kappaemme-git/codex-startup-user-simulator-skill — explained in plain English
Analysis updated 2026-05-18
Test a landing page against five simulated customer personas to spot conversion leaks.
Compare an old and new version of a landing page with the same personas in before-after mode.
Get a prioritized list of fixes ranked by impact, effort, and confidence.
Rewrite the highest impact copy sections of a page using the copy mode focus.
| kappaemme-git/codex-startup-user-simulator-skill | autolearnmem/automem | billy-ellis/exr-imageio-poc | |
|---|---|---|---|
| Stars | 32 | 32 | 32 |
| Language | Python | Python | Python |
| Setup difficulty | easy | hard | moderate |
| Complexity | 2/5 | 5/5 | 3/5 |
| Audience | pm founder | researcher | researcher |
Figures from each repo's GitHub metadata at analysis time.
Codex Startup User Simulator Skill is a Codex skill that tests a startup, SaaS product, app, or landing page by simulating five distinct customer personas, each with their own reasons for converting, hesitating, or leaving. Instead of just describing a page, it inspects the real page, builds plausible decision makers based on the product's actual positioning, walks through each persona's conversion journey, and returns a set of prioritized fixes a founder could act on. It can test live URLs, apps running on localhost, screenshots, prototypes, or pasted copy. For each of the five simulated personas, it looks at clarity, relevance, trust, friction, and call to action problems, and it keeps observed evidence separate from inference and unknowns so the reader can tell what is fact versus guesswork. The output includes a scorecard comparing the five personas and ranks five suggested improvements by impact, effort, and confidence. It clearly labels the results as simulation, not as real user research. The default report includes a one sentence verdict, the biggest conversion leak, the five persona scorecard, evidence backed decision journeys, five prioritized fixes, the fastest measurable experiment to try next, and a note on the simulation's limitations. Codex also generates a polished, responsive HTML report saved to the workspace outputs folder, complete with expandable persona journeys, score visuals, a printable layout, and a save to PDF option. The skill supports five modes: quick for a fast verdict and top fixes, deep for a full desktop and mobile evaluation including pricing and CTA destinations, mobile for a small screen focused pass, before-after for comparing two versions of a page with the same personas, and copy for focusing specifically on positioning and headline language. It installs through npx or by manually copying the skill folder into the Codex skills directory, and Codex must be restarted afterward. The project is released under the MIT license.
A Codex skill that tests landing pages using five simulated customer personas to find conversion problems and suggest fixes.
Mainly Python. The stack also includes Python, Codex, HTML.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly pm founder.
This repo across BitVibe Labs
double-check against the repo, no cap.