Skills Directory
Claude Skills by Intense-Visions
github.com/Intense-Visions> No implicit trust based on network position, VPN status, or previous authentication -- every request is authenticated, authorized, and encrypted regardless of origin
> LLM-judgment critique of spec quality (proposals + ADRs) against a curated rubric catalog from the spec-quality canon. Per-section critique with rubric-to-section mapping. Second member of the craft-pipeline initiative; highest-leverage craft skill because spec quality compounds across the entire planning → implementation → review lifecycle below it. Emits 3-axis findings (tier × impact × confidence per ADR 0019).
> Manage shared state with React Context and useReducer for prop-drilling avoidance and scoped state
> Build bottom-up atomic state with Jotai for granular, composable React state management
> Select and derive state efficiently to minimize component re-renders across any state management library
> Separate server state from client state and synchronize them with TanStack Query and local stores
> Debug Zustand stores with Redux DevTools integration for time-travel debugging and action inspection
> Write mutable-style state updates in Zustand stores with the Immer middleware for cleaner nested mutations
> Persist Zustand store to localStorage or custom storage with automatic rehydration and migration support
> Optimize Zustand re-renders with selectors, shallow comparison, useShallow, and transient subscriptions
> Split large Zustand stores into composable slice functions for modular, maintainable state management
> Create lightweight global stores with Zustand's create function for minimal-boilerplate state management
> Deploy SvelteKit to any platform by selecting and configuring the correct adapter in svelte.config.js
> Build flexible components in Svelte 5 using snippets, {@render}, typed children props, and named content areas
> Handle 404s, auth failures, and unexpected crashes in SvelteKit with +error.svelte, the error() helper, and handleError hooks
> Process HTML form submissions server-side using SvelteKit actions with progressive enhancement via use:enhance
> Fetch route data before rendering using SvelteKit's load functions — server-only, universal, streaming, and invalidation patterns
> Minimize bundle size, reduce perceived latency, and handle large datasets efficiently in SvelteKit applications
> Build SvelteKit routes using the file-system convention with +page.svelte, +layout.svelte, route groups, and dynamic segments
> Declare reactive state, derived values, and side effects in Svelte 5 using the runes API ($state, $derived, $effect, $props, $bindable)
> Intercept every request, populate locals, modify responses, and handle errors using SvelteKit's hooks.server.ts
> Choose the right state scope in SvelteKit: component-local runes, context API for subtree isolation, and module-level state for true singletons
> Share reactive state across any component tree using writable, readable, and derived stores with the Svelte store contract
> Test Svelte 5 components with Vitest and @testing-library/svelte — render, user events, store mocking, and async tick flushing
> Add enter/exit animations, list reordering motion, and spring physics to Svelte elements using built-in and custom transitions
> Directly read, write, and remove cache entries with QueryClient methods
> Chain queries that depend on each other's results using the enabled flag and useQueries
> Inspect cache state, query status, and network activity using the React Query DevTools panel
> Implement cursor-based pagination and "load more" UX with useInfiniteQuery
> Execute server-side mutations with useMutation, lifecycle callbacks, and retry configuration
> Update the UI immediately on mutation and roll back automatically if the server request fails
> Hydrate the client cache from server-fetched data using dehydrate/hydrate and HydrationBoundary
> Control cache freshness with invalidateQueries, staleTime, gcTime, and refetch strategies
> Structure query keys as type-safe factories to enable precise cache invalidation and scoped refetching
> Use useSuspenseQuery to integrate React's Suspense and error boundaries with TanStack Query
> Automate WCAG accessibility checks using axe-core with Playwright and jest-axe
> Test React components with Testing Library using user-centric queries and async utilities
> Test Svelte components with Testing Library using render, fireEvent, and waitFor
> Verify service compatibility using Pact consumer-provider contract tests
> Configure and interpret test coverage thresholds for meaningful quality signals
> LLM-judgment critique of test quality across vitest / jest / mocha / playwright / pytest. Fourth member of the craft-pipeline initiative. Per-`it`/`test` block critique with best-effort source pairing for contract-vs-implementation rubrics. Tests are often the worst-written code in a codebase precisely because the rule-based floor (coverage threshold) is so easy to clear. Emits 3-axis findings (tier × impact × confidence per ADR 0019).
> Choose the right test layer (unit/integration/E2E) and prevent flaky tests in CI
> Build maintainable test data using factory functions, builders, and faker.js
> Autonomous test-coverage backlog sweep — enumerate under-covered areas and uncovered critical paths, confirm a ranked target batch with the human in one up-front round, fan out worktree-isolated subagents that each author tests via the **real** authoring flow (tdd then test-craft), independently verify every result by added behavior-asserting tests plus a coverage delta plus all-OS CI, and hand back a batch of green test PRs for one bulk review. The fleet never auto-merges and never trusts ...
> Write integration tests that exercise real dependencies using test databases and containers
> Mock modules, functions, and timers in Vitest and Jest to isolate units under test
> Intercept HTTP requests in tests using Mock Service Worker handlers at the network level
> Measure and assert on code performance using vitest bench and timing budgets
> Write maintainable Playwright tests using page objects, fixtures, and parallel execution
> Configure Playwright test runner with fixtures, reporters, and browser contexts