The reverse-engineer-marketing skill produces a growth teardown: a history of a company in date order. It builds a monthly traffic curve and a log of dated moves from sitemaps, web archives, ad libraries, job posts, DNS records, reviews and the team’s public statements, then finds the clusters of moves that came before growth.
It compares the target with the user’s own numbers, lists where the target is weak, and ends with plays ranked by evidence, each routed to the skill that does the work with a paste-ready prompt. Modes include teardown, content, quick, funnel hack, landscape and watch.
What You Get
- A teardowns folder per target with playbook.md, traffic.csv, moves.csv and a moves-versus-growth table
- A timeline of dated moves with the growth inflection marked and clusters of moves credited together
- A gap table that sets the target against your own data by funnel stage
- A where-they-are-weak section built from low reviews, price rises and dropped markets
- Ranked plays, each with the skill that runs it, effort, metric and a paste-ready prompt, plus plays to skip
Common Use Cases
- Studying how a fast-growing competitor built its search and referral channels
- Preparing for a sales or partnership call with a quick first look at a company
- Comparing several competitors in a category to decide who to study
- Monitoring a competitor monthly and reporting only new moves
What Makes It Unique
It treats a competitor study as a dated history and refuses to credit a single move for a jump, crediting clusters of moves within six weeks and naming funding, press and seasonality as confounders. Code computes the arithmetic, a pack for each of more than thirty industries adds its own signals and limits, and the output ends in plays routed to the other skills in the set.
Under The Hood
The skill ships one Node script, teardown.mjs, that the agent runs from the shell with no dependencies. It makes network requests to fetch public web pages, sitemaps, web archive snapshots, DNS records, certificate logs and job boards of the target, and some commands read API credentials for the paid data services a user connects. It writes CSV files and reports under a teardowns folder in the project. The agent also works within a dollar cap for paid calls that it must quote first. Sign-ups, demos or purchases at the target happen only after the user approves each action, and fetched text is treated as data.