Competitive Analysis Skill for Claude Code

A Claude Code skill that performs structured competitive product analysis before you write a single line of code. Because with AI-assisted coding, the hard part isn't building — it's knowing what to build.

What It Does

Input your product idea → get a structured competitive analysis report with real market data.

Two modes:

Mode	When to Use	Output	Time
Quick Scan	Early brainstorming, "just a rough idea"	~1500 words: top competitors, gaps, first feature to ship	~90s
Deep Analysis	Ready to commit, need thorough research	Full report: Strategy Canvas, ERRC, moat analysis, threat assessment	~4min

Frameworks included:

• JTBD (Jobs-to-Be-Done) — finds competitors you wouldn't think of
• Strategy Canvas — maps where competitors cluster vs. white space
• ERRC (Eliminate/Reduce/Raise/Create) — actionable differentiation strategy
• Moat Assessment — evaluates which competitors have durable advantages
• Threat Assessment — ranks competitors by actual danger level

Installation

Option 1: Project-level skill (recommended)

Clone into your project and reference it in your project's CLAUDE.md:

# In your project root
git clone https://github.com/hoopyAI/competitive-analysis-skill.git .claude/skills/competitive-analysis

# Add to your CLAUDE.md
echo '- Skill: .claude/skills/competitive-analysis/SKILL.md' >> CLAUDE.md

Option 2: Global skill

Clone to your global Claude config so it's available in all projects:

git clone https://github.com/hoopyAI/competitive-analysis-skill.git ~/.claude/skills/competitive-analysis

Then reference it in your global CLAUDE.md or settings.

Setup data source scripts (optional but recommended)

The skill bundles scripts for richer data from app stores and Reddit. To enable them:

cd <skill-path>/scripts
npm install

This installs app-store-scraper and google-play-scraper — zero config, works immediately.

For Reddit data (optional): Create a free Reddit app at https://www.reddit.com/prefs/apps (select "script" type), then set env vars:

export REDDIT_CLIENT_ID="your_client_id"
export REDDIT_CLIENT_SECRET="your_client_secret"

Without these, the skill falls back to WebSearch for Reddit data — still works, just less structured.

Usage

Just describe your product idea in Claude Code:

# Triggers Deep Analysis (default)
"I want to build a project management tool for freelance designers. What's the competitive landscape?"

# Triggers Quick Scan
"我想做一个旅行计划app，快速看看有啥竞品"

# Also triggers (Chinese)
"帮我做个竞品分析，我想做一个AI写作助手"

The skill auto-detects language and responds accordingly. It searches in both English and your language for comprehensive coverage.

Trigger phrases

The skill activates when you mention:

• Competitive analysis, competitor research, market research
• "What's already out there", "is anyone doing this", "should I build this"
• 竞品分析, 市场调研, 有没有类似的产品
• Sharing an app idea and asking about the market

Quick Scan vs Deep Analysis

Say this...	Gets you...
"quick look", "快速看看", "rough idea"	Quick Scan
Everything else	Deep Analysis

After a Quick Scan, the skill offers to go deeper.

Data Sources

Source	Method	Setup	What You Get
Web	WebSearch	None	Articles, comparisons, pricing pages
Google Play	Script	npm install	Search, ratings, reviews, download counts
App Store	Script	npm install	Search, ratings, app details
Reddit	Script	Free OAuth	User discussions, recommendations, complaints
Product Hunt	WebSearch	None	Product pages, launch comments
知乎 / 人人都是产品经理	WebSearch	None	Chinese market analysis, user discussions

Running scripts manually

# Search Google Play
node scripts/search-playstore.mjs search "diet tracker"

# Get app reviews
node scripts/search-playstore.mjs reviews "com.myfitnesspal.android"

# Search App Store
node scripts/search-appstore.mjs search "project management"

# Search Reddit (needs env vars)
node scripts/search-reddit.mjs "best calorie tracking app" --subreddit=fitness

Limitations

Data quality:

• All data comes from public sources (WebSearch + scraper packages). No access to paid databases like SimilarWeb, Crunchbase, or Sensor Tower.
• Products with poor SEO or very new launches may be missed.
• Pricing and feature info may be outdated if the latest articles haven't been indexed.

Platform-specific:

• App Store reviews are currently broken in the scraper package (as of March 2026). Ratings and search still work. Google Play reviews work fine.
• 小红书 (Xiaohongshu) content is mostly behind JS rendering and login walls. The skill uses indirect search strategies (searching Chinese PM sites that reference 小红书 trends) instead of scraping directly.
• Product Hunt API lacks text search — only supports topic-based filtering, so the skill uses WebSearch instead.
• Reddit requires a free OAuth app registration. Without it, falls back to WebSearch with site:reddit.com queries.

Scope:

• This skill does competitive landscape analysis only. It does not cover market sizing (TAM/SAM/SOM), financial projections, technical feasibility, or go-to-market strategy.
• Deep Analysis uses ~40-50k tokens and takes ~4 minutes. Quick Scan uses ~25k tokens and takes ~90 seconds.

Language:

• The skill responds in whatever language you use. Search coverage is best for English and Chinese markets. Other languages work but may find fewer local competitors.

Output

Reports are saved as markdown files in your working directory (e.g., competitive-analysis-freelance-pm-tool.md) and also displayed in the conversation.

License

MIT