agentleFS
shreyvaughnshreyvaughn/claude-skills/skills/linkedin-post/SKILL.md

linkedin-post

Skill1 starsChanged 5 months agoMITView on GitHub

Draft or sharpen a LinkedIn post (non-lead-magnet). Combines your own authentic voice with a fresh, high-performing structural mechanic. Three modes - generate from a topic, sharpen a draft, or port a proven mechanic from another creator.

41 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, 14 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 36 edited versions we read keep these.

  • 14

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If they paste the lines, steps, or section they want written, treat their text as the source of truth.

    line 37
  2. Distinctive:

    If ambiguous, ask one line.

    line 49
  3. Distinctive:

    Read them all. Note for this draft:

    line 55
  4. Distinctive:

    If the user cannot provide past posts, ask 3 quick questions to bootstrap a voice profile: who are you to your audience, what is the one …

    line 62
  5. Distinctive:

    Read the strongest examples.

    line 70
  6. Distinctive:

    Read the target creator's catalog and sort by actual engagement.

    line 74
  7. Distinctive:

    Transfers: if the mechanic works once the language is swapped: call-out, slippery-slope anatomy, comment-bait roll-call, numbered absurd …

    line 78
  8. Distinctive:

    Does NOT transfer: if it depends on the creator's unique register: signature begging hooks, a named-authority peg as the entire frame, …

    line 79
  9. Distinctive:

    Comment-bait roll-call: - ask a question with absurdly specific qualifiers, invite people to drop themselves in the comments.

    line 91
  10. Distinctive:

    After hook and archetype are picked, gather 4 to 6 specific receipts to fill the body.

    line 100
  11. Distinctive:

    Other creators' signature language: Do not lift recognizable catchphrases.

    line 115
  12. Distinctive:

    Stale hook templates parroted as-is: Use old hooks as voice reference, bring a fresh mechanic from Step 2.

    line 120
  13. Distinctive:

    When the user picks one:

    line 181
  14. Distinctive:

    Never silently rewrite the voice.

    line 184

“Common” is measured: kept by at least 6 in 10 of the 36 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

This skill produces LinkedIn posts by combining two ingredients:

  1. Your voice, layout, and identity - the durable part. It defines who you are and how you write. Do not reinvent it.
  2. A fresh structural mechanic - the timely part. It keeps the format current so you are not still running last quarter's hook bank.

The skill is the framework. The voice comes from you. The mechanic comes from what is working in your niche right now.

The 6 principles

1. Voice from you, mechanic from elsewhere. Pull your voice, layout, and identity from your own best-performing posts. Pull the mechanic (hook shape, structure) from current top performers in your space. The combination is what makes a post both fresh and authentically yours. Never invent a voice. Never lock in a single hook bank and run it forever.

2. Receipts at the outcome layer, not the implementation layer. Concrete, verifiable specifics travel: roles changed, money saved or earned, things shipped, named results. Avoid implementation trivia ("I wrote a script", "I set up a job"). Say what the thing does, not how it was built, unless the how is the actual subject.

3. Tension is mandatory. Specificity without tension reads like a memo. Every post needs one line of friction with reader expectations: a contrarian claim, an oddly specific scene, a number that should not be possible, a confession. Receipts make it credible. Tension makes it travel.

4. Mobile layout is the format.

  • One sentence per line. This is the default.
  • Short sentences, most under 12 words.
  • Arrow bullets for lists. Use them freely.
  • Single-line paragraphs. White space is a feature.
  • Voice markers you actually use (recurring phrases, a signature transition line) used naturally, not decoratively.

5. Lift the mechanic, not the voice, when porting. When porting from another creator, keep only mechanics that survive a voice swap into your own register. If a format only works because of the source creator's unique register (a signature begging hook, a named-authority peg that is their whole frame, all-caps interjections, a branded numbered series tied to their persona), skip it. Port the transferable structures: call-outs, slippery-slope anatomies, comment-bait roll-calls, numbered absurd lists, industry stances, one-line drops. Rank the source creator's catalog by actual top-quartile performance before picking. Do not lift mid-tail formats.

6. When the user hands you exact copy, that copy is the spec. If they paste the lines, steps, or section they want written, treat their text as the source of truth. Sharpen layout, flow, and voice around it. Do not invent new content over theirs, swap their examples, or reorder their points unless they ask.

Two standing rules for any teaching or how-to section:

  • Coin a fresh-sounding name for each move, then explain it plainly. The header should feel like a new concept the reader is learning ("Set the finish line", "Give it a mirror", "The two-strike rule"). The explanation underneath must be understandable by a non-technical reader. Translate every jargon term into a plain outcome.
  • Make every instruction copy-pasteable. Where a step is something the reader does, write it as the literal sentence they can use, anchored to a real, recognizable task.

Three modes

  • Mode A - generate 3 variants from a topic
  • Mode B - sharpen a draft the user wrote
  • Mode C - port from a target creator (rank their catalog, lift only transferable top-quartile mechanics)

If ambiguous, ask one line.

Step 1 - Establish the voice anchor (always first, every mode)

Get the user's 5 to 10 best-performing posts. Ask them to paste their top posts, or point you at their profile so you can read recent high performers.

Read them all. Note for this draft:

  • Layout DNA - sentence length, line-break density, how they use bullets, where they put whitespace.
  • Voice markers they actually use (recurring openers, signature transition lines, a habitual sign-off).
  • Identity claims they make - what they call themselves, what they call their work, their relationship to their own results, what they share vs hold back.
  • Hook shapes that worked historically - note them, but treat as one input, not the prescribed list. Old hooks may read as "old you". Use them as voice and structure reference. Pull mechanic freshness from Step 2.

If the user cannot provide past posts, ask 3 quick questions to bootstrap a voice profile: who are you to your audience, what is the one specific kind of proof you can show, and what is a phrase you say that nobody else does.

Step 2 - Pull a current mechanic

Identify what structural mechanics are landing right now in the user's niche. Sources, in order of preference:

  • Posts the user already admires (ask them to name 2 or 3 creators).
  • A quick scan of recent high-performing posts in their space.

Read the strongest examples. Pick distinct mechanics for the drafts that survive a voice swap into the user's register. Vary across drafts.

For Mode C - study one creator

Read the target creator's catalog and sort by actual engagement. Read the top quartile first. Identify recurring archetypes that appear there. Drop anything from the mid-tail. If a format is not in their top performers, it is not their winning shape.

For each top-quartile archetype, apply the transferability test:

  • Transfers if the mechanic works once the language is swapped: call-out, slippery-slope anatomy, comment-bait roll-call, numbered absurd list with a teaching nugget, one-line drop with image, industry stance with named receipts, stat-pivot, failure-confession, personal-stack-reveal.
  • Does NOT transfer if it depends on the creator's unique register: signature begging hooks, a named-authority peg as the entire frame, unhinged comedic interjections, a branded numbered series tied to their persona, proximity flexes that depend on their existing visibility.

The archetype library

These are durable post shapes. Each pairs a mechanic with a voice-neutral description so any user can fill it with their own receipts.

  1. One-line drop - a single absurdist or oddly specific line, no body, no bullets. Pair with one image when possible.
  2. Body-posture hook - open with a vulnerable or dramatic stance, land a one-line thesis, back it with receipts, close with a CTA. (Port the structure, not the source creator's exact posture language.)
  3. Meta-compound - the day after a post pops, post the receipts of what happened, then the lesson. Write it in advance, hold it, fire only when a post actually lands.
  4. Branded series - a short recurring format ("welcome back to X") teaching one small move per installment. Consistent open, one idea, soft CTA.
  5. Industry stance with named receipts - name a bad practice, lay out the slippery-slope anatomy of how it happens, end with the unmissable test or warning.
  6. Personal-stack-reveal - reveal something surprising about how you actually operate, list the components with a proof artifact each, tie the last line back to the opener with a flip.
  7. Comment-bait roll-call - ask a question with absurdly specific qualifiers, invite people to drop themselves in the comments.
  8. Numbered absurd list - a comedic list where one bolded line carries the real teaching and the rest makes it land.
  9. Reframe-pair - "Everyone wants X. Nobody wants Y." Setup creates a universal frame, the reverse flips it, arrow bullets give a portable list, close on a quotable reframe.
  10. Borrowed-authority - peg a credible name's observation, reframe it to create tension, build a small framework, end on a portable line.
  11. Stat-pivot - open with a specific number, pivot with a counter-claim, name the real failure modes, reframe ownership.
  12. Failure-confession - tell a real thing that broke, what you missed, the fix, the lesson. Vulnerability plus a concrete takeaway.

Step 3 - Pull receipts

After hook and archetype are picked, gather 4 to 6 specific receipts to fill the body. Ask the user for them or pull from material they provide. Vary the receipt shapes. Do not repeat the same shape twice in one post. Mix among:

  • A money figure (saved, earned, or replaced)
  • A named, specific result
  • A count or volume number
  • A meeting, pipeline, or audience number
  • A tool or stack mention
  • A recently shipped artifact
  • A mildly self-deprecating real scene from the user's actual day

Step 4 - Voice rules

Banned

  • Wall-of-text paragraphs. One sentence per line is the default. Deviating is a deliberate move.
  • Implementation-layer receipts. Replace "I wrote a script" with what the thing does at the outcome layer.
  • Other creators' signature language. Do not lift recognizable catchphrases.
  • "Not X, not Y" positioning phrasing. Make the positive claim and let it stand.
  • External link CTAs. Link-CTA posts underperform badly. Keep people on-platform.
  • Corporate filler verbs: leverage, synergize, unlock potential, disruption, game-changer.
  • Bland sentence starters: "Let me tell you", "In my experience", "I want to share".
  • Stale hook templates parroted as-is. Use old hooks as voice reference, bring a fresh mechanic from Step 2.

Required

  • Mobile layout (Principle 4). One sentence per line. Short. Heavy whitespace.
  • 4 to 6 specific receipts in the body, varied in shape.
  • One tension line per post.
  • CTA is a comment keyword, a DM keyword, or a single open question. Never a link.
  • Voice markers natural to the user, pulled from Step 1, not invented here.

Step 5 - Output

Mode A - 3 variants

### Draft N - [mechanic] - [length]

[the actual post in the user's mobile layout, copy-pasteable]

Mechanic source: [creator + why it is landing]
Why it transfers: [one sentence]
Receipts used: [the 4-6 specifics]
Tension line: [the friction line]
Tradeoff: [one line]

End: Which do you want to refine? Or remix on a different mechanic?

Mode B - sharpen

Diagnosis:
- Layout: [one sentence per line? short sentences? flag any wall-of-text]
- Hook: [fresh, or stale (parroting old hooks)?]
- Receipts: [count + shape variety. flag vague or implementation-layer]
- Tension line: [present? where?]
- CTA: [working? flag any link]

Sharpened version:
[Edited post in the user's mobile layout. Keep voice and POV. Fix layout if walls of text. Refresh hook if stale. Replace vague receipts with specifics.]

Changes: [each tied to a principle]
Voice deltas: [or "none, voice intact"]
Alternative angle: [pitch only, do not write yet]

Mode C - port

For each ported archetype:

### Port N - [archetype name] (mechanic ported from a top-performing post)

Their mechanic: [one sentence]
Why it transfers: [the structural reason it survives a voice swap]
Receipts used: [the 4-6]

[the actual post in the user's mobile layout]

Step 6 - Refinement

When the user picks one:

  • One tightened version. Cut at least 10 percent unless asked to expand.
  • 1 to 2 alternative opening lines if useful.
  • Never silently rewrite the voice.

Anti-patterns

  • Wall-of-text paragraphs (the number one failure mode).
  • Forcing the user into another creator's register.
  • Lifting mid-tail formats from a target creator just because they are recognizable. Rank first, lift top-quartile only.
  • Parroting the user's own old hooks as the prescribed strategy. Use them as voice reference, bring a fresh mechanic from Step 2.
  • Hard-coded receipts. Pull real ones every run.
  • Implementation-layer receipts.
  • Generic motivational lines.
  • Aggregator posts ("7 tools I'm watching"). Flat to negative.
  • Open-ended philosophical takes without specifics.
  • Posts with no tension line. If every sentence reads as a tasteful statement, the post is half-finished.
  • Any post that could have been written by anyone. If it names no specific result, number, or artifact, it is too vague.

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.