georgeshani-skills

Guidance for a single Angular + Ionic + Capacitor codebase shipping to Web, iOS, and Android — from initial project setup through architecture and store release. Use this whenever scaffolding a new Ionic/Capacitor project, adding Capacitor or Ionic UI to an existing Angular app, choosing an initialization approach, picking state management or testing tools, deciding where code should live (shared vs. platform-specific), designing or editing CI/CD pipelines, discussing branch strategy, or handling mobile store distribution (TestFlight, Google Play Internal Testing, App Store, Google Play Production). Also covers Capacitor plugins, native permission/config syncing, Fastlane, versioning/build numbers, and GitHub Actions workflows for this stack. Trigger even if the user just says "the mobile pipeline," "the release process," or "the app" without naming Ionic or Capacitor directly, once the stack is Angular/Ionic/Capacitor.

Audit a mobile app against Apple App Store and Google Play publication requirements before submission, and diagnose rejections after them. Covers privacy policy and account-deletion requirements, SDK and target-API floors, store-listing metadata, data-disclosure forms, permissions policy, age ratings, and regional law (GDPR, EU Accessibility Act, DSA trader status, age assurance). Use this whenever someone is preparing to publish or update an iOS or Android app, asks what is needed to get an app approved, got a rejection from App Review or Play Console, is adding accounts / subscriptions / login / analytics / AI features to a mobile app, or is starting a React Native, Expo, Flutter, Ionic, or Capacitor project that will eventually ship to a store — even if they do not mention "compliance," "review," or "rejection" by name. Also use it when reviewing backend or data-model design for a mobile app, because account deletion and data disclosure are far cheaper to design in than to retrofit.

A universal engineering-partner, mentor, and code-reviewer skill for anyone writing, debugging, reviewing, or architecting software, and for anyone who wants to learn a technical concept deeply rather than just get a task done - e.g. 'how does the TypeScript compiler actually work.' Runs in two modes: Build Mode (production-quality code, explained trade-offs, honest pushback on weak designs, root-cause debugging) for implementation and review requests, and Learn Mode (first-principles, mechanism-level explanations with primary sources and a hands-on next step) for requests whose real goal is understanding rather than shipping. Falls back to web search when a helpful MCP server isn't configured, then suggests one and writes setup steps. Uses only the portable Agent Skills frontmatter, so it works unmodified in Claude Code, Codex, Cursor, or any compliant AI agent harness. Use for building software or for understanding something technical in depth.