saurun

Use proactively for .NET backend implementation work. Implements features, fixes bugs, and refactors code following DDD patterns with rich domain models, Result<T>, and DTOs at API boundaries. Does NOT handle frontend code, infrastructure, or deployment.

Validate whether a gap identified by Codex is a genuine completeness gap that needs fixing. Compares against source plan.

Use proactively for React frontend implementation work. Builds components, pages, and features using Tailwind CSS v4, shadcn/ui, and Zustand. Applies the frontend-design skill for distinctive visual design. Does NOT handle backend code, React Native, or deployment.

Transform raw ideas or business goals into structured, actionable product plans. Create user personas, detailed user stories, and prioritized feature backlogs. Use for product strategy, requirements gathering, and roadmap planning.

Design user experiences and visual interfaces for applications. Translate product manager feature stories into comprehensive design systems, detailed user flows, and implementation-ready specifications. Create style guides, state briefs, and ensure products are beautiful, accessible, and intuitive.

Use when user wants brutally honest feedback on business ideas, plans, pitches, or strategies. Triggers on "brutal feedback", "honest review", "devil's advocate", "red team", "tear apart", "critique", "what's wrong with this", "find holes in", "challenge this idea", or requests for harsh/critical evaluation.

Catchup — Resume Work After Time Away

Use when user wants to ask Codex, get GPT's opinion, get a second opinion from OpenAI, compare perspectives, have Codex review code or docs, or let Codex fix/refactor/implement something.

Use when reviewing Danish copy for anglicisms, technical jargon, defensive tone, vague formulations, English words, and overly formal language. Scans uncommitted files by default or specific files with --file flag.

Use when user wants to deploy/publish a skill to saurun-marketplace, says "deploy [skill] to marketplace", or asks to publish a skill to the plugin marketplace.

---

Use when implementing ASP.NET Core + EF Core features or bugfixes with xUnit + NSubstitute tests, before writing implementation code.

Use when validating implemented User Flows against product spec, generating demo videos for stakeholders, or running pre-release E2E QA. Triggers: spec complete with User Flows section, need to verify user journeys work end-to-end, or preparing demo artifacts.

Verify that a completed implementation matches its plan. Use after implementation subagents finish, when checking if a plan was fully executed, or when the user asks to verify/check/audit implementation progress against a plan file. Takes a plan file path as argument. Reads the plan, analyzes git diff, runs verify conditions, and outputs a structured JSON verdict. Read-only auditor — never modifies code.

Use when a plan needs quality assurance before execution, especially after writing-plans completes. Catches intent gaps (missing user requirements) and quality gaps (vague, inconsistent, or unexecutable tasks).

Use when CLAUDE.md is bloated or ineffective. Use when Claude ignores instructions, file exceeds 150 lines, response formatting rules aren't working, or user asks to optimize/restructure their Claude Code configuration.

Use when building React frontends with Zustand, TanStack Query, Tailwind v4, or shadcn/ui. Mandatory constraints for component architecture, state management, styling, and pre-commit verification.

Use when architecting or implementing React frontends with Zustand, TanStack Query, or component hierarchies. Symptoms - unclear where state lives, stores causing re-renders, no API caching strategy, inconsistent error handling.

Use when building React components with Tailwind CSS v4, shadcn/ui, or Radix primitives. Use when writing className strings, cva variants, or CSS variable values in Tailwind.

Use when implementing React + Vite + TypeScript features or bugfixes with Vitest + React Testing Library tests, before writing implementation code.

Use when improving, polishing, cleaning up, or fixing inconsistencies in any text artifact. Use when asked to "polish", "refine", "clean up", "make more concise", "fix inconsistencies", "improve this doc", "review this code", "refine config", or "refine plan".

Scaffold a complete, deploy-ready .NET 10 + React 19 project from scratch. Creates backend (ASP.NET Core + SQLite), frontend (Vite + Tailwind v4 + shadcn/ui), Dockerfile, GitHub repo, and Railway deployment — all in one skill call. Use whenever the user wants to create a new greenfield project, start a new app, scaffold a project, bootstrap a codebase, or mentions "new project" or "greenfield". Even if they don't say "scaffold" explicitly — if they want a fresh .NET + React project set up, this is the skill to use.

Deploy a new project to Railway via GraphQL API. Creates project and service named after the required project-name parameter, connects GitHub repo, sets Dockerfile builder (mandatory — Railpack ignores Dockerfiles), creates persistent SQLite volume at /data by default, configures env vars, and public domain. Use when deploying to Railway, when scaffold invokes Railway setup, or when user says "deploy to Railway".

Use when generating videos with Google Veo 3.1, writing video prompts, or creating image-to-video animations. Triggers on "generate video", "veo", "video generation", "animate image", "text-to-video".

Use when user wants to explore, compare, or A/B test multiple design styles, visual directions, or mockups for a web page before implementation begins. Triggers on "show me design options", "style exploration", "design comparison", "visual options".