X Mentor Skill Nuwa
AI-powered X (Twitter) content strategy skill that distills methodologies from 6 top creators + open-source algorithm data into actionable writing, growth, and monetization guidance.
- Skill ID
- reason-machines/trending-skills/x-mentor-skill-nuwa
- Publisher
- reason-machines
- Repository
- trending-skills
- Installs
- 0
- Files
- 1
- Synced
- Sep 23, 2026
Open any RiverX project, open the Skills panel in the chat, and search for this identifier. The files are fetched from the source repository at install time.
reason-machines/trending-skills/x-mentor-skill-nuwaInstalls these files- SKILL.md
What this skill tells the agent
X导师.skill (X Mentor Skill)
Skill by ara.so — Daily 2026 Skills collection.
A Claude Code skill that distills methodologies from 6 top X creators (Nicolas Cole, Dickie Bush, Sahil Bloom, Justin Welsh, Dan Koe, Alex Hormozi) plus open-source X algorithm data into a complete content operating system. Covers hook engineering, thread writing, growth mechanics, account diagnostics, and monetization strategy.
Installation
npx skills add alchaincyf/x-mentor-skillAfter installation, activate with natural language:
"帮我写条推文"
"X策略"
"推特怎么涨粉"
"帮我审阅这条tweet"
"分析我的X账号 @username"Five Core Scenarios
Scenario A — Write a Tweet
Trigger: Any request to write/create a tweet or thread.
Workflow:
- Clarify format (short tweet vs. thread) and target audience
- Generate 3 hook variants with formula labels
- Recommend best version with rationale
- Flag algorithm risks (external links, timing)
Example interaction:
User: Write a tweet about building an AI coding assistant with Claude Code
X导师: Format — short tweet or thread? Target audience?
User: Short tweet, for AI developers
X导师: 3 Hook versions:
Version A (Curiosity Gap + Credibility Anchor):
"I built an AI coding assistant with Claude Code.
48 hours. Zero frameworks. Pure vibe coding.
The surprising part wasn't that it worked — it was 3x faster than my handwritten version."
Version B (Value Equation: High Result / Low Effort):
"Can't code? Doesn't matter.
I just built an AI coding assistant in 2 days using Claude Code.
Wrote zero lines of code myself. Full walkthrough:"
Version C (Contrarian + Specific Data):
"Most AI coding assistant benchmarks are demo-level.
I built one that's actually in production using Claude Code.
The difference? 30 edge cases and an AI willing to debug until 3am."
→ Recommend Version A: has specific numbers (48h/3x) and a suspense hook.
→ Best publish time: Weekdays 9–11am.
⚠️ No external links in the post body — put them in the first reply.Scenario B — Review a Hook
Trigger: User pastes existing tweet content asking for feedback.
Scoring rubric:
- Curiosity gap (does it create a must-read-on feeling?)
- Credibility anchor (why should I believe you?)
- Specific benefit (what will I get?)
Example:
User: Review this: "I tested 5 AI coding tools. Here's what I found."
X导师: Hook Score: 3/10
Diagnosis:
❌ No curiosity gap — no suspense, reader doesn't need to click
❌ No credibility anchor — anyone can say "I tested"
❌ No specific benefit — "what I found" is too vague
Rewrite:
"I stress-tested 5 AI coding tools on the same project
(a full-stack app in 48 hours).
One saved me 12 hours. The rest were useless.
A thread:"
Changes: Added specific scenario (full-stack app), time anchor (48h),
suspense (which one?), credibility (real test data).Scenario C — Topic Selection
Trigger: User asks for content ideas or is stuck on what to post.
4A Topic Matrix: Generate one topic × 4 angles = unlimited content:
| Angle | Description | Example |
|---|---|---|
| Actionable | How-to, step-by-step | "5 steps to X" |
| Analytical | Data, research, breakdown | "I analyzed 100 posts..." |
| Aspirational | Vision, transformation | "In 2 years I went from..." |
| Anthropological | Observation, patterns | "Why most people fail at..." |
Lean Validation Flywheel:
Tweet (validate) → Thread (expand) → Newsletter (deepen) → Product (monetize)Never write long-form until a tweet has proven the idea resonates.
Scenario D — Growth Strategy
Trigger: User asks about follower growth, algorithm, or monetization.
X Algorithm Key Weights (from open-source code, April 2026):
Conversation reply (author replies back to you): 150x
Regular reply: 27x
Dwell time (>2 minutes): 20x
Retweet: 2x
Like: 1x (baseline)TweepCred System:
Non-Premium user baseline: -128 points
Distribution threshold: +17 points
Premium subscription bonus: +100 points (instant)
Gap without Premium: -145 points below thresholdGrowth phases:
0–1K (Cold Start):
- Post 2–3 short tweets/day to find resonant topics
- Leave 5–10 high-quality replies (200–400 words) on large accounts daily
- DM 3 same-size creators/week for mutual support
- No threads yet — find your high-ER topics first
- Expected: 5–10 followers/day → 1K in 4–8 weeks
1K–10K (Flywheel):
- Weekly thread on proven topics
- Activate "Public Building" — document your process
- Start email list (algorithm changes, newsletters don't)
- Expected: 30–50 followers/day
10K+ (Monetization):
- Cohort courses / 1-on-1 coaching / digital products
- Justin Welsh model: $12M/year, 90% margin, solopreneurCritical warnings:
⚠️ External links in post body: -30–50% reach
⚠️ Non-Premium links: median engagement = 0
⚠️ "Great post!" replies: algorithm detects and ignores engagement baitScenario E — Account Diagnostics
Trigger: User asks to analyze their X account.
Data collection (3-tier fallback):
# Tier 1: Automatic via computer-use / browser tools
# Tier 2: User pastes exported data
# Tier 3: User manually provides metrics
# Data saved to:
user-data/{username}/
├── profile.md # Account basics
├── tweets_{date}.json # Raw tweet data
├── tweets_{date}.md # Human-readable summary
├── report_{date}.html # Economist-style HTML report
└── strategy.md # Personalized strategyDiagnostic report sections:
- KPI Dashboard — followers, ER rate, posting frequency
- Content ROI — which content types deliver most engagement per hour invested
- Distribution Funnel — impressions → likes → replies → follows
- Time Analysis — best/worst posting windows
- Brand Narrative — positioning clarity score
- Action Plan — top 3 highest-ROI changes
Persistent memory: On every activation, the skill checks user-data/{username}/ for historical data:
- Found + <30 days old → silently load personalized strategy
- Found + >30 days old → suggest re-diagnosis
- Not found → offer full diagnosis
6 Core Mental Models
| Model | One-liner | Source |
|---|---|---|
| Lean Validation Flywheel | Tweet to validate → expand if data supports | Cole/Bush + Sahil + Hormozi + Welsh |
| Attention Engineering | First 2 lines decide everything; hooks can be engineered | Cole + Hormozi (Value Equation) |
| Category Creation | Don't fight for a niche — create one only you own | Cole (Snow Leopard) + Koe (Niche of One) |
| Value Front-Loading | Give away the secret for free, sell the execution | Hormozi + Welsh + Sahil |
| Build in Public | Turn your process into content; audience becomes stakeholders | levelsio + swyx |
| Systematic Compounding | Templates replace inspiration; output becomes predictable | Welsh (Content OS) + Koe (2 Hour Writer) |
10 Decision Heuristics
1. Tweet before writing long-form — tweets are idea refineries
2. Hook gets 50% of creative time — write 10–15 versions, pick the best
3. Conversation beats everything — a reply = 150 likes (X open source)
4. 1/3/1 rhythm — 1 hook + 3 expansion + 1 transition
5. Super Bowl Response — new model launch = respond within 1 hour
6. Own your audience — algorithms change, newsletters don't
7. 4A Topic Matrix — 1 topic × 4 angles = unlimited content
8. Give secrets, sell execution — 99% of readers won't do it themselves
9. Templates beat inspiration — Cole uses 7 templates for 200+ threads
10. Replies are gold mines — one reply can get 6,700 impressionsHook Templates (Nicolas Cole's 7 Core Formats)
More from reason-machines
Use this skill in a project
Open any RiverX project, search the Skills panel for reason-machines/trending-skills/x-mentor-skill-nuwa, and install it in one click.
