Competitor Teardown

Most competitor write-ups mix facts, marketing claims and guesses, and you can't tell which is which. Six weeks later nobody remembers where a number came from.

This skill gives your AI a research method for one competitor at a time. It covers positioning, audience, pricing and packaging, features, onboarding, review sentiment themes, distribution channels, strengths and weaknesses, and opportunities for your own product. Every finding cites a source URL with an access date and a High, Medium or Low confidence label. Marketing claims are attributed, review themes are counted, and anything it couldn't find is listed as Unknown along with where it looked.

Inside: SKILL.md with an 8-step procedure, a research method (evidence log, confidence rules, how to handle conflicting sources, where to look for each section, ethical ground rules), a scoring rubric with anchors for 8 dimensions and an opportunity ranking, a teardown template, and a Node checker that flags uncited findings, missing confidence labels, unknown source IDs, bad URLs and stale access dates. The example is a full teardown of a fictional company built offline from 8 page snapshots, so you can see the output before you run it.

For founders, product managers, marketers and sales teams. It uses only public pages your AI can reach, doesn't sign up for trials or contact the competitor, and is not legal advice for publishing comparisons.