convez.ai
SkillDados & AnalyticsIntermediário
Mahanaicoach

Competitor X-Ray

Mahanaicoach/competitor-x-ray

Analise competidores em ICP, funil e monetização com evidências verificadas

Ver no GitHub
InicialPluginScore · de 0 a 100

Skill que analisa competidores em profundidade, extraindo ICP, funil de vendas e fluxos de monetização com evidências verificadas. Gera relatórios em PDF, planilhas ou documentos Word com dados quantitativos (seguidores, tráfego, financiamento) e qualificação de cada fonte como confirmada, reportada ou inferida.

Instalação

$npx convez-ai Mahanaicoach/competitor-x-ray

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
19
Forks
7
Acompanhando
0
Saúde
42%
Último commit
há 2 semanas
Versão
v1.0.0
Issues abertas
0
Contribuidores
1

Casos de uso

  • Análise competitiva para startups e PMEs identificarem ameaças reais
  • Pesquisa de mercado para entender modelos de negócio de concorrentes
  • Due diligence de investidores e analistas de mercado
  • Estratégia de pricing comparativa entre players do mesmo segmento

Veredito

Skill bem executada que diferencia-se por rigor metodológico (camadas de evidência, separação público vs. cliente pagador) e saída formatada profissional. Útil para quem precisa de inteligência competitiva estruturada, não de achismo.

README

A tradução para o português deste README ainda está na fila.

Competitor X-Ray — x-ray any competitor into ICP, funnel & monetization

Competitor X-Ray 🔎

A free Claude skill from Mahan AI.

Hand it a list of competitors — Instagram/TikTok/X/LinkedIn handles, company names, or website URLs — and it decodes each one into the three things that actually matter:

  • ICP — who they really sell to (audience vs. paying customer)
  • Funnel — how a stranger becomes a customer, stage by stage
  • Monetization — every revenue stream, with price points

Then it zooms out into a cross-competitor threat ranking so you can see who to actually worry about.

It runs one research agent per competitor in parallel, verifies every claim against real sources, layers in a quantitative snapshot (followers, traffic, growth, funding), and can deliver the result as a chat summary, a spreadsheet, a Word doc, or a branded, designed PDF report.

👋 Built and given away by Mahan AI. If this is useful, follow @mahanaicoach for more AI + automation builds.


⚡ Quick start (60 seconds)

  1. Download competitor-x-ray.skill from this repo. (Click the file → the Download button.)
  2. Open it in the Claude desktop app (Cowork) and click Save skill.
  3. In any chat, just say:
research these competitors for me: @sabrina_ramonov @thedankoe gregisenberg

That's it — you don't even have to name the skill. Claude picks it up automatically.

Using Claude Code instead? See INSTALL.md for the folder-copy method.

📖 Want the deep dive? GUIDE.md explains exactly how it works and how to get the most out of it.


What makes it different

Most "competitor research" is vibes. This skill holds a quality bar:

  • Audience ≠ customer. It separates the people who follow someone from the people who pay them — usually the single most useful insight.
  • Three evidence tiers. Every claim is tagged CONFIRMED (seen on an owned page), REPORTED (self-reported / third-party estimate, unaudited), or INFERRED (a reasoned guess). Self-reported revenue never gets passed off as fact.
  • Real numbers. A quantitative layer pulls followers/growth (Social Blade, HypeAuditor), and for companies: funding/headcount/G2/traffic.
  • Gated profiles don't stop it. A concrete fallback chain pulls facts from owned sites, link-in-bio hubs, communities, podcasts, and analytics when Instagram/LinkedIn are login-walled.
  • It never fabricates. "Not found" is reported as "not found," not invented.
  • Presentation matters. The designed-PDF output looks like a real intelligence brief, not a wall of text.

Usage examples

Just describe what you want and paste the targets — you don't need to name the skill:

research these competitors for me: @sabrina_ramonov @thedankoe gregisenberg
break down bonusly and nectar — ICP, funnel, and how they make money, with the numbers
reverse-engineer this creator's funnel: instagram.com/<handle>

The skill will ask which output format you want (designed PDF, chat, spreadsheet, or Word doc), research each target in parallel, and deliver.


What you get

For each competitor:

  • ICP (audience + buyer), pain points
  • Funnel mapped top → middle → bottom
  • Every monetization stream with prices
  • A quantitative snapshot (sourced + dated, evidence-tiered)
  • A links directory — their site, socials, podcast, community, store
  • A confidence note

Plus a synthesis: shared patterns, where they diverge, a threat ranking by maturity × momentum, and the single sharpest takeaway.


How it works

targets → normalize (handle / company / URL)
        → ask output format
        → fan out: one research agent per competitor (parallel)
        → quantitative layer + evidence tiers
        → synthesize threat ranking
        → render in the chosen format

The research standard each agent follows lives in competitor-x-ray/references/research-prompt.md. Output templates are in competitor-x-ray/references/output-formats.md, and the designed-PDF system is in competitor-x-ray/references/designed-pdf.md.


Repo contents

PathWhat it is
competitor-x-ray.skillThe packaged skill — download this to install in Cowork
competitor-x-ray/SKILL.mdThe skill's brain: the full step-by-step instructions Claude follows
competitor-x-ray/references/Research-agent prompt, output templates, and the designed-PDF system
competitor-x-ray/evals/evals.jsonTest cases proving the skill works on real targets
INSTALL.mdDetailed install for both Cowork and Claude Code

Optional dependency

The designed PDF output renders with WeasyPrint:

pip install weasyprint

Everything else (chat summary, spreadsheet, Word doc) works without it.


Limitations & honesty

  • Without a logged-in browser, follower/engagement numbers for gated profiles may be estimates (tagged REPORTED).
  • Exact website traffic is usually an order-of-magnitude range unless you have a paid SEO data source.
  • It tells you what it couldn't verify instead of guessing — by design.

License

MIT — see LICENSE. Use it, fork it, ship it.


Made by Mahan AI · Follow for more AI + automation builds → @mahanaicoach

Ver o original no GitHub

Bom para

Relacionados