Taste-skill is not a software application in the traditional sense; it is a portable library of AI-agent 'skills' - extremely detailed Markdown instruction files (SKILL.md) distributed via the `npx skills add` CLI (Vercel agent-skills spec) and a .claude-plugin marketplace. The repo's product is the prompt content itself: 13 frontend design-direction skills that agents like Claude Code, Codex, and ChatGPT load to avoid generic 'AI slop' output. The flagship is `taste-skill` v2 (`design-taste-frontend`), a ~5000-line instruction document that codifies anti-pattern bans (zero em-dashes, no section-number eyebrows, no three-equal-card rows), a brief-inference protocol, design-system mapping to official packages (Fluent, Carbon, Polaris, GOV.UK), canonical GSAP/scroll-reactive code skeletons, and a hard pre-flight checklist with mechanical pass/fail criteria. It is complemented by image-generation-only skills (brandkit, imagegen-frontend-web/mobile) that produce reference comps for downstream coding agents, plus variant skills (brutalist, minimalist, soft, stitch, output-enforcement). The repository also contains a `research/laziness/` folder documenting LLM output-truncation causes and remediation, and a small set of Node/Sharp image-processing scripts used to build the README assets. Those scripts contain one genuinely valuable, extractable algorithm: a flood-fill outer-background removal with auto-cropping to content bounds. The scripts are one-off (hardcoded `C:/Users/User/Downloads` source paths, tightly coupled to specific sponsor assets), so they are not drop-in reusable without adaptation, but the background-removal technique itself is portable. This is a content/instruction repository rather than a codebase; software-engineering quality bars (tests, CI) do not apply. Its craftsmanship lives in the rigor, specificity, and mechanical enforcement of the prompt files, which is exceptional for the prompt-engineering domain. The active development cadence, current dependency references (Tailwind v4, Motion/react, React Server Components), and 92k-star traction confirm it is the leading project in its niche.
Onlook is a browser-based visual editor for building and modifying Next.js + Tailwind CSS websites. It provides a Figma-like canvas where users can directly manipulate React components, with changes written back to source code via AST-based transformation. The application integrates an AI chat assistant that can generate code, edit files, and run terminal commands through a set of well-defined tools (e.g., search-replace, file edit, bash). The monorepo architecture separates concerns into packages for AI agents, code parsing, font management, UI components, database access (Drizzle + Supabase), git operations, and sandbox management (CodeSandbox). It also includes a complete billing system (Stripe), GitHub integration for project import, custom domain management, and a publish/deployment pipeline. The codebase demonstrates strong engineering practices with type-safe tRPC APIs, MobX state management in the editor, extensive use of Radix UI primitives, and a consistent monorepo toolchain (Bun workspaces, ESLint, Prettier). While the editor engine is somewhat monolithic, individual packages like the AST parser, font helpers, and utility library are well-isolated and testable.
VibeTunnel is a production-grade, cross-platform remote terminal access platform. It allows users to connect to terminal sessions from a browser, macOS native app, or iOS app. The Node.js/TypeScript server manages PTY sessions via a custom forked node-pty, supports session recording in asciinema format, and integrates with multiple tunneling services (Tailscale, ngrok, Cloudflare, Pinggy, Pangolin) for remote access. A Rust forwarder handles terminal title sanitisation, and the CLI wrapper (`vt`) enables transparent session creation. Authentication supports PAM, SSH keys, JWT, Tailscale identity, and local bypass. The macOS app (SwiftUI) provides menu bar control, system permission management, Git worktree follow-mode, and Sparkle-based updates. The web frontend (Lit + Tailwind CSS) offers a full terminal UI with quick keys, file browser, and push notifications. The project demonstrates strong engineering with extensive test coverage (unit, integration, Playwright e2e) across all platforms, though multi-platform duplication and vendored native dependencies introduce some maintenance overhead.
This repository is a curated collection of design system analyses for over 70 brands, each documented in a standardized DESIGN.md file. The analyses reverse-engineer public websites to extract design tokens (colors, typography, spacing, components) and present them in a machine-readable YAML front matter format alongside human-readable prose sections. The repo serves as a reference dataset for design system research, AI-assisted UI generation, and cross-brand comparison. It is not a software application or a traditional data pipeline; rather, it is a hybrid content repository that blends design research with structured data extraction. The project is maintained by VoltAgent and includes a commercial layer for requesting new analyses. While the data is well-structured and consistent, the analyses are static snapshots that may drift from live sites, and some files contain placeholder markers or typos.
This repository powers a multi-domain interactive programming tutorial platform (learnpython.org, learnjavaonline.org, etc.). The core is a Flask web application that serves Markdown-based lessons, provides an in-browser code editor via a vendored CodeMirror 3.x, and executes user code through the Ideone SOAP API. It supports 18+ programming languages with per-domain configuration (editor modes, language IDs, default code). A newer 'vibe coding' feature uses a local browser-based AI model (SmolLM2-360M) to generate code from prompts. The application is deployed via Docker and relies on environment variables for credentials and domain selection. Architecturally, the app is a monolith with a thin Flask layer, static asset serving, and template rendering via Jinja2. Tutorial content is stored as structured Markdown files with consistent headings (Tutorial, Exercise, Tutorial Code, Expected Output, Solution) and is parsed and rendered at runtime. The system is highly internationalized, with 20+ language translations of the core Python tutorials. User progress tracking and authentication are stubbed (501 responses). The codebase includes vendored copies of Python-Markdown, the suds SOAP library, and many CodeMirror modes/addons. The platform has grown organically with significant technical debt: outdated vendored libraries (CodeMirror v3, suds, Markdown), Python 2/3 compatibility hacks, incomplete translations, many empty tutorial pages, and a fragile translation pipeline dependent on a specific OpenAI model. The lack of automated testing beyond basic coverage and the reliance on a third-party code execution API (Ideone) are notable risks. However, the domain configuration pattern in constants.py and the Ideone client wrapper are well-structured and could be reused in similar projects.