x-post-creator-skill（x 帖子创建者技能）

X Post Creator for Cardiology Thought Leadership

Generate batches of 10 scientifically accurate, engaging X posts that drive shares and followers.

Core Workflow

1. Load feedback → Read references/feedback-log.md for accumulated learnings
2. Select strategic combinations → Use combination engine below
3. Verify scientific accuracy → Every claim must be defensible in peer review
4. Apply writing frameworks → Use frameworks from references/copywriting-frameworks.md
5. Quality check → Run checklist before output
6. Output batch → Present 10 posts with metadata
7. Collect feedback → Update feedback log

Reference Files

Scientific Accuracy (NON-NEGOTIABLE)

This is directly associated with the cardiologist's reputation. Good content may or may not help career. Bad science WILL doom it.

Requirements:

• State ONLY what peer-reviewed evidence supports
• Use appropriate hedging: "research suggests," "studies show," "evidence indicates"
• Never overstate benefits or understate risks
• Include mechanism when possible (builds credibility)
• When uncertain, flag for verification rather than guess
• Cite study types when relevant: RCT, meta-analysis, cohort

Never produce:

• Unsubstantiated claims or "miracle cures"
• Cherry-picked data without context
• Fear-mongering without solutions
• Advice contradicting clinical guidelines without justification

Combination Engine

Each post uses: Seed(s) + Modifier(s) + Audience + Awareness Level + Framework = Unique Post

Variety requirements per batch of 10:

• Minimum 5 different seed categories
• Minimum 4 different modifier categories
• All 5 audiences represented at least once across batch
• Mix of 4A frameworks (Actionable, Analytical, Aspirational, Anthropological)
• At least 2 different Magical Multiplier angles
• No repetitive openings (vary first 3 words)

Output Format

[1] {post text}
---
Seeds: {seeds used}
Modifiers: {modifiers used}
Audience: {primary audience}
Awareness: {level}
Framework: {4A type} + {Multiplier if used}
Chars: {count}/280

[2] {post text}
...

Continue to [10].

Feedback Integration Protocol

After output, ask: "Any feedback on this batch? Rate 1-5 and note what worked/didn't."

When feedback received:

1. Acknowledge specific change needed
2. Append to references/feedback-log.md with date
3. Apply immediately to all future generations
4. Confirm understanding back to user

Quality Checklist (Run Before Every Output)

For EACH post verify:

• Scientifically accurate (defensible in peer review)
• No AI-typical phrases (see writing-rules.md)
• No em dashes
• Under 280 characters
• Engaging hook in first line
• Clear value to specific audience
• Would not embarrass cardiologist professionally
• Different opening from other posts in batch
• Framework applied correctly