LinkedIn Post from Blog

What this skill does

Reads a blog post and produces a single LinkedIn post: a 2-3 line opener that hooks readers above the "see more" cut, a one-paragraph framing, a 3-5-bullet insight list, and a soft CTA. Output is sized to land within LinkedIn's 1300-character optimal window.

Inputs

• post_path: markdown blog post.
• Optional link: URL to the published post.
• Optional voice: practical (default), narrative, provocative.
• Optional hashtags: array of up to 3 (LinkedIn ranks better with few specific tags).

Steps

1. Read the post. Extract the title, the most concrete sentence from the lede (the one with a number or contrast), and three to five concrete takeaways from the body.
2. Compose the opener (first 210 chars — the see-more cutoff on most clients):
   • Line 1: a hook sentence (claim or surprising stat).
   • Line 2: a one-line stake ("Here's what changed.").
   • Blank line.
3. Compose the framing paragraph: 3-4 sentences narrating the problem and the move toward the takeaways. Use plain language; avoid "leveraging" / "synergy" / corporate filler.
4. Compose the insight bullets: 3-5 lines, each starting with a glyph (-, •, or numbered if sequential). Each bullet is one full sentence; no fragments.
5. Compose the soft CTA: a question for the reader plus the link to the post.
6. Append at most 3 hashtags on the final line (specific, not generic — #PostgresPerformance, not #tech).
7. Validate length: total under 1300 characters; opener under 210.
8. Save to linkedin-<slug>.md.

Output

linkedin-<slug>.md with the post body in a single fenced block (so it's paste-ready) plus a header block listing character counts: opener length, total length. Stdout prints both numbers.

Verification

Paste the output into LinkedIn's compose box (or a simulator) and confirm the see-more cutoff lands after the opener, not mid-sentence. Read the bullets aloud and confirm each is concrete enough to act on (avoid "be intentional", "communicate clearly" — too abstract). Confirm the CTA's link previews to the post (LinkedIn fetches the link's OG image automatically).

Edge cases

• Posts that are mostly tutorials (heavy code): summarize the lesson in the bullets, never paste code; LinkedIn formatting mangles indentation.
• Posts that are interview write-ups: lead with a quote from the interview rather than a stat.
• Posts in a niche where hashtags are conventional (e.g., #dataengineering): include them without exceeding the cap of 3.
• Posts that touch sensitive topics (layoffs, grief): switch voice to narrative, soften CTA, never end with a sales pitch.