Agent skill
carousel-copy
Install this agent skill to your Project
npx add-skill https://github.com/drshailesh88/integrated_content_OS/tree/main/skills/cardiology/carousel-copy
SKILL.md
Carousel Copy Skill
Transform research and knowledge base material into compelling carousel slide copy that sounds like an interventional cardiologist—authoritative, clear, evidence-grounded, human.
When to Use This Skill
Use when you have:
- Research output (PubMed findings, trial data, clinical guidelines)
- Long-form content (newsletters, articles, scripts) to repurpose
- A topic + knowledge base that needs carousel conversion
- Draft slides that feel "flat" or "generic"
The Core Principle
You are not writing "social media content." You are a specialist distilling clinical knowledge for intelligent laypeople.
Every slide should feel like a cardiologist explaining something important to a patient in clinic—direct, clear, evidence-backed, respectful of their intelligence.
Voice Calibration
The Target Zone
TOO CASUAL TARGET TOO FORMAL
─────────────────────────────────●─────────────────────────────────────
"Yo, statins are "Statins reduce heart "HMG-CoA reductase
actually fire for attacks by 25-30%. inhibitors demonstrate
your heart lol" The evidence is clear." significant cardiovascular
risk reduction."
Voice Markers
| Do This | Not This |
|---|---|
| "The evidence shows..." | "Studies have shown..." (passive, generic) |
| "In most patients..." | "Generally speaking..." (hedging) |
| "This matters because..." | "It's important to note that..." (AI tell) |
| "The risk drops by 30%" | "There is a significant reduction" (vague) |
| "Here's what I tell my patients" | "One should consider" (distant) |
| "The data is clear" | "Research suggests" (weak) |
Sentence Rhythm
- Lead with the finding, not the setup
- Short declarative sentences for key points
- Occasional question to the reader (sparingly)
- Never more than 2 sentences of similar length in a row
Example:
FLAT: Studies have shown that walking can be beneficial for heart health,
with research indicating improvements in cardiovascular outcomes.
BETTER: Walking 30 minutes daily cuts heart disease risk by 20%.
You don't need a gym. You need consistency.
Knowledge Base → Slide Copy Transformation
Step 1: Extract the Core Insight
From your research/knowledge base, identify:
- The key finding (what changed, by how much)
- The clinical relevance (why should the reader care)
- The actionable element (what can they do with this)
Step 2: Compress Without Losing Meaning
The 15-Word Discipline
Each slide body should convey ONE idea in ≤15 words. This forces clarity.
| Research Finding | First Draft (too long) | Final Copy (≤15 words) |
|---|---|---|
| "SGLT2 inhibitors demonstrated a 26% relative risk reduction in cardiovascular death or hospitalization for heart failure" | SGLT2 inhibitors reduce the combined risk of cardiovascular death and heart failure hospitalization by 26% | SGLT2 inhibitors cut heart failure hospitalization by 26%. One pill, major impact. |
| "Meta-analysis of 19 RCTs showed statin-associated muscle symptoms occur in 5-10% of patients, with most able to continue therapy" | While 5-10% of patients on statins experience muscle symptoms, the majority can continue their therapy | Muscle pain affects 5-10%. Most can continue statins with adjustments. |
Step 3: Add Clinical Authority
Sprinkle markers that signal "doctor speaking" not "content creator speaking":
Authority Markers (use 1-2 per carousel):
- "In my practice, I see this regularly"
- "The clinical reality is..."
- "What the guidelines miss..."
- "Here's what the data actually shows"
- "This is what I tell my patients"
- "After thousands of cases..."
Evidence Markers (use throughout):
- Specific numbers: "25%", "HR 0.74", "NNT of 20"
- Trial names: "In PARADIGM-HF...", "The EMPEROR-Reduced trial showed..."
- Timeframes: "Over 3.5 years...", "At 12-month follow-up..."
Slide Type Copy Patterns
HOOK Slide (Slide 1)
Purpose: Stop the scroll. Create curiosity gap.
Copy Structure:
[Provocative statement or question]
[Subtext that promises value]
Patterns that work:
| Pattern | Example |
|---|---|
| Counterintuitive fact | "Your LDL might be normal. Your risk isn't." |
| Specific number + surprise | "67% of heart attacks happen with 'normal' cholesterol" |
| Direct question | "Statins causing muscle pain? Here's what to do." |
| Myth setup | "Everything you've heard about salt is wrong" |
| Patient scenario | "Fit, 45, no symptoms. Triple vessel disease." |
Avoid:
- Generic questions: "Want to know about heart health?"
- Clickbait without substance: "Doctors don't want you to know this"
- Fear-only hooks: "You could die from this" (no hope)
TIPS Slide
Purpose: Actionable takeaway.
Copy Structure:
[Tip headline - 5-8 words]
[Supporting detail - 1 sentence]
Example:
Tip #3: Check LDL, not just total cholesterol
Total cholesterol can hide high LDL. Ask for the breakdown.
Compression technique: Start with the action verb.
- "Check your...", "Ask for...", "Avoid...", "Choose...", "Monitor..."
STATS Slide
Purpose: Make data memorable.
Copy Structure:
[BIG NUMBER]
[What it means in plain terms]
[Source - footer or small text]
Example:
26%
Lower risk of heart failure hospitalization
DAPA-HF Trial, NEJM 2019
Where to put the citation:
- Footer (preferred): Clean slide, source visible but not dominant
- In-body: Only when the source itself is the point ("NEJM just published...")
MYTH Slide
Purpose: Correct misconception with evidence.
Copy Structure:
MYTH: [State the misconception exactly as people believe it]
TRUTH: [Direct correction + brief evidence]
Example:
MYTH: "Statins cause permanent muscle damage"
TRUTH: Muscle symptoms affect 5-10%. They're reversible.
Most patients continue therapy successfully.
Key: State the myth in the reader's own words/beliefs, not a strawman.
COMPARISON Slide
Purpose: Clarify choices or show contrast.
Copy Structure:
[Option A] vs [Option B]
[Key differentiator] [Key differentiator]
[When to choose] [When to choose]
Keep parallel: Same structure on both sides.
STORY Slide
Purpose: Human connection, case illustration.
Copy Structure:
[Patient descriptor - age, situation]
[What happened]
[Insight or lesson]
Example:
52-year-old executive. Perfect BMI. Non-smoker.
Came in for a routine checkup.
CAC score: 450. Three-vessel disease.
Risk factors don't tell the whole story.
Privacy note: Keep details generic enough to protect identity while specific enough to feel real.
STEPS Slide
Purpose: Process or sequence.
Copy Structure:
Step [N]: [Action in 3-5 words]
[One sentence of detail]
Example:
Step 2: Get baseline labs
LDL, kidney function, liver enzymes—before starting therapy.
CTA Slide (Final)
Purpose: Clear next action.
Copy Structure:
[Value reminder - what they learned]
[Specific CTA]
@handle
Example:
Heart health is simpler than you think.
Follow for evidence-based cardiology.
@heartdocshailesh
CTA options:
- "Follow for more" (growth)
- "Save this for later" (engagement)
- "Share with someone who needs this" (reach)
- "Questions? Comment below" (engagement)
Archetype Adaptation
Same information, different framing based on who you're writing for.
Example: SGLT2 Inhibitors Reduce Heart Failure Risk
For Heart Patients (Archetype 1):
Hook: "On heart failure meds? This one might not be on your list yet."
Body: SGLT2 inhibitors reduce hospitalization by 26%.
Ask your cardiologist if you're a candidate.
For Biohackers (Archetype 3):
Hook: "SGLT2 inhibitors: Not just for diabetes anymore"
Body: DAPA-HF showed 26% reduction in HF hospitalization
independent of diabetes status. HR 0.74 (95% CI 0.65-0.85).
Mechanism: sodium-glucose cotransporter inhibition +
cardiac remodeling effects.
For Lifestyle Disease Group (Archetype 2):
Hook: "Diabetic? Your heart meds might need an upgrade."
Body: SGLT2 inhibitors do double duty—blood sugar AND
heart protection. 26% fewer hospitalizations in trials.
For Caregivers (Archetype 5):
Hook: "Is your family member on the right heart meds?"
Body: SGLT2 inhibitors are now guideline-recommended for
heart failure. If they're not on one, ask why at the next visit.
Adaptation Quick Reference
| Archetype | Hook Style | Data Depth | Tone | CTA |
|---|---|---|---|---|
| Heart Patients | Direct relevance | Outcome-focused | Reassuring + clear | "Ask your doctor" |
| Lifestyle Disease | Wake-up call | Consequence-focused | Firm but hopeful | "Start here" |
| Biohackers | Novel angle | Deep (HR, CI, mechanism) | Peer-to-peer | "Optimize this" |
| Health Enthusiasts | Myth-buster | Evidence summary | Curious guide | "Share with family" |
| Caregivers | Protective concern | Action-focused | Supportive | "Bring this to appointment" |
| Young & Anxious | Reassurance frame | Context (normal vs. not) | Calm authority | "When to actually worry" |
Citation Placement Rules
Footer placement (default):
- Keeps slide clean
- Use for: Trial names, journal references, guideline citations
- Format:
Trial Name, Journal YearorSource: ACC/AHA Guidelines 2023
In-body placement:
- When the source is the point of the slide
- When establishing authority for controversial claim
- Format: Integrate naturally: "The PARADIGM-HF trial showed..."
Skip citation when:
- Stating clinical common knowledge
- Giving practical advice from experience
- The claim is uncontroversial among cardiologists
Quality Checklist Before Finalizing
- Hook creates genuine curiosity (not clickbait)
- Each slide has ONE clear idea
- Body text ≤15 words per slide
- At least 2 specific numbers/data points in carousel
- At least 1 authority marker ("In my practice...", "What I tell patients...")
- No AI tells ("It's important to note", "In conclusion", "vital role")
- CTA is specific, not generic
- Citations placed for clarity, not clutter
- Reads like a doctor, not a content creator
Common Mistakes to Avoid
| Mistake | Example | Fix |
|---|---|---|
| Hedging everything | "May potentially help reduce..." | "Reduces risk by 25%" |
| Generic claims | "Heart health is important" | "LDL above 100 doubles your risk" |
| Passive voice | "It has been shown that..." | "Three trials showed..." |
| Missing specificity | "Significant improvement" | "26% reduction in hospitalization" |
| Overexplaining | 3 sentences where 1 works | Cut to the core insight |
| Forgetting the human | All data, no "so what" | Add "This means for you..." |
Positioning Markers
These phrases signal "interventional cardiologist" vs. generic health page:
Use these (sparingly, 1-2 per carousel):
- "After thousands of angiograms..."
- "What I see in the cath lab..."
- "The guidelines say X, but clinically..."
- "In my practice..."
- "Before I put in a stent, I always..."
- "Here's what your cardiologist is thinking..."
- "The cardiology community is shifting toward..."
Avoid these (sounds like influencer):
- "Listen up..."
- "This is a game-changer..."
- "Doctors hate this..."
- "The truth they don't tell you..."
- "You need to know this..."
Example Transformation
Input: Knowledge Base Excerpt
DAPA-HF Trial Summary:
- Dapagliflozin vs placebo in HFrEF patients (EF ≤40%)
- Primary endpoint: CV death or worsening HF
- Result: HR 0.74 (95% CI 0.65-0.85), p<0.001
- NNT: 21 over median 18.2 months
- Benefit seen regardless of diabetes status
- Most common side effect: Volume depletion (7.5% vs 6.8%)
Output: Carousel Copy
Slide 1 (HOOK):
A diabetes drug that prevents heart failure
—even if you don't have diabetes
The DAPA-HF story
Slide 2 (STATS):
26%
Reduction in heart failure hospitalization or CV death
HR 0.74 (95% CI 0.65-0.85)
DAPA-HF Trial
Slide 3 (TIPS):
Who benefits?
Heart failure with EF ≤40%
—whether you have diabetes or not.
The effect was consistent across both groups.
Slide 4 (STATS):
NNT: 21
Treat 21 patients for 18 months
to prevent one major event.
That's a strong number.
Slide 5 (MYTH):
MYTH: "It's a diabetes drug—I don't have diabetes"
TRUTH: Dapagliflozin is now FDA-approved for heart failure
independent of diabetes. The heart benefits stand alone.
Slide 6 (TIPS):
What to watch for
Volume depletion: dizziness, lightheadedness
Happens in ~7% of patients. Usually manageable.
Slide 7 (CTA):
SGLT2 inhibitors changed how we treat heart failure.
If you have HFrEF and you're not on one,
ask your cardiologist why.
@heartdocshailesh
Part of Dr. Shailesh Singh's Integrated Content OS
Didn't find tool you were looking for?