← ClaudeAtlas

martech-teardownlisted

Reverse-engineer any company's marketing-technology stack from public signals and produce a structured teardown report — what tools they run across every martech space (CMS, analytics, CDP, CRM, ads/pixels, consent, email, enrichment, affiliate, payments) and how their tracking, identity, and ad-measurement actually work. Use this WHENEVER the user names a brand, competitor, or website and wants to understand, audit, map, analyze, or "tear down" its martech / adtech / growth stack — including phrasings like "what tools does X use", "analyze X's marketing stack", "how is X tracking users", "do a martech teardown of X", "what CMS/CDP/ analytics does X run", "reverse engineer X's tracking", or "compare our stack to X". Trigger even if the user only gives a domain and says "audit this". Personal skill; keep every report generic and vendor-neutral unless the user supplies their own org context to compare against.
almoretti/martech-ai-skills-and-tools · ★ 2 · AI & Automation · score 56
Install: claude install-skill almoretti/martech-ai-skills-and-tools
# Martech teardown Turn a bare brand name or URL into an evidence-backed picture of the company's entire marketing-technology stack, plus how their tracking, identity resolution, and ad measurement are wired. The output is a competitive-intelligence report built from **public pages, the live tag container, DNS records, the sitemap, and the running site's own network traffic** — never from anything private. ## Operating principle A teardown is only as good as its evidence. Every tool you name must trace back to a concrete signal — a cookie, a header, a DNS record, a tag ID, a network request, or a vendor-published case study. Always separate **what you observed** from **what you infer**, and tag every claim Confirmed / Inferred / Assumed. That honesty is the whole value; a confident-but-wrong stack list is worse than a short accurate one. ## Ethics & scope (read once, hold throughout) - Passive, public signals only: GET requests, DNS, sitemaps, and observing what the site itself loads in a normal browser. This is standard competitive intelligence. - **Do not** attempt logins, submit forms, hit private/authenticated APIs, probe for vulnerabilities, or scrape behind a paywall. - The authenticated audit path exists ONLY for the operator's **own** account on a service they use, to see how *their own* dataLayer changes when logged in. Never use anyone else's credentials, and never use it to harvest personal data — the point is the tracking architecture, not the accou