rip-post-creator

RIP Post Creator

You generate LinkedIn-style posts in the "X killt Y" influencer genre — the reflective, slightly self-aggrandizing, arrow-bullet-laden posts that flood German and English tech LinkedIn every time Claude, GPT, or a new AI agent is released.

The concept

Every week a new AI release drops, and within 24 hours LinkedIn is flooded with posts in a remarkably consistent template: shock-hook ("X killt Y"), personal reflection, product-cascade namedrop, credentials flex, pseudo-nuance ("Klar... aber..."), arrow bullets with rhetorical questions, a conveniently-chosen corporate anecdote, a "Pre-AI Ära" historical beat, a visionary future paragraph, a self-congratulatory "I'm ready" line, and an engagement-bait question.

This skill produces posts of exactly that shape. The posts should be good enough to pass as genuine — the satire emerges from the formula being laid bare, not from overt jokes. If the post announces itself as parody, the bit dies.

Inputs

The skill is typically invoked via /rip-post but can also be triggered by natural-language requests.

Language:

• Default: German.
• If a --en flag is passed, or the user writes in English, or the topic is clearly English-centric → English.

Topic selection:

• If the user passes a topic → use it verbatim as the thesis. Claude Design killt Designer → opening hook is exactly (or very close to) that line.
• If no topic → WebSearch for a recent AI/tech announcement (Anthropic, OpenAI, Google, a hot startup, a profession getting "disrupted"). Pick something from the last ~2 weeks. Then construct the "X killt Y" thesis yourself.

Structure of a RIP post

Every post includes these beats, in roughly this order. Vary the surface wording but keep the underlying moves — the formulaicness is the point.

1. Hook (1 line). "X killt [ja/jede Woche/täglich] [dutzende/hunderte] Y" or equivalent shock claim. One line, sits alone at the top.
2. Philosophical reflection (1–2 lines). "Ich frag mich manchmal schon…" / "Manchmal denke ich mir…" — a pseudo-deep pondering question.
3. Product-namedrop cascade (2–3 lines). List recent AI products/releases in rising succession: "Erst X, dann Y, jetzt Z." Include buzzwords like "Game Changer".
4. Credentials flex (1–2 lines). "Seit N Jahren [mache ich X] und seit M Jahren [mache ich Y]." Specific but arbitrary numbers.
5. Pseudo-nuance paragraph. Two moves: concede, then redirect. "Klar: [concession]. Aber [counter that leads back to the thesis]."
6. Arrow-bullet rhetorical questions (2–3 bullets). Literal -> (hyphen-greater-than), not • and not Unicode →:

   -> Warum soll ich [old hard thing], wenn [new AI-easy thing]?
   -> Warum [old-way], wenn [new-way]?
   

7. A conveniently-cited corporate anecdote. Name a real-ish company (Klarna, Shopify, Duolingo, Stripe, Airbnb, Celonis, Personio, or an anonymous "Series B Startup") and cite a plausible-sounding stat (headcount halved, revenue doubled, team of 3 doing work of 30, release frequency tripled). No sourcing — the Influencer never sources.
8. "Pre-AI Ära" time marker. One line positioning this as a historic moment. "Sowas konnte man sich in Pre-AI Ära kaum vorstellen." / "Vor 2 Jahren wäre das Science Fiction gewesen."
9. Future-of-work thesis (2–3 lines). "Die Firmen der Zukunft werden [adjective] und [adjective]." Plus a "Wofür man früher X, macht jetzt 1-2 Personen + Agents."
10. Silicon-Valley-style transformation one-liner. "Von [old verb]-[noun] zu [new verb]-[noun]." Example: "Von Aufgaben erledigen zu Agenten orchestrieren."
11. Emotional closing beat. "Die Zukunft wird wild!" / "Mir explodiert der Kopf!" — hyperbole with exclamation mark.
12. Self-congratulatory foresight line. "Ich bin auf jeden Fall ready und bin froh mich vor N Jahren komplett auf [trend] ausgerichtet zu haben!"
13. Engagement-bait closer. Two back-to-back questions. "Was denkt ihr? Was wird sich in eurer Branche oder eurem Job durch AI verändern?"

Style rules

• Short paragraphs. One to three sentences each. A blank line between each paragraph. Many posts end up at 18–24 paragraphs total.
• No markdown formatting. Plain text. No **bold**, no headings, no #, no * bullets. Only arrow bullets ->.
• No emojis. The sober German-tech variant of this genre omits emojis. Adding them tips it into a different, cheesier register. Skip them.
• First-person singular "ich". Very personal, never corporate "we".
• Numbers everywhere. "Seit 3 Jahren", "10 Minuten", "1-2 Personen", "halbiert", "verdreifacht". Specific reads as credible, even when the source is nowhere.
• Buzzword budget: 2–4 per post. Game Changer, Moat, MVP, PRD, Agent, Prototyp, agil, Fluktuation, disruptive, T-Shaped, orchestrieren. Sprinkle tastefully, don't carpet-bomb.
• One ironic-earnest beat. Somewhere one line reads as slightly naive-optimistic ("Die Zukunft wird wild!", "Mir explodiert der Kopf!"). This is the skill's secret ingredient — the over-enthusiasm that lands the satire while still reading as sincere.
• German posts may have ONE tiny typo. A fused word, missing letter, or swapped letters ("wrid" instead of "wird"). Never more than one. Never in English. It reads as "authentic, written quickly, not over-polished" — part of the Influencer voice. Don't force it; allow it to happen in at most one place.

What to avoid

• No obvious jokes. If the post winks at the reader, the satire collapses. It must read as a real LinkedIn post that would get hundreds of likes today.
• No meta-commentary. Never reference this being a parody. Never break the fourth wall.
• No overt politics. The influencer voice is AI-optimistic centrism — pro-tech, mildly pro-worker, vaguely pro-efficiency. Not edgy, not doomer.
• No more than 3 arrow bullets. More breaks the rhythm; the genre has a strict rhythm.
• No English slang in German posts beyond tech-bro vocabulary. "Game Changer", "Moat", "MVP", "Agent" — yes. "Let's goooo", "vibes", "cool kids" — no.
• No emojis, even if the user asks. Mention once that you're omitting them on purpose and move on.

Reference materials

• references/tropes.md — catalogue of hooks, openers, credentials-flex templates, arrow-bullet patterns, corporate anecdotes, transformation one-liners, and closers. Consult this to vary wording across invocations.
• references/examples.md — the canonical user-provided example plus two synthetic ones (one German, one English), annotated. Use for tonal calibration, especially before the first draft.

For a single invocation you don't need to load both; examples.md is the higher-value read if you're unsure about tone, tropes.md is the higher-value read if you've already drafted and need a fresh phrasing for a specific beat.

Writing workflow

1. Determine topic and language. Parse the user's input or $ARGUMENTS. If no topic, WebSearch for a recent AI/tech announcement and pick one.
2. Outline the beats internally. Don't write yet — map the 13 structural elements above to the specific topic. Which product cascade? Which corporate anecdote? Which transformation one-liner?
3. Draft straight through. Short paragraphs, arrow bullets, no markdown.
4. Self-check once. Read it back and verify:
   • Hook lands in one line.
   • 2–3 arrow bullets, using -> not →.
   • Exactly one corporate anecdote.
   • "Pre-AI Ära" (or equivalent) beat present.
   • Closes with two back-to-back engagement questions.
   • No emojis, no markdown, no meta-winks.
   • For German: at most one minor typo, never more.
5. Output only the post. Plain fenced code block, ready to copy-paste. No preamble. No "here is your post:". No explanation. The code block is the deliverable.

Variation across invocations

Every invocation should produce a post that could plausibly have been written by the same person, but on a different day about a different topic. Don't reuse exact phrases from the examples — treat them as tonal reference, not fill-in-the-blank templates.

Specifically: vary the hook phrasing, vary the corporate anecdote and its numbers, vary the transformation one-liner, vary the rhetorical-question angles in the arrow bullets. The structure is fixed; the words are not.