0xbigboss/claude-code
4 skills7.6k installs196 starsGitHub
Install
npx skills add https://github.com/0xbigboss/claude-codeSkills in this repo
1React Best PracticesReact Best Practices is an agent skill backed by a patterns reference that shows how solo builders should structure modern React code without common foot-guns. It targets indie developers shipping SaaS dashboards, marketing sites, or browser extensions where Claude or Cursor generates components faster than a human can audit every Effect. The material contrasts bad habits—storing derived full names in state via useEffect, recomputing filters in effects, resetting comments with effects on userId—with idiomatic fixes: calculate during render, memoize heavy work with useMemo, and remount subtrees with key. Agents load the reference when they need verbatim implementations, not vague “follow React docs” advice. Use it throughout Build as you scaffold features, and again in Ship when reviewing agent-written UI for subtle state loops. The skill is pattern-first rather than framework-version-specific, so it stays useful across Create React App, Next.js client components, and Vite SPAs as long as hooks are in play.2.8kinstalls2Typescript Best PracticesTypeScript Best Practices is an agent skill for solo and indie builders who want their coding agent to write and review TypeScript and JavaScript with the same discipline a senior engineer would expect: discriminated unions instead of boolean soup, branded IDs so UserId and OrderId cannot be swapped, and const assertions for finite role or status unions. The skill is intentionally narrow—it covers language-level patterns only, not framework mechanics. When you or your agent open React files, the skill tells you to pair it with react-best-practices so hooks, effects, and component boundaries stay in the right document. Use it during feature work, refactors, and pre-merge reads of typed modules so invalid states fail in the type checker rather than in production. It fits Claude Code repos that already encode standards in CLAUDE.md and want those rules applied file-by-file without re-pasting the whole style guide every session.2.5kinstalls3Python Best PracticesPython Best Practices is a journey-wide agent skill for solo builders shipping typed Python APIs, CLIs, and agent tooling. It encodes how to model domains safely: immutable frozen dataclasses, discriminated Success/Failure unions with exhaustive matching, NewType wrappers so IDs do not cross wires, and Protocol-based interfaces for testable boundaries. Invoke it whenever you or the agent touch Python modules or packaging files so refactors stay consistent with a type-first, functional style and deliberate error paths rather than silent mutation or stringly-typed APIs. It is intentionally narrow—Python language idioms from the parent repo’s engineering guide—not a full linter replacement. Pair it with tests and CI in Ship, and with monitoring fixes in Operate, without re-tagging the skill each time.1.6kinstalls4Web Fetchweb-fetch is a small Bun and TypeScript fetch pipeline bundled as an agent skill for solo builders who need trustworthy page text without hand-copying from bloated marketing sites. It requests a URL, parses HTML, scores candidate containers (article, main, ARIA main, common content class/id selectors), picks the longest text block, prunes chrome elements that pollute LLM context, then converts the remainder to markdown. That pattern matters when you are validating an idea against competitor pages, skimming API reference sites, or archiving release notes during Build or Ship—you still install it from the research phase because ingestion precedes implementation. It is not a headless browser for SPAs that require JavaScript rendering; it is a fast integration for document-heavy pages. Pair it with your agent’s normal browsing tools when you only need the substantive body copy in-repo.858installs