replica-entrepreneur collects 100 or more public reviews of an app across at least three sources, such as the App Store, Google Play, G2, Reddit and Hacker News. A bundled script, reviews.py, sorts them into themes and ranks them by rating and recency.
From the ranking the agent writes three lists (what they hate, what is missing, what is unsolved), a fix plan of five to eight items and three positioning options with evidence counts.
What You Get
- A review sheet where every row carries its source, link, date, rating and exact text
- A ranked report of complaint and request themes with review counts, source counts, average stars and linked quotes
- Three lists: what they hate, what is missing and what is unsolved
- A fix plan of five to eight items with size, owning skill and evidence, added to the feature matrix
- Three positioning options in a fixed template, with one recommended
Common Use Cases
- Finding out whether an app is worth cloning by reading what its users complain about
- Turning negative reviews into a roadmap and a one-line positioning statement
- Spotting pricing and billing complaints to pass to the launch step
What Makes It Unique
It forbids invented evidence: every quote is verbatim with a URL, and the script drops any row without a link, marks themes with fewer than three reviews or one source as thin, and warns when the sample is under 30. Reviewer quotes are treated as research and never as testimonials.
Under The Hood
reviews.py is a standard-library Python script with a themes.json file of regular expressions. It reads your CSV, ranks themes and writes a markdown report, with no network calls of its own. The skill’s instructions have the agent read review pages in the browser and may call public endpoints: Apple’s customer reviews feed, the Hacker News Algolia API and Reddit’s official API. It then writes the CSV, feedback, fixes and feature matrix files in your project.