01What this is
A Claude Skill trained on 50 decoded viral LinkedIn posts and data from 600,000 LinkedIn posts. It writes, rewrites, and explains LinkedIn strategy, built around how the 2026 algorithm actually works.
One skill. Three modes. No generic output.
Step 1, Install it
The skill file is included in the PDF download. Save it as SKILL.md in a folder named linkedin-post-generator, then zip that folder.
- Go to claude.ai
- Click your profile → Settings → Customize → Skills
- Click + and upload the zip file
- Toggle the skill on
Takes 60 seconds. Works in every conversation after that.
Step 2, Understand the three modes
The skill automatically detects what you need based on how you ask.
Write mode
You have a topic or idea. You want a post.
How to trigger it:
"Write a LinkedIn post about [topic]"
"I want to post about [idea], help me draft something"
"I had a client situation this week: [describe it]. Turn this into a post."
What you'll get:
- A full draft in your goal's format (thought leadership, lead gen, engagement, shares)
- 2 alternative hooks in case the first direction doesn't feel right
- A visual suggestion (image, carousel, or graphic)
- One specific action to take in the first 90 minutes after posting
Rewrite mode
You have a draft that isn't working. You want it better.
How to trigger it:
"Here's a post I wrote, can you punch it up: [paste post]"
"This doesn't sound like me. Fix it."
"Improve this: [paste post]"
What you'll get:
- Diagnosis of what's working and what's weak
- A rewritten version with stronger hook, better structure, sharper closer
- 2 alternative hooks
- A note on what changed and why
Strategy mode
You have questions about reach, timing, or why something underperformed.
How to trigger it:
"Why did my post tank?"
"When should I post?"
"Should I put the link in the body or comments?"
"Why is my reach dropping?"
"What engagement signals matter most to LinkedIn right now?"
What you'll get:
- Answers grounded in the actual 360Brew algorithm mechanics
- Specific actions, not generic advice
Step 3, Get better output faster
Tell it your goal upfront. The skill writes differently depending on what you want:
- Reach/shares → structures for strangers to share it
- Comments/engagement → structures for specific discussion
- DMs/leads → structures with soft next steps
"Write a LinkedIn post about [topic]. Goal is shares, I want to reach new people."
Give it your best past posts. The more context, the more it sounds like you.
"Here are 3 posts I've written that I liked: [paste them]. Now write a post about [topic] in the same voice."
Be specific about your credibility. Generic topics produce generic posts.
"Write a post about cold email. My angle is: I've reviewed 50,000 cold emails in the last 3 years and the same 4 mistakes keep appearing."
Step 4, After you post
The first 90 minutes matter more than anything else. LinkedIn's algorithm tests your post with a small seed audience in that window.
Do this immediately after posting:
- Stay on LinkedIn and engage other people's content (author stickiness is a real signal)
- Reply to the first comments that come in, fast
- Leave 20 to 30% of comments unanswered, fully answered looks like the conversation is over
- Put any links in the first comment, never the post body
06Quick reference, what kills reach in 2026
| ❌ Do not do this | Why |
|---|---|
| Links in the post body | 38% reach drop, put in first comment |
| Edit after publishing | Resets algorithmic distribution entirely |
| More than 3 to 5 hashtags | Broad hashtags are now a spam signal |
| Post and immediately leave | Kills author stickiness score |
| AI-sounding content | 360Brew detects it and suppresses it |
07Quick reference, engagement weights
LinkedIn scores your content based on what it earns:
| Action | Weight |
|---|---|
| Like | 1 |
| Comment | 2 |
| Long comment (10+ words) | 2.5 |
| Share | 4 |
| Share with caption | 4.5 |
| DM about the post | Highest signal |
Shares are worth 4x a like. Write posts that make people want to share, not just react.
08Example prompts to copy
Start from scratch:
"Write a LinkedIn post. Topic: [your topic]. My take: [your specific angle]. Goal: shares, I want to reach new people outside my network."
Rewrite a draft:
"Here's a LinkedIn post I wrote. Rewrite it with a stronger hook and cleaner structure. Keep my voice: [paste post]"
Check if it sounds human:
"Does this post sound AI-generated? What would you change: [paste post]"
Algorithm question:
"I posted yesterday at 9am and got almost no reach. What could have caused this?"
Need any help setting this up?
Software, accelerator and startup background. Runs the GTM architecture side: how the system is designed, what it is built on, and where AI actually earns its place.
LinkedIn ↗YouTube ↗