mogkit
Open-source toolkit for product managers
About
mogkit installs PM craft into the terminal: 13 methodology-backed skills that run in Claude Code, a thin CLI (published on npm as mogkit) that scaffolds a working corpus, and a workflow library of implementable AI-for-PM recipes. MIT licensed, live at mogkit.com.
The design thesis is strict: interrogator, not generator. No skill outputs a finished deliverable. Skills output reasoning scaffolds: what your evidence supports, what you are assuming, and what to validate next. It will not write your PRD; it makes you write a better one.
Process
The skills
13 skills in 2 tiers: 7 standalone skills that work on pasted input (metrics trees, spec stress-tests, tradeoff framing) and a 6-skill Discovery engine that builds and queries an evidence graph over a real research corpus.
The CLI
A TypeScript CLI scaffolds the workspace in 1 command (npx mogkit init), then keeps it honest: a corpus health report and a doctor command that verifies the setup. Schema validators run in CI against the graph format.
The site and workflow library
An Astro and Tailwind site with an MDX workflow library, plus a from-zero setup guide written for PMs who have never opened a terminal.
The principles doc
Every skill, command, and page derives from a written set of non-negotiable rules, so contributions stay coherent with the thesis instead of drifting into another prompt pack.
Outcome
Shipped v0.1 end to end as a solo project: 13 skills, the published CLI, the Discovery engine with an exemplar graph, CI schema validation, and the public site.