agentleFS
kristian-pressokristian-presso/claude-marketing-skills/skills/linkedin-post/SKILL.md

linkedin-post

Skill0 starsChanged 31 days agoMITView on GitHub

Drafts LinkedIn posts in the person's own voice, learned from three to five of their past posts rather than a generic template, from a rough idea, a one-liner, notes or a longer article. Also cuts one article into two or three standalone posts. Use when someone says "write a LinkedIn post", "turn this into a LinkedIn post", "draft a post about", "make this sound like me on LinkedIn", "write a short thought", or pastes an idea and asks for a post.

44 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use youtube-upload, proposal-writer and 3 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • BlackLotus1974Skill

    youtube-upload

    Upload videos to Atchalta's YouTube channel (office@atchalta.com). Use when the user asks to upload a video to YouTube, list YouTube videos, manage playlists, or update video metadata. Trigger phrases: 'upload to youtube', 'youtube video', 'publish video', 'post to youtube', 'list my videos', 'youtube playlists'.

    Also in 3 projects that use this

  • ncklrsSkill

    proposal-writer

    Expert sales proposal and pricing presentation strategist. Use when writing proposals, executive summaries, ROI business cases, pricing presentations, SOWs (Statement of Work), RFP responses, or competitive positioning documents. Covers proposal structure, terms positioning, design formatting, follow-up strategy, and win-rate optimization.

    Also in 3 projects that use this

  • Microsoft AzureSkill

    lint

    Run all linting and spell checking for TypeScript and Rust code. Run this before committing to catch style issues, clippy warnings, and misspellings.

    Also in 3 projects that use this

  • OpenAISkill

    brainstorming

    You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

    Also in 3 projects that use this

  • GoogleSkill

    grill-me

    A relentless interview to sharpen a plan or design.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 13 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 38 edited versions we read keep these.

  • 13

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Look for, in this order:

    line 17
  2. Distinctive:

    If there is nothing, ask once: "Paste three to five posts you have written that sound like you, and I will match them." If they would …

    line 24
  3. Distinctive:

    Every post makes one point.

    line 38
  4. Distinctive:

    If the brief is vague, ask one question: "What is the single point, and is it a claim, a number or a story?"

    line 46
  5. Distinctive:

    Default to the tier the person's own posts mostly use.

    line 57
  6. Distinctive:

    No engagement bait ("Agree?", "Thoughts?", "What do you think?").

    line 82
  7. Distinctive:

    No hashtags unless the person uses them.

    line 83
  8. Distinctive:

    No emoji unless the person uses them.

    line 84
  9. Very specific:

    Read references/hooks.md for opening patterns that work and ones to drop.

    line 86
  10. Distinctive:

    If they ask for alternatives, give a second version with a different opening line only.

    line 103
  11. Distinctive:

    When the input is an article or long piece and the person wants several posts:

    line 107
  12. Distinctive:

    Ask for an explicit yes to that exact text.

    line 126
  13. Distinctive:

    Only then post or schedule, and report back the link or scheduled time.

    line 127
  14. Distinctive:

    Never publish on an ambiguous "post it" without that confirmation.

    line 129

“Common” is measured: kept by at least 6 in 10 of the 38 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

LinkedIn post

Most AI-drafted LinkedIn posts sound the same: a one-word hook, a line break after every sentence, a lesson, a question at the end. People scroll past them because they recognise the shape.

This skill writes in the person's voice. It learns that voice from their own posts, records it as a short profile, and checks every draft against it.

Step 1: get the voice

Look for, in this order:

  1. A voice profile already in the conversation or project (from an earlier run).
  2. Three to five of their past LinkedIn posts, pasted or in a file. Ask for ones they are happy with, ideally a mix of short and long.
  3. Other writing of theirs: blog posts, newsletters, emails.

If there is nothing, ask once: "Paste three to five posts you have written that sound like you, and I will match them." If they would rather not, write in a plain, direct default (short sentences, concrete detail, no hype) and say that is what you did.

Build a voice profile from the samples. Read references/voice-profile.md for the method and the template. It takes a minute and covers: typical length, how posts open, line and paragraph rhythm, vocabulary and words they never use, how they handle numbers and names, tone, how posts end, and formatting habits (emoji, hashtags, lists, dashes).

Show the profile in five to eight lines and offer to save it so the next run can skip this step. Never copy sentences from the samples into new drafts.

Step 2: find the one point

Every post makes one point. From the person's idea, notes or article, write down:

  • The point, in one sentence.
  • The proof: the number, story, example or before-and-after that makes it believable. It must come from the person. Never invent a statistic, client, result or anecdote; if the post needs one, ask, or leave [ADD: ...].
  • Who it is for, if it is not obvious.

If the brief is vague, ask one question: "What is the single point, and is it a claim, a number or a story?"

Step 3: pick a length

TierWordsUse for
Short40 to 120One observation, one sharp line. A "thought".
Medium150 to 250A story or a take with one specific number.
Long300 to 450A real multi-step argument or how-to.

Default to the tier the person's own posts mostly use. If they have no samples, default to short. LinkedIn shows roughly the first 200 characters before "see more", so whatever the length, the opening line has to earn the click.

Step 4: draft

Shape:

Opening line: the point, a specific number, or the most surprising detail.

One or two lines that set it up, only if needed.

The body: the proof, in their rhythm.

The landing: one line the reader could repeat.

Optional: one question or call to action, only if their posts usually have one.

Apply the profile, then these defaults unless the profile says otherwise:

  • The point in the first line. No warm-up.
  • Concrete over abstract: names, numbers, places, prices.
  • One idea. If two ideas compete, offer to split them into two posts.
  • No engagement bait ("Agree?", "Thoughts?", "What do you think?").
  • No hashtags unless the person uses them; if they do, three at most at the end.
  • No emoji unless the person uses them.

Read references/hooks.md for opening patterns that work and ones to drop.

Step 5: check before showing

  • Reads like their samples, not like a template. Read it against two of them.
  • One point, stated in the opening line.
  • Every fact, number and anecdote came from the person. Do not invent what customers did or said ("most never opened it") to fill a gap; mark it [ADD: ...].
  • Within the chosen length.
  • No phrase from the "drop" list in references/hooks.md.
  • Spelling and punctuation follow their locale and habits.

Output

One draft, not three. Print it ready to paste. Under it, one line: tier, word count, and anything marked [ADD: ...].

If they ask for alternatives, give a second version with a different opening line only.

Cut mode: one article, several posts

When the input is an article or long piece and the person wants several posts:

  • Find the genuinely distinct points. Two or three is typical.
  • Each post stands alone: no "part 1", no "link in comments for the rest", no teaser that holds back the payoff.
  • Different angles, not the same point reworded. If the article has one real point, write one post and say so.
  • Short or medium tier for each.
  • Number them and suggest spacing them at least a few days apart.

The content-atomisation skill in this plugin plans the wider multi-channel version.

Publishing

This skill drafts. It does not post. If the person has a LinkedIn or social scheduling tool connected and asks to publish:

  1. Show the exact final text and the target (personal profile or company page). Ask which if they did not say.
  2. Ask for an explicit yes to that exact text.
  3. Only then post or schedule, and report back the link or scheduled time.

Never publish on an ambiguous "post it" without that confirmation.

Example

User: Post about how we stopped doing discovery calls for orders under 500 pounds. Here are three of my old posts: (pasted)

Claude: Your profile: short posts (80 to 120 words), open with a plain statement of what you did, no questions at the end, British spelling, numbers written as figures, occasional dry aside in brackets, never uses hashtags.

We stopped doing discovery calls for orders under £500.

They were taking 40 minutes each. Most ended with the client picking the package they had already looked at on the website.

Now small orders go straight to a short form and a booking link. Bigger jobs still get a call (those are the ones where the call changes the answer).

[ADD: what happened to conversion or time saved]

The call was for us, not for them.

Short tier, 78 words. Needs one result from you where marked.

Configuration

Two things vary per person:

  • Their voice samples or saved profile. The skill is only as close to their voice as the samples it learned from.
  • Locale. British or American spelling, currency and date format. Take it from their samples if not stated.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.