /generate-post:

Open scripts/generate_post.py and add to TONE_GUIDE:

TONE_GUIDE = {
    ...
    "humorous": "Use wit and light humour. Keep it professional but make the reader smile.",
}

Then update .claude/skills/generate-post/SKILL.md to include humorous as an option.

To add a brand new command

Create a new folder: .claude/skills/your-command-name/

Create SKILL.md inside it with the format:

---
description: What this command does. Usage: /your-command-name <args>
allowed-tools: Bash
---
## Step 1 — Parse Arguments
...
## Step 2 — Run Script
```bash
python3 scripts/your_script.py --arg "$ARGUMENTS"

Create scripts/your_script.py with your logic
Restart Claude Code — the command is live immediately

💡 Usage Examples

# Generate a controversial post about AI in the tech niche
/generate-post AI agents replacing junior developers in tech, tone: controversial

# Create a 10-slide carousel about LinkedIn growth hacks
/generate-carousel 10 LinkedIn growth hacks nobody tells you, 10 slides, listicle style

# Plan a full month of content with a lead generation goal
/generate-calendar AI & Technology niche, 30 days, goal: leads, 4 times a week

# Write a deep-dive newsletter edition
/generate-newsletter The state of AI agents in 2025, long format, niche: Technology

# Save what worked to memory
/feedback That storytelling post about my first startup failure got 800 reactions — save the narrative hook style

# Check what the skill remembers about you
/show-memory

🔄 Reinforcement Learning — How It Works

You generate a post
        ↓
You post it on LinkedIn
        ↓
It performs well
        ↓
You run: /feedback <what worked>
        ↓
Claude runs: python3 scripts/memory_manager.py add ...
        ↓
memory.md is updated with a timestamped entry
        ↓
Every future generation reads memory.md
        ↓
Content becomes more aligned with YOUR voice over time

The memory.md file grows richer with every piece of feedback. After 10-20 feedback entries, the skill generates content that sounds genuinely like you.

Script Reference

`scripts/generate_post.py`

python3 scripts/generate_post.py --topic "<topic>" --niche "<niche>" [--tone <tone>] [--style <style>]

Flag	Options	Default
`--tone`	`professional` `storytelling` `controversial` `educational` `motivational`	`professional`
`--style`	`list-based` `text-only` `data-driven` `contrarian` `storytelling`	`list-based`

`scripts/generate_carousel.py`

python3 scripts/generate_carousel.py --topic "<topic>" --niche "<niche>" [--slides <n>] [--style <style>]

Flag	Options	Default
`--slides`	Integer 3–12	`7`
`--style`	`how-to` `listicle` `myth-busting` `framework` `story-arc`	`listicle`

`scripts/generate_newsletter.py`

python3 scripts/generate_newsletter.py --topic "<topic>" --niche "<niche>" [--length <length>] [--title "<title>"]

Flag	Options	Default
`--length`	`short` (~700w) `medium` (~1200w) `long` (~2000w)	`medium`

`scripts/generate_calendar.py`

python3 scripts/generate_calendar.py --niche "<niche>" [--days <n>] [--frequency "<freq>"] [--goal <goal>]

Flag	Options	Default
`--days`	Any integer	`30`
`--frequency`	e.g. `"daily"` `"3 times a week"`	`"3 times a week"`
`--goal`	`awareness` `engagement` `leads` `authority` `growth`	`growth`

`scripts/memory_manager.py`

python3 scripts/memory_manager.py add --id "<id>" --feedback "<text>" [--tags "<tags>"]
python3 scripts/memory_manager.py read
python3 scripts/memory_manager.py clear

Platform Compatibility

Platform	Support
Claude Code (CLI)	✅ Full native support — slash commands auto-register
Claude Projects (Web)	⚠️ Partial — paste `SKILL.md` contents into Custom Instructions; upload `memory.md` to Knowledge
GitHub Copilot	⚠️ Partial — add `SKILL.md` content to `.github/copilot-instructions.md`
Cursor	⚠️ Partial — add `SKILL.md` content to `.cursorrules`

Contributing

Contributions are welcome. To add a new command or improve an existing one:

Fork this repository
Create a branch: git checkout -b feature/new-command
Make your changes following the structure above
Submit a Pull Request with a description of what you changed

📄 License

MIT License — free to use, modify, and distribute. See LICENSE for details.

Built for LinkedIn creators who want AI that learns their voice, not generic content.