Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user wants to create LinkedIn post copy or optimize for LinkedIn. Also use when the user mentions "LinkedIn post," "LinkedIn article," "professional post," "post to LinkedIn," "LinkedIn content," "LinkedIn copy," "B2B LinkedIn," "LinkedIn engagement," "LinkedIn feed," "share box," "document post," "poll," "Newsletter," "reshare," or "LinkedIn marketing." For LinkedIn ads, use linkedin-ads.
.claude/skills/mkurman-linkedin-posts/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 56% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 127% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 136% | 0% |
|----------------| | Start a post | Short update; can include link preview if you paste a URL. Same feed format as other updates. | | Photo | Single or multiple images (carousel in feed). | | Video | Uploaded file (distinct from LinkedIn Live, which is live streaming and has separate gating). | | Write article | Article = long-form editor, separate from the short post box; long URL, better for depth and some off-site discoverability. | | Document | PDF / PPT / DOC (slides in feed). Official limits (check current help): on the order of ~100MB / ~300 pages per file—verify when publishing. | | Poll | Engagement driver; keep question and options scannable. | | More (menu) | Often includes celebrations, hiring-style share, Find an expert, etc. (varies by product/region). | | Reshare | Reshare or quote another member’s post with your take—adds context; avoid empty reshares. | | Newsletter | Series subscription; not the same as a one-off post but compound reach over time. | | Event | Create/promote events via a dedicated flow, not the same as a plain text post. |
Product detail: Get started with posting on LinkedIn · Upload and share documents
Why it matters for copy: Match CTA and length to the form (e.g. a document deck vs a 5-line hot take). Do not treat a short post and an Article as interchangeable.
Writing implications: Strong first line and on-topic depth; comments that add substance; avoid templates that look automated or “pod” coordinated.
| Type | Limit | Notes | |------|-------|-------| | Post | 3,000 characters | Optimal: 1,300–1,600 | | First line (critical) | 210–235 chars | Visible before "See more"; 60–80% decide here | | Short posts | 100–200 chars | Polls, announcements, quotes |
| Type | Characters | Use | |------|------------|-----| | Short | 100–200 | Polls, announcements, quotes | | Medium | 300–1,200 | Case studies, tips, BTS | | Long | 1,200–2,000 | Thought leadership, analysis | | Sweet spot | 1,300–1,600 | Highest engagement | | Avoid | >2,000 | ~35% engagement drop |
| Format | Dimensions | Use | |--------|------------|-----| | Single image | 1200×627 (1.91:1) | Feed; link previews | | Square | 1200×1200 | Single image | | Carousel (organic) | Up to 20 images | Multi-image post | | File | ≤10 MB; JPG/PNG | Native uploads perform better | | Vertical | Preferred | 88% browse on mobile |
When generating LinkedIn copy, provide:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 12,251 | 4,627 | -62% | 1 | 1 | 0% | 1,790 | 2,446 | +37% | 0 | 0 | — |
case-01 | fail→pass | 14,836 | 11,862 | -20% | 1 | 1 | 0% | 2,303 | 3,593 | +56% | 0 | 0 | — |
case-02 | fail→pass | 13,989 | 8,666 | -38% | 1 | 1 | 0% | 2,082 | 3,068 | +47% | 0 | 0 | — |
case-03 | pass→pass | 6,327 | 5,899 | -7% | 1 | 1 | 0% | 1,101 | 2,766 | +151% | 0 | 0 | — |
case-04 | fail→fail | 7,456 | 6,809 | -9% | 1 | 1 | 0% | 1,344 | 2,882 | +114% | 0 | 0 | — |
case-05 | fail→pass | 5,406 | 2,311 | -57% | 1 | 1 | 0% | 940 | 2,132 | +127% | 0 | 0 | — |
case-07 | fail→pass | 14,391 | 9,599 | -33% | 1 | 1 | 0% | 2,078 | 3,195 | +54% | 0 | 0 | — |
case-08 | fail→pass | 11,178 | 12,686 | +13% | 1 | 1 | 0% | 1,622 | 3,834 | +136% | 0 | 0 | — |
case-09 | pass→pass | 10,753 | 10,064 | -6% | 1 | 1 | 0% | 1,753 | 3,331 | +90% | 0 | 0 | — |
case-10 | pass→pass | 13,836 | 12,848 | -7% | 1 | 1 | 0% | 2,025 | 3,568 | +76% | 0 | 0 | — |
case-11 | pass→pass | 9,734 | 9,402 | -3% | 1 | 1 | 0% | 1,561 | 3,208 | +106% | 0 | 0 | — |
case-12 | fail→pass | 10,777 | 11,643 | +8% | 1 | 1 | 0% | 1,593 | 3,488 | +119% | 0 | 0 | — |
case-13 | pass→pass | 18,075 | 13,166 | -27% | 1 | 1 | 0% | 2,675 | 3,731 | +39% | 0 | 0 | — |
case-14 | fail→pass | 6,305 | 2,784 | -56% | 1 | 1 | 0% | 978 | 2,135 | +118% | 0 | 0 | — |
case-15 | fail→pass | 6,002 | 5,985 | -0% | 1 | 1 | 0% | 901 | 2,771 | +208% | 0 | 0 | — |
case-16 | pass→pass | 9,593 | 6,830 | -29% | 1 | 1 | 0% | 1,489 | 2,789 | +87% | 0 | 0 | — |
case-17 | fail→pass | 12,129 | 15,383 | +27% | 1 | 1 | 0% | 1,768 | 3,952 | +124% | 0 | 0 | — |
case-18 | fail→fail | 17,348 | 13,503 | -22% | 1 | 1 | 0% | 2,455 | 3,740 | +52% | 0 | 0 | — |
case-19 | pass→pass | 8,897 | 4,424 | -50% | 1 | 1 | 0% | 1,414 | 2,466 | +74% | 0 | 0 | — |
case-20 | pass→pass | 15,325 | 15,546 | +1% | 1 | 1 | 0% | 2,292 | 4,074 | +78% | 0 | 0 | — |
case-21 | pass→pass | 6,186 | 6,793 | +10% | 1 | 1 | 0% | 1,104 | 3,116 | +182% | 0 | 0 | — |
case-22 | pass→pass | 41,728 | 40,592 | -3% | 1 | 1 | 0% | 6,185 | 7,573 | +22% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +41 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.