Skip to content
Marketing
Skill

/write-post

Use when user wants to draft a single LinkedIn text post and has a specific angle, or passes a seed from find-ideas. Trigger phrases include "write a LinkedIn post," "draft a post about X," "turn this into a post," "post this on LinkedIn." If no specific angle, route to

From plugin
linkedin-maxxing
417 skills17 commands
Install
$ npx -y skills add warpirate/linkedin-maxxing --skill write-post --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/write-post

Context preview

The summary Claude sees to decide when to auto-load this skill.

Use when user wants to draft a single LinkedIn text post and has a specific angle, or passes a seed from find-ideas. Trigger phrases include "write a LinkedIn post," "draft a post about X," "turn this into a post," "post this on LinkedIn." If no specific angle, route to

SKILL.md

write-post.SKILL.md
name: write-post
description: |
  Use when user wants to draft a single LinkedIn text post and has a
  specific angle, or passes a seed from find-ideas. Trigger phrases
  include "write a LinkedIn post," "draft a post about X," "turn this
  into a post," "post this on LinkedIn." If no specific angle, route to
  find-ideas first instead.
license: MIT

Write post

This is the most-used skill in the repo and the easiest one to do badly. Most LinkedIn AI tools produce posts that sound exactly like other LinkedIn AI tools. This skill exists to produce posts that sound like the user, on a specific angle they actually want to say, optimized for how LinkedIn ranks content in 2026.

Why this skill exists

LinkedIn ranks content on dwell time (how long readers spend before scrolling), the "see more" expand rate (whether they click past the fold), and comment quality (specific, substantive comments). None of those reward templated content. They reward content with a real point, written in a way that holds attention.

The default failure mode of LinkedIn AI writing: hook-value-CTA structure, em dashes, "the real question is," rule-of-three phrasing, generic enthusiasm, and a closing question that feels like engagement bait. LinkedIn has explicitly downranked all of these. This skill exists to not do them.

When to use this skill

Trigger when:

  • The user has a specific angle they want to post about
  • The user passes a seed from find-ideas
  • The user pastes a rough draft and wants it sharpened
  • The user uploads a piece of work (PR description, decision doc, customer email thread) and says "post this"

Do NOT trigger when:

  • The user has no specific angle. Send them to find-ideas first.
  • The user wants a slide carousel. Send them to build-carousel.
  • The user wants a 600+ word article. Send them to write-longform.
  • The user wants a video script. Send them to write-video-script.

Before drafting

Do these in order:

1. **Read voice-profile.md if it exists.** It will be at the workspace root or `~/.linkedin-maxxing/voice-profile.md`. If present, use it. If absent, do not invent one; just ask the user one question about tone (see "voice fallback" below).

2. **Identify the angle.** If the user passed a seed from find-ideas, use it directly. If they passed a vague idea, refuse to draft and send them to find-ideas. If they pasted a rough draft, find the angle within their draft (the strongest sentence is usually it).

3. **Ask up to three sharp questions, batched, only if material is missing.** What specifically happened, who was involved, what was the actual number, what is the takeaway the user wants. Do not ask about audience, goals, hashtags, or CTA. Those are noise.

4. **Skip questions whose answers are in context.** If the user already told you the role, company, or specific detail, do not ask again.

How to draft

LinkedIn text posts in 2026 work best at 1,000-1,300 characters. The "see more" cut is at 210 characters, so the first two lines have to earn the click.

Write the draft following these rules:

**Lines 1-2 (the hook, 210 character budget).** The hook is the single most important part of the post. It has to make someone scrolling at speed stop and click "see more." Hooks that work in 2026: a specific number, a confession ("I shipped X without doing Y"), a contrarian claim, a story setup that promises a payoff, an unusual fact. Hooks that fail: generic questions ("Have you ever thought about X?"), motivational openers, "Here's what I learned this week," anything starting with "Let's talk about."

**The middle (the substance).** This is where dwell time is earned or lost. Tell the actual story or make the actual argument. Specific details, named tools, real numbers, exact quotes. Short paragraphs, one or two sentences each. Whitespace between paragraphs (LinkedIn renders blank lines as visible breaks). Vary sentence length: a few short sentences, then a longer one, then short again.

**The close.** Three options, in order of preference: a single concrete sentence that lands the point, an honest question that you actually want answered (not "thoughts?"), or just stop after the strongest line. Avoid: "If this resonated...", "Agree?", "What do you think?", any CTA to subscribe or DM, any reference to "thought leadership."

Things that are explicitly out

  • Em dashes and en dashes. Humanizer §14, hard rule.
  • The "it's not X, it's Y" construction. LinkedIn now explicitly detects this and downranks.
  • Rule-of-three lists where the third item is filler.
  • "Here is what I learned" / "Three lessons from..." openers.
  • Bold text for emphasis on random phrases.
  • Bullet points unless the content is genuinely a list (steps, items, comparisons). Most posts should not have bullets.
  • Hashtags more than 1-3. None at all is also fine. Never use ten.
  • External links in the body. Mentioning a link costs about 60% of reach. If a link is essential, mention "link in comments" once and put it in the first comment.
  • Emojis as decoration. Emojis as occasional emphasis are fine if the voice profile shows the user does that. Most users should not.
  • "Conclusion:" or "TL;DR:" headers in a 1,200-character post.

Things that work

  • One specific, hard-to-fabricate detail in the first three lines.
  • A change of mind, a specific moment, or a number nobody else would have.
  • Short paragraphs. Visible whitespace.
  • A line that admits something the user is not supposed to admit.
  • Ending without a question, just the strongest sentence.

Voice fallback (if voice-profile.md is missing)

Ask ONE question, then commit: "I don't have a voice profile yet. Quick: when you write at work, do you tend toward (a) short and punchy, (b) longer and more thoughtful, or (c) somewhere in between? Anything you definitely don't want in the post (emojis, hashtags, em dashes, CTAs)?"

Get the answer, draft. Suggest at the end: "Want me to run train-voice on some past posts so I match your voice better next

Read more
Ships withlinkedin-maxxing

17 Claude Code skills + slash commands for substance-first LinkedIn growth. Profile audit, content drafting (posts, carousels, longform, video, DMs, comments), performance analysis, and a Wikipedia-based humanizer. Anti-template, anti-slop, open source, MIT.

Get the whole plugin

Other skills on linkedin-maxxing.