Lauren Tan's (poteto) MIT-licensed plugin of the skills used day to day to ship code at Cursor. One entry skill, `/poteto-mode`, matches a task to one of twenty-three playbooks and pulls in the other skills as the steps need them. Installs in Cursor and in Claude Code.
The goal is less code of higher quality, so that you can trust several agents running in parallel. Start a task with /poteto-mode and it picks a playbook (investigation, bug fix, perf, hillclimb, feature, refactoring, babysit, shipping, autonomous run and others), copies its steps into a todo list and routes to skills such as how, why, architect, arena, swarm, interrogate, tdd and unslop. A set of principle-* skills encodes rules like fix root causes, prove it works and subtract before you add.
Several skills are multi-model: /interrogate has different models try to break a diff and keeps only findings a run proves, and /setup-pstack lets you choose which models handle code delegates and which handle judgment. It also ships create-formal-verification-skill for modelling state machines and statechart for visualizing user flows.
In Cursor, run /add-plugin pstack. In Claude Code, /plugin marketplace add cursor/plugins, then /plugin install pstack@cursor-plugins; commands get the pstack: prefix there, and steps that need Cursor-only features are skipped.