pinned-comment exists to make a particular kind of LinkedIn humour repeatable. The pinned comment is where Charlie Hills drops the polished creator mask and lets a small absurdity carry the personality of the post. The skill is built around one rule that holds the format together, which is that the image carries the joke and the comment captions the image as if reporting the news.
The pinned comment is the most underused trust-building surface on LinkedIn. The post itself does the value-delivery work. The pinned comment does the personality work, which is what makes the writer feel like a person rather than a content account. pinned-comment is the skill that codifies one specific approach to that personality: a deadpan image-led format that the team can reproduce without losing the voice.
§01What it does
The skill works from a published or draft post and pulls a single quiet admission from it. Cowork does most of my actual job now. I am embarrassingly dependent on Anthropic. I gave away a 9-month product for free. If the post does not contain an admission the skill can name in one sentence, the skill stops; the post is not pinned-comment material yet.
Then it builds the image prompt before writing any text. Three image rules: one clear visual gag the eye lands on in under a second, a completely straight expression with no winking or thumbs-up, and the writer in the lower-status position relative to whatever is in the frame (Claude wins, the logo wins, the mum wins).
Only after the image prompt is built does the skill write the four-line comment. Each line is one complete sentence, forty characters maximum, with the pin emoji on line one. Line one names the absurd thing as fact. Line two flips the writer's status downward. Line three is a sad flex, the smallest possible win. Line four is resigned acceptance with no punchline reach. Then five tests run before the output ships: image gag describable in five words, line one captions the image, writer is the loser in every line, line four does not strain for cleverness, and the four lines are boring without the image attached.
§02The boring-on-its-own test is the trick
The instinct when writing a comment that is supposed to be funny is to make the words funny. This skill inverts that. If the four lines are funny without the image, the format is broken; the image is meant to do the heavy lifting and the comment is meant to read as deadpan reporting on what the image shows. A comment that lands without its image is a comment that is doing too much, which means the image is doing too little.
That asymmetry is what makes the format reproducible by a team rather than dependent on one specific writer's joke instincts. The comment-writing job becomes mechanical (caption the image as fact, hit four lines, hit the four roles), and the personality lives in the image gag, which is easier to brief than humour in prose is.
§03Setup
# Trigger phrases:
# "write me a pinned comment"
# "pin comment for this post"
# "first comment for the post"
The image prompt is built in the standard use the person in the attached reference image format, which means the writer needs a reference photo on hand (the same one used for youtube-thumbnail works fine). Generate in Gemini with Create Image enabled.
§04Caveats
The format only works on posts that contain an admission worth flipping. A purely informational post, a how-to, or a celebration does not have the right material; running the skill on those will either produce a stretched joke or stop with a not pinned-comment material yet message. That message is correct. Most posts are not.
The status-asymmetry rule (writer always loses) is non-negotiable. Letting the writer win in any of the four lines collapses the format from charming to boastful instantly. The five-test pass exists for that reason.