Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user wants to create X (Twitter) post copy, threads, or optimize for X platform. Also use when the user mentions "X post," "X thread," "Twitter post," "Twitter thread," "tweet," "tweet copy," "thread," "X marketing," "X content," "post to X," "create X post," or "X post copy." For long-form source, use article-content.
.claude/skills/mkurman-twitter-x-posts/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 66% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 60% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 127% | 0% |
|-------|-------| | Standard post | 280 characters | Most users | | Premium | 25,000 characters | X Premium subscribers | | URL | Counts as 23 chars | t.co shortens all links | | Emoji | ~2 chars each | Varies by emoji |
| Use Case | Characters | Purpose | |----------|------------|---------| | General | 71-100 | Sweet spot for engagement | | Promotional | 120-130 | Product/offer posts | | Question | 60-80 | Drive replies | | Retweet-friendly | ~116 | Leaves room for "RT @user:" |
| Part | Ratio | Guideline | |------|-------|-----------| | Hook | 10% | First 1-2 lines; question, fact, or emotion; ~50 chars | | Value | 70% | Practical info; use thread for depth | | CTA | 20% | Call for reply, question, or action |
End with open question to drive replies. Thread structure can extend impressions ~10x.
| Signal | Impact | |--------|--------| | Replies | Highest (~54-75x likes); quality > quantity | | Author replies | Active reply chains ~75x visibility | | Bookmark | 2026 signal; saves boost recommendation | | Media | Images/video ~2x; video 2-4x exposure | | External links | Reduce score ~50%; prefer internal refs | | Post limit | 5-8/day; >10/day reduces later visibility ~80% |
Content: Open questions, controversial views, stories; avoid "RT if agree" bait.
| Threshold | Effect | |-----------|--------| | < 65 | Only last 3 posts enter For You | | New accounts | Need +17 to appear in feed | | Blue V | Auto 100 points |
Monitor account health; avoid spam, bulk ops, excessive links.
Posts tested in small group first; early interaction (first 30 min) decides reach. Niche consistency helps SimClusters match.
| Practice | Guideline | |----------|-----------| | Penalty | External links ~-50%; Regular accounts: link posts 0% engagement since 2025 | | Placement | Put link in reply, not main post | | Ratio | Max 1 link per 5 posts | | Premium | Premium+ safest for links (~0.25-0.3% engagement) | | Preview | Title ≤70 chars, description ≤200 chars; 1200×628 image |
| Tier | Reach | Link posts | |------|-------|------------| | Regular | <100/post | 0% engagement | | Premium ($8) | ~600/post | ~0.25-0.3% | | Premium+ ($40) | ~1550/post | Safest for links |
Premium ~10x reach vs Regular. If X is core channel, Premium helps.
| Type | Share | Use | |------|-------|-----| | Value | 70% | Education, how-to, insights | | Interaction | 20% | Polls, questions, AMA | | Promo | 10% | Product, offers |
Cards (og:image, title, description) boost CTR ~64%, engagement ~26%. Use for Social SEO; X traffic can accelerate Google indexing. For programmatic SEO (template + data pages at scale), use programmatic-seo.
| Format | Dimensions | Use | |-------|------------|-----| | Single image | 1200×675 (16:9) | Best visibility; no crop | | Square | 800×800 | Single image | | Profile | 400×400 | Avatar | | Header | 1500×500 (3:1) | Banner | | File | ~5MB; JPG/PNG | Over 5MB may compress |
When generating X copy, provide:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 14,517 | 11,196 | -23% | 1 | 1 | 0% | 2,258 | 3,188 | +41% | 0 | 0 | — |
case-21 | pass→pass | 6,164 | 2,587 | -58% | 1 | 1 | 0% | 984 | 1,775 | +80% | 0 | 0 | — |
case-01 | fail→fail | 11,327 | 11,902 | +5% | 1 | 1 | 0% | 1,868 | 3,503 | +88% | 0 | 0 | — |
case-02 | fail→fail | 17,438 | 16,926 | -3% | 1 | 1 | 0% | 3,036 | 4,151 | +37% | 0 | 0 | — |
case-04 | pass→pass | 14,435 | 13,481 | -7% | 1 | 1 | 0% | 2,113 | 3,397 | +61% | 0 | 0 | — |
case-05 | pass→pass | 4,848 | 8,523 | +76% | 1 | 1 | 0% | 757 | 2,975 | +293% | 0 | 0 | — |
case-06 | fail→pass | 19,974 | 22,539 | +13% | 1 | 1 | 0% | 3,181 | 5,285 | +66% | 0 | 0 | — |
case-07 | fail→pass | 9,728 | 5,356 | -45% | 1 | 1 | 0% | 1,399 | 2,223 | +59% | 0 | 0 | — |
case-08 | fail→pass | 9,447 | 5,176 | -45% | 1 | 1 | 0% | 1,352 | 2,165 | +60% | 0 | 0 | — |
case-09 | fail→pass | 9,152 | 4,015 | -56% | 1 | 1 | 0% | 1,338 | 1,919 | +43% | 0 | 0 | — |
case-10 | pass→pass | 13,354 | 8,421 | -37% | 1 | 1 | 0% | 2,088 | 2,618 | +25% | 0 | 0 | — |
case-11 | pass→pass | 8,244 | 4,302 | -48% | 1 | 1 | 0% | 1,243 | 2,060 | +66% | 0 | 0 | — |
case-12 | fail→pass | 4,961 | 3,667 | -26% | 1 | 1 | 0% | 835 | 1,893 | +127% | 0 | 0 | — |
case-13 | fail→pass | 11,176 | 6,481 | -42% | 1 | 1 | 0% | 1,698 | 2,329 | +37% | 0 | 0 | — |
case-14 | fail→pass | 14,374 | 7,743 | -46% | 1 | 1 | 0% | 2,210 | 2,654 | +20% | 0 | 0 | — |
case-15 | pass→pass | 9,047 | 7,356 | -19% | 1 | 1 | 0% | 1,436 | 2,528 | +76% | 0 | 0 | — |
case-16 | fail→pass | 14,639 | 5,165 | -65% | 1 | 1 | 0% | 2,026 | 2,133 | +5% | 0 | 0 | — |
case-17 | fail→pass | 13,345 | 4,805 | -64% | 1 | 1 | 0% | 1,928 | 2,030 | +5% | 0 | 0 | — |
case-18 | fail→pass | 14,925 | 8,264 | -45% | 1 | 1 | 0% | 2,272 | 2,734 | +20% | 0 | 0 | — |
case-19 | pass→pass | 10,389 | 2,893 | -72% | 1 | 1 | 0% | 1,621 | 1,855 | +14% | 0 | 0 | — |
case-20 | pass→pass | 13,492 | 6,523 | -52% | 1 | 1 | 0% | 1,987 | 2,314 | +16% | 0 | 0 | — |
case-22 | pass→pass | 9,614 | 2,999 | -69% | 1 | 1 | 0% | 1,623 | 1,852 | +14% | 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 +45 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.