concorrentes is an open-source Agent Skill for Claude Code that runs a self-expanding competitor analysis: it sweeps the Meta Ad Library (Facebook and Instagram ads), Google Play, the App Store and competitor landing pages, finds competitors you didn't know about, and ends with pivots, positioning and decisions for your own project.
Run /concorrentes and the skill reads your repo first, then loops through searches until a full round finds no new names. It writes the result to docs/CONCORRENTES.md and replies in chat with 10 lines at most.
Competitor analysis · competitive research · competitor discovery · Meta Ad Library · Facebook Ad Library · app store research · Google Play competitor analysis · market validation · product pivot · positioning · go-to-market research
/concorrentes # infers the niche from the repo
/concorrentes couples finance BR # niche + country
/concorrentes Pierre, Fikko # names you already know
Every competitor analysis I did stopped at the 3 names I already knew. What changes decisions is the 4th, 5th or 12th name: the one in an ad's copy, in a "best apps" article, or in the store's "similar apps" list. This skill forces the search loop until it runs dry.
| Plain prompt | concorrentes |
|---|---|
| Lists the competitors you already named | Expansion loop: every advertiser, article, feature and adjacency becomes a new search term (3 rounds minimum) |
| Guesses who advertises | Reads the Meta Ad Library in the browser: creatives, start date, repeated copy, CTA, domain |
| Generic market overview | Compares every competitor against what your project already decided (README, PRODUCT, ROADMAP) |
| Store data from memory | Opens each Google Play / App Store listing: downloads, rating, reviews, last update, in-app price |
| Ends with "consider differentiating" | Ends with pivots, risks and decisions, each with a trigger to revisit |
Skill (recommended), via skills CLI
npx skills add WednyFernandes/competitor-research-skillManual, as a global command
git clone https://github.com/WednyFernandes/competitor-research-skill
cp competitor-research-skill/skills/concorrentes/SKILL.md ~/.claude/commands/concorrentes.mdManual, as a project skill
mkdir -p .claude/skills/concorrentes
cp competitor-research-skill/skills/concorrentes/SKILL.md .claude/skills/concorrentes/Requires: Claude Code with the built-in browser (desktop app) or the Claude in Chrome extension, plus WebSearch and WebFetch.
Type /concorrentes with an optional argument: a niche, a country, or competitor names you already know. With no argument, the skill infers the niche from the repo. It also triggers when you ask for a competitive analysis, a benchmark, "who else does this", whether a market is worth entering, a pivot, positioning, or before spending on paid media.
| Step | What happens |
|---|---|
| 0. Context | Reads README.md, docs/PRODUCT.md, ROADMAP.md and any existing CONCORRENTES.md. Sums up the project in 5 lines: what it is, who it's for, how data gets in, price, and what was decided not to do |
| 1. Sources | Meta Ad Library via the browser (headless scrapers get a 403), WebSearch, Play Store and App Store listings, WebFetch of each landing page, and the MCP registry or AdWhispr when connected |
| 2. Expansion loop | Names cited in ad copy, "best apps" articles, "similar apps", competitor blogs, features found in 3+ competitors and adjacent products all become new search terms. It stops only when a full round finds nothing new |
| 3. One row per competitor | Name, audience, how data gets in, privacy, price and trial, paid media (creatives, running since), store stats, distribution channel |
| 4. Analysis | What the ads teach, what the stores teach, top 2–3 vs. us, what to copy, where they lose, pivots, risks, decisions with triggers |
| 5. Output | Writes docs/CONCORRENTES.md with the new section on top and history preserved below, plus a source list. It doesn't commit unless asked |
- An ad running 60+ days is a creative that converts.
- 10k+ downloads with reviews means real distribution.
- A name cited in a competitor's copy ("vs X") is the next search term.
Excerpt from a real docs/CONCORRENTES.md generated for a couples finance app (Brazil, Sep 2026), translated:
One-line summary: couples + privacy + manual entry gets paid for (R$ 14.90–19.90, proven by Duofi, Fikko, Balance). Nobody in Brazil imports bank statements. The winning paid-media pitch became "send it on WhatsApp"; we lose there and won't enter.
Couples Privacy How data gets in Price Paid media Where Pierre "no fights" creative since Jun 30 no Open Finance + WhatsApp 7-day trial 30+ creatives since May '26 Play 4.7 Duofi "yours, mine and ours" yes 100% manual R$ 14.90 1 creative, Sep 11 PWA Balance native partial manual 7-day trial — Play 50k+, 4.7, 1,320 reviews
skills/concorrentes/SKILL.md # the skill: context, sources, expansion loop, analysis, output
LICENSE
Why is the command called /concorrentes?
It's Portuguese for "competitors". The skill was written in Portuguese and defaults to Brazil-focused searches (BRL prices, country=BR in the Ad Library), but you can pass any niche or country as the argument.
Why does it need a browser? Headless scrapers get a 403 from the Meta Ad Library. The skill opens the Ad Library and the store listings in the built-in browser and reads the page text instead.
Does it change my repo?
It writes or updates docs/CONCORRENTES.md and nothing else. It doesn't commit without asking.
How is it different from what-am-i-missing?
what-am-i-missing checks a plan against laws, standards and frameworks. concorrentes looks outward at who is already in the market, what they pay to advertise, and what that means for your positioning.
MIT © Wedny Fernandes