# Skill: Content Engine

A LinkedIn + newsletter content engine. Train it on your past posts once. Give it 3 ideas. Get back 3 publish-ready posts in your voice in 2 minutes.

---

## How to build this skill in Claude Code

1. Open **Claude Code** inside the Claude App (Mac / Windows) or Antigravity.
2. In your project folder, create this path: `.claude/skills/content-engine/`
3. Inside that folder, create a file named `SKILL.md`
4. Paste this entire document into that file and save.
5. Train your voice (next section).
6. Reload Claude Code so it picks up the new skill.
7. Invoke with `/content-engine` followed by your topic ideas.

---

## Train your voice (one-time)

Inside the same folder, create a sub-folder: `voice-training/`

Drop in:
- 10-20 of your best past LinkedIn posts as `.md` files (one post per file, just the post body)
- Optional: a `sign-off.md` with your standard sign-off block
- Optional: a `source-material.md` with your client wins, frameworks, signature stories

The more samples you give it, the closer the drafts match your tone.

**No external tools or MCPs needed for this skill.** It runs purely on the voice training set.

---

## How to invoke

```
/content-engine 3 ideas:
1. [topic 1]
2. [topic 2]
3. [topic 3]
```

The skill outputs 3 posts, ready to edit and publish.

---

## What the skill does

For each idea, the skill:

1. Reads the topic
2. Reads your voice training set (past posts, frameworks, named clients, tone)
3. Picks the right post structure (hook → setup → body → CTA → sign-off)
4. Drafts in your exact voice
5. Outputs a clean, publish-ready post

---

## The 5 voice rules (HARD)

1. **Hook on its own line.** Bold stat, claim or 2-4 word opener. Never an intro sentence.
2. **One sentence per line.** White space is structural. No walls of text.
3. **Real specifics.** Real names, numbers, timelines.
4. **Max 1 em dash per post.** Default to zero.
5. **No hashtags. Ever.**

---

## The 4 content pillars

Tag every post into one of these:

- **Knowledge / Value** — tools, frameworks, AI use cases
- **Personal / Journey** — founder story, behind-the-scenes
- **Authority / Social Proof** — client case studies, real numbers
- **Promotional (max 30%)** — community, newsletter, lead magnets

---

## What NOT to do (anti-AI patterns)

The skill catches and removes these before output:

- "I wanted to reach out" / "I hope this finds you well"
- Three-part lists ("fast, reliable, and scalable")
- Negative parallelism ("It's not just X, it's Y")
- AI vocab: leverage, delve, crucial, pivotal, landscape, testament
- Generic CTAs ("Let me know what you think!")
- More than 1 em dash per post

---

## The 5x output move

Once you have a strong post, the skill repurposes it:

1. **LinkedIn post** (the original)
2. **Newsletter section** (expanded, with one extra story)
3. **YouTube short script** (90 seconds, hook + 3 beats + CTA)
4. **Skool community post** (conversational, ends with a question)
5. **X / Twitter thread** (5-7 tweets)

Same idea. Five distribution surfaces. Forty minutes of work, not five hours.

---

## Need help configuring this for your voice?

Join the RecruiterGTM community at **https://skool.com/recruitergtm** — Daniyal sets up the Content skill with your past posts, sign-off, 4 pillars and client wins in a 60-minute call. Free for TIYYS attendees this week.

---

Built by Reyhan Khan · RecruiterGTM