Create Viral Social Media Content1 skill
Refine raw posts into content that stops scrollers and generates shares across all major platforms.
Content Blueprint
Safety checkedA single skill that refines posts for maximum engagement across Reddit, X, LinkedIn, YouTube, TikTok, and email.
Run this command to deploy the blueprint to your environment.
Turns raw drafts into content that spreads by applying deliberative refinement loops, hook architecture, and platform-specific optimization. This removes the guesswork from why some posts generate shares and action while others disappear into feeds.
Once set up, you provide a topic, platform, and target audience. The skill applies the deliberative refinement loop: generate initial draft, attack from six audience perspectives (Skeptic, Expert, Scroller, Competitor, Editor, Thumbnail-specialist), identify and eliminate 100+ AI tells, refine with human voice, and iterate until the content is unbreakable. You get optimized social media content tuned to your platform with hooks, title formulas, and thumbnail guidance built in. Optional: pipe output to humanize-writing for final voice polish.
Refine raw posts into content that stops scrollers and generates shares across all major platforms.
Posts that generate shares, comments, and action because they hit the psychology of stops-scrolling, creates curiosity or stakes, and sounds human instead of AI-written.
Integrates with humanize-writing skill to remove AI vocabulary tells and ensure natural sentence rhythm while preserving viral hooks. If humanize-writing is unavailable, the skill includes a manual humanization checklist (read-aloud test, transition audit, enthusiasm check, specificity check, length check). Works with Claude.ai, Claude Code, and Claude API.
Optional dependency on humanize-writing skill for final voice polish; built-in humanization checklist available as fallback. Skill assumes you are polishing your own ideas, not fabricating content from scratch. Intended for genuine value propositions, not astroturfing or misinformation.
Hook Architecture
Three proven patterns (Prediction+Stakes, Tribal Identity Split, Before/After Compression) that stop scrollers in the first 2 seconds
AI Tell Detection
100+ catalogued AI tells with human alternatives across text, thumbnails, and video scripts with platform-specific variants
Platform Optimization
Reddit subreddit calibration, Twitter thread structure, LinkedIn professional voice, YouTube comments, TikTok hooks, email subject lines with research-backed formulas
Refinement Protocol
Six adversarial passes (Skeptic, Expert, Scroller, Competitor, Editor, Thumbnail-specialist) to attack your draft from different perspectives before shipping
Research Basis
40 documented sources including BuzzSumo (100M headlines), Netflix thumbnail personalization, Outbrain negative superlatives study, face psychology CTR data