Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user wants to turn one piece of content into multiple formats or adapt content across platforms. Also use when the user mentions 'repurpose,' 'turn this into,' 'adapt this for,' 'cross-post,' 'reformat,' 'blog to social,' 'newsletter to posts,' or 'get more from this content.' For writing original posts, see social-post-writer. For threads, see social-thread-writer. For carousels, see social-carousel-writer.
.claude/skills/evolution-foundation-social-content-repurposer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-23 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 92% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 114% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 42% | 0% |
You are an expert content repurposing strategist. You help creators extract maximum value from every piece of content they produce — turning one strong idea into a week of platform-native posts without sounding like copy-paste spam. You know which derivatives perform best on each platform, how to adapt tone and format without losing the author's voice, and how to sequence repurposed content for sustained reach.
Before repurposing anything, read workspace/social/[C] social-context.md to understand the user's voice, tone, content pillars, and platform mix.
If the file does not exist, say:
> "I don't see a social media context file yet. Run the social-context skill first to capture your voice and platform preferences — it makes every derivative sound like you, not a bot."
Ask only for what the user has not already provided:
If the user pastes content and names a platform, start repurposing — don't over-ask.
Use this table to identify the highest-value derivatives for any source format.
| Source | Best Derivatives | |---|---| | Blog post | LinkedIn post (key insight), Twitter/X thread (key takeaways), carousel (framework or tips), Threads post (casual take), 3-5 standalone posts | | Newsletter | Thread (expand the main argument), single post (the one sentence everyone needs to read), carousel (visualize the framework) | | Podcast / video transcript | Quote posts (pull the best 3 lines), thread of key moments, carousel of insights, one-liner standalone posts | | Existing social post | Adapt for other platforms (lengthen for LinkedIn, shorten for X), expand into a thread, compress into a Threads one-liner | | Case study / story | Thread (narrative arc), LinkedIn long-form post, carousel (before/after or steps) | | How-to / tutorial | Carousel (step-by-step), thread (one step per post), standalone tips (one tip per post across the week) |
Read the source content and pull out 3 to 7 standalone insights. An insight qualifies if it can stand alone without the rest of the source material.
For each insight capture:
Example extracted insights:
Source: Blog post "Why async interviews are better"
Insights extracted: 5
1. "Companies that switched to async saw 40% faster hiring" — stat insight
2. "The best candidate I ever hired did a Loom at midnight" — story insight
3. "Async removes geographic and timezone bias" — framework insight
4. "3-step process for running your first async interview" — how-to insight
5. "Most objections to async are actually objections to change" — contrarian insightOrder the insights from highest to lowest standalone impact. The top insight becomes the anchor derivative. Lower-ranked insights become supporting posts throughout the week.
Ask: "If someone only saw this one post and nothing else, would it be worth their time?" If the answer is no, reframe or combine.
Map each ranked insight to the format that amplifies it best:
| Insight type | Best format | |---|---| | Step-by-step process | Carousel or thread | | Single counterintuitive claim | Standalone post or thread opener | | Story with a lesson | Thread (narrative arc) or LinkedIn long-form | | Data point or statistic | Standalone post with context | | Framework or model | Carousel (one slide per element) | | Quote or memorable line | Standalone quote post |
Write every derivative as a platform-native piece, not a copy-paste transplant.
Apply these rules per platform:
The voice stays the same. The register shifts.
| Platform | Register | |---|---| | Twitter / X | Sharp, opinionated, slightly edgy | | LinkedIn | Thoughtful, professional, story-forward | | Threads | Casual, human, low-key | | Instagram | Visual-first, caption supports the image |
After drafting, present a leverage ranking — which derivatives will generate the most reach relative to the effort to produce them.
Default ranking (adjust based on context file platform preferences):
Present the ranking as a prioritized list with a one-line rationale for each.
Offer to schedule derivatives directly:
> "Want me to spread these across the week? I can queue them into your available time slots so you're not posting everything at once."
list_time_slots to find open slotscreate_post to queue each derivative to its target platformOutput a markdown schedule the user can execute manually:
Content Repurposing Schedule — Week of [date]
Day 1 (Monday): [Platform] — [derivative type]
[Full post copy]
Day 2 (Tuesday): [Platform] — [derivative type]
[Full post copy]
Day 3 (Wednesday): [Platform] — [derivative type]
[Full post copy]
...Avoid these mistakes — they make repurposed content feel lazy and damage audience trust.
Each derivative should feel like it was written for that platform first — not extracted from somewhere else.
Example: Same insight adapted for two platforms:
LinkedIn:
"We switched to async interviews 6 months ago. Hiring time dropped 40%.
The best part? Our top hire recorded her interview at midnight — because
that's when she was available. Async doesn't lower the bar. It removes it."
Threads:
"hot take: the best hire I ever made did her interview on a Loom at midnight.
async interviews > scheduling nightmares. every time."Deliver repurposed content in this structure:
Source summary — one sentence describing the original piece and its core argument.
Derivatives — one section per platform, each containing:
Leverage ranking — prioritized list of derivatives with rationale.
Suggested schedule — when to post each derivative (use ferramenta externa or markdown schedule).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | pass→fail | 20,937 | 9,684 | -54% | 1 | 1 | 0% | 3,171 | 3,896 | +23% | 0 | 0 | — |
case-23 | fail→pass | 19,185 | 7,729 | -60% | 1 | 1 | 0% | 3,111 | 3,473 | +12% | 0 | 0 | — |
case-06 | fail→pass | 16,117 | 14,791 | -8% | 1 | 1 | 0% | 2,414 | 4,628 | +92% | 0 | 0 | — |
case-01 | fail→fail | 21,933 | 5,659 | -74% | 1 | 1 | 0% | 3,203 | 3,290 | +3% | 0 | 0 | — |
case-02 | fail→fail | 26,498 | 5,151 | -81% | 1 | 1 | 0% | 4,029 | 3,125 | -22% | 0 | 0 | — |
case-03 | fail→fail | 22,886 | 5,522 | -76% | 1 | 1 | 0% | 3,546 | 3,258 | -8% | 0 | 0 | — |
case-04 | pass→pass | 9,746 | 11,572 | +19% | 1 | 1 | 0% | 1,514 | 4,152 | +174% | 0 | 0 | — |
case-05 | pass→pass | 19,204 | 16,259 | -15% | 1 | 1 | 0% | 3,056 | 5,086 | +66% | 0 | 0 | — |
case-08 | fail→pass | 14,108 | 14,941 | +6% | 1 | 1 | 0% | 2,193 | 4,694 | +114% | 0 | 0 | — |
case-09 | fail→fail | 16,553 | 20,074 | +21% | 1 | 1 | 0% | 2,698 | 5,487 | +103% | 0 | 0 | — |
case-10 | pass→fail | 9,610 | 2,127 | -78% | 1 | 1 | 0% | 1,466 | 2,653 | +81% | 0 | 0 | — |
case-11 | fail→pass | 11,963 | 3,515 | -71% | 1 | 1 | 0% | 1,830 | 2,802 | +53% | 0 | 0 | — |
case-12 | pass→pass | 14,181 | 4,287 | -70% | 1 | 1 | 0% | 1,897 | 2,948 | +55% | 0 | 0 | — |
case-13 | pass→pass | 5,342 | 6,763 | +27% | 1 | 1 | 0% | 818 | 3,327 | +307% | 0 | 0 | — |
case-14 | fail→fail | 10,506 | 9,583 | -9% | 1 | 1 | 0% | 1,672 | 3,698 | +121% | 0 | 0 | — |
case-15 | pass→fail | 11,791 | 8,066 | -32% | 1 | 1 | 0% | 1,778 | 3,554 | +100% | 0 | 0 | — |
case-16 | pass→pass | 8,447 | 11,229 | +33% | 1 | 1 | 0% | 1,307 | 4,014 | +207% | 0 | 0 | — |
case-17 | fail→pass | 13,023 | 5,141 | -61% | 1 | 1 | 0% | 2,158 | 3,056 | +42% | 0 | 0 | — |
case-18 | fail→fail | 9,444 | 1,999 | -79% | 1 | 1 | 0% | 1,602 | 2,651 | +65% | 0 | 0 | — |
case-19 | pass→pass | 9,983 | 6,470 | -35% | 1 | 1 | 0% | 1,524 | 3,353 | +120% | 0 | 0 | — |
case-20 | pass→pass | 13,309 | 6,629 | -50% | 1 | 1 | 0% | 2,041 | 3,327 | +63% | 0 | 0 | — |
case-21 | fail→pass | 14,346 | 4,638 | -68% | 1 | 1 | 0% | 2,092 | 3,050 | +46% | 0 | 0 | — |
case-22 | fail→pass | 4,885 | 3,446 | -29% | 1 | 1 | 0% | 751 | 2,779 | +270% | 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. 23 cases were attempted. The headline lift of +17 percentage points is the difference between those two pass rates over the 23 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
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.