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
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
pufferlib
This skill should be used when working with reinforcement learning tasks including high-performance RL training, custom environment development, vectorized parallel simulation, multi-agent systems, or integration with existing RL environments (Gymnasium, PettingZoo, Atari, Procgen, etc.). Use this skill for implementing PPO training, creating PufferEnv environments, optimizing RL performance, or developing policies with CNNs/LSTMs.
fluidsim
Framework for computational fluid dynamics simulations using Python. Use when running fluid dynamics simulations including Navier-Stokes equations (2D/3D), shallow water equations, stratified flows, or when analyzing turbulence, vortex dynamics, or geophysical flows. Provides pseudospectral methods with FFT, HPC support, and comprehensive output analysis.
metabolomics-workbench-database
Access NIH Metabolomics Workbench via REST API (4,200+ studies). Query metabolites, RefMet nomenclature, MS/NMR data, m/z searches, study metadata, for metabolomics and biomarker discovery.
geniml
This skill should be used when working with genomic interval data (BED files) for machine learning tasks. Use for training region embeddings (Region2Vec, BEDspace), single-cell ATAC-seq analysis (scEmbed), building consensus peaks (universes), or any ML-based analysis of genomic regions. Applies to BED file collections, scATAC-seq data, chromatin accessibility datasets, and region-based genomic feature learning.
zinc-database
Access ZINC (230M+ purchasable compounds). Search by ZINC ID/SMILES, similarity searches, 3D-ready structures for docking, analog discovery, for virtual screening and drug discovery.
astropy
Comprehensive Python library for astronomy and astrophysics. This skill should be used when working with astronomical data including celestial coordinates, physical units, FITS files, cosmological calculations, time systems, tables, world coordinate systems (WCS), and astronomical data analysis. Use when tasks involve coordinate transformations, unit conversions, FITS file manipulation, cosmological distance calculations, time scale conversions, or astronomical data processing.
Didn't find tool you were looking for?