elonmusk-repost
npx skills add https://github.com/mingyouagi/elonmusk-repost-skill --skill elonmusk-repost
Agent 安装分布
Skill 文档
Elon Musk Repost Skill
Analyze what Elon Musk reposts on X and generate tweets optimized for his engagement.
How to Use
When invoked, follow this workflow:
- Check prerequisites – Verify account readiness
- Gather context – Ask the user about their product/content
- Analyze fit – Score how well it matches Musk’s repost patterns
- Generate tweets – Create optimized tweet drafts
- Plan post-publish – Give a follow-up playbook
Prerequisites (Account-Level)
Before optimizing tweet content, confirm the basics:
- Profile looks credible (real bio, avatar, some post history)
- At least a few hundred organic followers
- Account is not brand new (>30 days old)
- No recent spam/suspension flags
A zero-follower egg-avatar account will be ignored regardless of content quality.
Step 1: Gather Product/Content Info
Ask the user:
- What is the product/project name?
- What does it do? (one sentence)
- Is there a demo URL or video?
- What’s the visual hook? (screenshot, GIF, video)
Step 2: Musk Repost Pattern Analysis
Core Repost Triggers (Score each 0-10)
Evaluate the user’s content against these proven patterns:
| Pattern | Weight | Description |
|---|---|---|
| AI Visual Demo | 10 | Visually stunning AI-generated output (images, video, 3D, games) |
| AI Agent | 9 | Autonomous AI that completes complex tasks end-to-end |
| “Wow Factor” | 9 | Instantly impressive, no explanation needed |
| Robotics / Physical AI | 9 | Humanoid robots, embodied AI, real-world autonomous systems |
| Video / Multimodal Generation | 8 | AI-generated video, voice, music, or cross-modal output |
| Playable/Interactive | 8 | Users can try it immediately (game, tool, demo) |
| Grok/xAI Adjacent | 8 | Content that relates to or complements Grok capabilities |
| Open-Source AI | 7 | Open-weight models, open-source tools (Musk pushed Grok open-source) |
| Underdog / Unexpected Origin | 7 | Surprising team origin â small team, non-obvious country, solo dev |
| Creator Tool | 7 | Empowers individuals to create professional-grade output |
| Speed/Scale | 6 | “X million users in Y hours” type virality metrics |
| Government Efficiency / DOGE | 6 | Tools that cut bureaucratic waste or align with efficiency themes |
| Meme Potential | 6 | Content that can be remixed into memes |
| Anti-Establishment | 5 | Challenges incumbents (Adobe, Google, OpenAI, etc.) |
| Space / Mars | 5 | Space exploration, Mars colonization, rocket tech |
Scoring Formula
For each pattern the content matches:
match_score(i) = relevance(0-10) Ã weight(i)
Repost Score = Σ match_score(i) / Σ weight(i) for ALL patterns à 100
relevance= how strongly the content matches that pattern (0 = no match, 10 = perfect match)weight= the number in the table above- Denominator uses ALL 15 patterns (sum of all weights = 108), so a perfect score requires hitting every pattern at 10/10.
- In practice, 70+ is strong; 85+ is exceptional.
Successful Case Studies
Lovart (AI Design Agent) â Score: 81/100
- AI Agent: 9/10 Ã 9 = 81 â End-to-end design agent
- AI Visual Demo: 10/10 Ã 10 = 100 â Stunning design outputs
- Underdog Origin: 8/10 Ã 7 = 56 â LiblibAI (China), unexpected entrant
- Creator Tool: 8/10 Ã 7 = 56 â Agency-grade design for $90/mo
- Speed/Scale: 7/10 Ã 6 = 42 â 20,000 queued on day 1, 800K users in 2 months
- Wow Factor: 9/10 Ã 9 = 81 â “World’s first Design Agent”
- Anti-Establishment: 6/10 Ã 5 = 30 â Challenges Adobe/Canva
- Total: 446 / (7 matched patterns’ max = 756) â normalized across all 15 â ~81/100
Midjourney â Score: 78/100
- AI Visual Demo: 10/10 Ã 10 = 100 â Iconic AI image generation
- Wow Factor: 9/10 Ã 9 = 81 â Jaw-dropping visuals from text
- Creator Tool: 9/10 Ã 7 = 63 â Anyone can create stunning art
- Anti-Establishment: 7/10 Ã 5 = 35 â Challenges Adobe/stock photography
- Speed/Scale: 8/10 Ã 6 = 48 â 16M+ users, organic growth
- Meme Potential: 8/10 Ã 6 = 48 â Endless remixable output
Cursor (AI Code Editor) â Score: 72/100
- AI Agent: 8/10 Ã 9 = 72 â AI writes and edits code autonomously
- Wow Factor: 8/10 Ã 9 = 72 â “It just wrote the whole feature”
- Creator Tool: 9/10 Ã 7 = 63 â Turns anyone into a 10x developer
- Anti-Establishment: 8/10 Ã 5 = 40 â Challenges VS Code / JetBrains
- Speed/Scale: 7/10 Ã 6 = 42 â Explosive developer adoption
- Grok/xAI Adjacent: 5/10 Ã 8 = 40 â Coding + AI intersection
Suno (AI Music) â Score: 76/100
- AI Visual Demo: 8/10 Ã 10 = 80 â Audio is the “visual”; instantly shareable
- Wow Factor: 9/10 Ã 9 = 81 â Full songs from a text prompt
- Video/Multimodal: 9/10 Ã 8 = 72 â Audio generation, cross-modal AI
- Creator Tool: 9/10 Ã 7 = 63 â Anyone can make radio-quality music
- Playable/Interactive: 8/10 Ã 8 = 64 â Try it instantly, share results
- Meme Potential: 8/10 Ã 6 = 48 â Meme songs go viral
Loopit (AI Game Generator) â Score: 79/100
- Playable/Interactive: 9/10 Ã 8 = 72 â “Make everything playable”
- AI Visual Demo: 9/10 Ã 10 = 90 â AI-generated mini-games
- Underdog Origin: 8/10 Ã 7 = 56 â SeedLeap (China)
- Creator Tool: 8/10 Ã 7 = 56 â Anyone can create games
- Wow Factor: 9/10 Ã 9 = 81 â Instant game generation
- Grok/xAI Adjacent: 6/10 Ã 8 = 48 â Aligns with Musk’s AI gaming interest
DeepSeek â Score: 73/100
- Open-Source AI: 10/10 Ã 7 = 70 â Fully open-weight LLM
- Underdog Origin: 9/10 Ã 7 = 63 â Chinese hedge fund â top-tier LLM
- Anti-Establishment: 8/10 Ã 5 = 40 â Challenges OpenAI
- Wow Factor: 9/10 Ã 9 = 81 â GPT-4 performance at a fraction of cost
- Speed/Scale: 7/10 Ã 6 = 42 â Viral global adoption
Manus (AI Agent) â Score: 74/100
- AI Agent: 9/10 Ã 9 = 81 â General-purpose autonomous agent
- Underdog Origin: 8/10 Ã 7 = 56 â Chinese team
- Wow Factor: 9/10 Ã 9 = 81 â End-to-end autonomous task completion
- Speed/Scale: 7/10 Ã 6 = 42 â Viral launch, invite-only hype
What Musk Ignores (Anti-Patterns)
Avoid these â they virtually guarantee being ignored:
- ð« Beg-tagging: “@elonmusk PLEASE see this!! ððð”
- ð« Hashtag spam: #AI #MachineLearning #Startup #Tech #Innovation
- ð« Wall of text with no visual: Long explanations, no demo
- ð« Obvious paid promotion / shill tone: “This amazing product will change everything!!!”
- ð« Political hot-takes unrelated to tech: Partisan rants with no tech substance
- ð« Copycat / derivative products: “Like ChatGPT but⦔ with no unique angle
- ð« Excessive self-promotion: Multiple @elonmusk tags in the same thread
- ð« Low-effort screenshots: Static UI screenshots with no motion or interaction
Step 3: Generate Optimized Tweets
Tweet Formula
Based on analysis of content Musk engages with, use this structure:
[Hook - 1 line, max impact] + [Visual media] + [Optional: metric/claim]
Rules for Musk-Optimized Tweets
- Lead with the demo, not the pitch – Show, don’t tell
- Keep text under 50 words – Musk engages with brief, punchy content
- Attach video/GIF – Visual content gets 10x more engagement
- Use superlatives carefully – “World’s first X” works if defensible
- Tag @elonmusk only if organic – Don’t beg; make content worth sharing
- Post timing – See timing windows below
- Thread format – First tweet = hook + demo, reply with details
- No hashtags – Musk never engages with hashtag-heavy posts
- Mention Grok/xAI angle if relevant – “Built with Grok” or “Works with Grok” increases odds
- Include virality metric if available – “1M users in 24 hours” type stats
Posting Timing Windows (US Pacific)
| Window | Time (PT) | Notes |
|---|---|---|
| Prime | 10 PM â 2 AM | Musk’s late-night scroll; highest engagement probability |
| Secondary | 6 PM â 10 PM | Evening wind-down; decent reach |
| Weekend | Sunday 8 PM â 1 AM | Variable but Sunday evening is consistently strong |
| Avoid | 6 AM â 12 PM | Low engagement; posts get buried before Musk is active |
Tweet Templates
Template A: Product Launch
[Product Name] â [bold claim in â¤8 words]
[1-2 sentence description of what it does]
[Attach: 30-60s demo video or GIF]
Example:
Lovart â the world's first AI design agent.
Give it a brief. It delivers agency-grade creative in minutes.
20,000 people queued on launch day.
[45s demo video attached]
Template B: Milestone/Traction
[Metric] in [timeframe].
[Product Name] lets you [core value prop].
[Attach: demo video showing the "wow" moment]
Example:
1M songs created in 72 hours.
Suno turns a one-line prompt into a full, radio-quality track.
[30s audio/video demo attached]
Template C: Challenge the Status Quo
[Incumbent] charges $X/mo for [thing].
[Product Name] does it with AI for [fraction of price / free].
[Attach: side-by-side comparison video]
Example:
Adobe charges $55/mo for design tools most people never learn.
Lovart does it with one sentence for $9/mo.
[side-by-side comparison: 2 hours in Photoshop vs 30 seconds in Lovart]
Template D: Pure Demo (Highest Conversion)
[No text or minimal text â just the demo video/GIF]
[Reply to own tweet with product name + link]
Example:
[60s screen recording: type a prompt â watch a playable 3D game appear in real-time]
â³ Reply: "This is Loopit. Try it â loopit.ai"
Template E: Reply to Musk’s Tweet
[Monitor Musk's tweets on a relevant topic]
[Reply with your demo + one punchy line â no @tag needed, you're already in his thread]
Example:
Musk tweets: "AI is getting insane"
â³ You reply: "Made a game in 10 seconds with pure AI. No code." + [15s GIF of Loopit generating a game]
This is often the highest-conversion path â you’re already in his notification stream.
Output Format
For each generation, provide:
- Repost Score: X/100 with per-pattern breakdown (show the math)
- 3 Tweet Variants: Using different templates above
- Recommended Media: What visual to attach (format, length, content)
- Posting Strategy: Best time window, thread structure, follow-up plan
- Risk Assessment: What could prevent a repost
- Improvement Suggestions: How to increase the score
Musk Engagement Style Reference
When Musk reposts or engages, his style is:
- Single word: “Wow”, “Interesting”, “True”, “Cool”
- Single emoji: ð¥, ð, ð¤¯
- Brief comment: “This is the future”, “AI is moving fast”
- Quote tweet with minimal addition
Content that elicits these minimal-effort responses has the highest repost probability.
Step 4: Post-Publish Playbook
After posting, follow this sequence:
- 0â5 min: Reply to your own tweet with link + one extra detail or metric
- 5â30 min: Engage with early replies; like/retweet anyone who shares it
- 30â60 min: DM or tag 3â5 relevant mutual followers to start organic amplification
- 1â4 hours: If no traction, reply to a relevant Musk tweet with the same demo (Template E)
- 4â12 hours: If still no traction, consider reposting at the next Prime timing window with a different template
- Next day: Analyze impressions; iterate on the hook or media if engagement was low
Do not:
- Delete and repost the same tweet repeatedly
- Tag Musk in follow-up tweets if the first one didn’t land
- Spam the same content across multiple accounts
Important Notes
- This skill analyzes public patterns; no guarantee of actual repost
- Musk’s interests shift â AI/tech/robotics/efficiency is the current focus (2025-2026)
- Quality of the actual product matters more than tweet optimization
- Organic virality > manufactured virality
- The best strategy: build something genuinely impressive, then present it well