X Grow

Created By

BlockRunAI14 days ago

# x-grow

# twitter-optimization

Overview Content Tools Comments

Content

X Grow

A Claude Code skill that helps you write high-performing X (Twitter) posts using the actual X algorithm.

Works standalone. Even better with BlockRun MCP for real-time research + image generation.

Features

Mode	Command	Description
Draft	`/x-grow draft "topic"`	Research and write optimized posts
Review	`/x-grow review`	Analyze and score your draft
Image	`/x-grow image "desc"`	Generate post visuals

What You Get

Draft Mode:

Optimized post ready to copy
3 variations (different hooks, formats, tones)
Algorithm explanation (why it should perform)
Best practices applied
Image suggestions

Review Mode:

Score out of 10
Checklist audit (hook, length, engagement triggers, etc.)
Optimized rewrite
Explanation of changes
Actionable tips

Image Mode:

AI-generated visuals optimized for X
Posts with images get 2x engagement
~$0.05 per image via BlockRun

Installation

As a Claude Code Skill

# Clone to your skills directory
git clone https://github.com/blockrunai/x-grow ~/.claude/skills/x-grow

Optional: BlockRun MCP (Enhanced Features)

For real-time research and image generation:

# Install BlockRun MCP
claude mcp add blockrun -- npx @blockrun/mcp

# Fund your wallet with USDC on Base network
# The MCP will generate a wallet and show you a QR code

Usage

Draft a Post

/x-grow draft "launching my new AI tool"

Output:

Researched post optimized for engagement
Multiple variations
Algorithm reasoning
Best practices applied

Review Your Draft

/x-grow review

[paste your draft]

Output:

Score: 7/10
Checklist showing what's working and what's not
Optimized version
Specific improvements explained

Generate an Image

/x-grow image "abstract visualization of AI agents making payments"

Output:

Generated image optimized for X
Suggested post to accompany it

How It Works

Standalone (No BlockRun)

Algorithm knowledge baked in from x-algorithm repo
Best practices from high-performing posts
Drafting and review work fully

With BlockRun MCP

Real-time research via Grok API (~$0.03)
Live examples of what's working right now
Image generation via Nano Banana (~$0.05)

Algorithm Insights

The skill knows what X's ranking algorithm actually values:

Action	Weight	What This Means
You reply to comments	+75.0	150x more valuable than a like
Getting replies	+13.5	Conversation > vanity metrics
Profile clicks	+12.0	Curiosity about you = reach
Retweets	+1.0	Distribution, but less than replies
Likes	+0.5	Lowest priority signal
"Show less" clicks	-74.0	Kills your reach
Reports	-369.0	Devastating penalty

Key insight: One author reply to a comment is worth 150 likes.

Cost

Action	Without BlockRun	With BlockRun
Draft	Free	~$0.03 (research)
Review	Free	Free
Image	N/A	~$0.05

Directory Structure

x-grow/
├── skill.md              # Main skill (Claude Code loads this)
├── knowledge/
│   ├── algorithm.md      # X algorithm reference
│   └── best-practices.md # Proven patterns
├── prompts/
│   ├── draft.md          # Draft mode template
│   ├── review.md         # Review mode template
│   └── image.md          # Image mode template
├── docs/
│   └── plans/            # Design documents
└── README.md             # You're reading this