How to Write a LinkedIn Post That Is Not a Humblebrag

Anyone about to post a win, a hire, or a lesson can write a LinkedIn post that is not a humblebrag in one sitting: one scene, one fact, one lesson you paid for. Run it in ChatGPT, Claude, or Gemini, then cut every line you would not say to a peer at a table.

A humblebrag is a win wearing a gratitude costume. "Humbled to announce" is a press release with a stoop. You need a shape that forbids theater, plus a stop so the model cannot "punch it up."

If you still paste "write a LinkedIn post about my launch" into a blank chat, read what a skill prompt actually is and save the card.

Write a LinkedIn post that is not a humblebrag

Ban the theater first. Scene second. Lesson last, and only if you paid for it.

1. Ban the humblebrag before you draft

Paste this list and make it non-negotiable:

humbled to announce, I don’t usually share, blessed to, I almost didn’t post this, key takeaway, proud to share, we crushed it, excited to announce used as a whole post.

Then add your own crimes. Instruct Claude, ChatGPT, or Gemini: "If a draft uses a banned phrase, rewrite the whole post. Do not swap in a synonym."

Synonym swaps are how "humbled" becomes "honored." The same ban-first habit that keeps captions from sounding like ads applies to the feed.

2. Feed one scene, not a career arc

Your input is not your bio. It is a moment.

• Where you were, and what broke or worked.
• One fact that would survive a skeptic: a number, a date, a miss, a constraint.
• The cost of the lesson. Time, money, a wrong call. If there was no cost, you do not have a lesson. You have a win. Post the win as a fact, or do not post.

Example: "Thursday standup. Killed the waitlist after 11 days. 40 signups, 3 replies. Lesson: a form is not a conversation." Keep moments in the Promptcrates library of reusable prompts.

3. Force the three-part shape

Every post, same bones:

• Scene — a concrete thing in the world. Time, place, friction. No verdict.
• Fact — one claim you can defend. The win or the miss may live here, once.
• Lesson — one sentence you paid for. It cannot flatter you. It can indict you.

Tell the model: "80 to 160 words. Scene cannot contain the job title. Fact may contain it once. Lesson cannot flatter you or hire anyone. No hashtag block."

4. Generate three, pick one, do not blend

Ask for three variants that share the fact and the lesson, but change only the scene. Pick one. Do not blend. Blended posts have a carousel, not a voice. If ChatGPT hooks with "I almost didn’t share this," refuse it.

5. One human pass at table speed

Read the post at the pace you talk. If you would not say it to one peer, it is still a humblebrag.

Kill:

• The fake almost. You were going to post.
• The fake we, if you were one person.
• The fake lesson that is actually a slogan.

Then stop. If you are hiring, that is a second post, written by you after the skill stops.

Reusable skill card

Trigger: You are about to post a win or a lesson, and the draft already sounds like a press release in a hoodie.

Input: One scene, one skeptic-proof fact, one lesson you paid for or "no lesson, fact only," plus the banned-phrase list.

Output: Three posts, 80–160 words, scene / fact / lesson, title at most once, zero banned phrases.

Stop: After you pick one variant and read it at table speed. No blending. No recruiter paragraph. No "I almost didn’t post this."

Takeaways

• Ban humblebrag phrases before you generate. Synonym swaps still sound like a podium.
• One scene, one fact, one lesson you paid for. A free lesson is a slogan.
• Generate three scenes. Pick one. Never blend.