prototypebench/prototypebench
Open benchmark for AI coding agents on full-stack feature shipping (React+Vite+Tailwind/FastAPI+SQLModel). 71 PR-mined tasks · 32k tests · execution-based scoring (pytest+Playwright) · no LLM-as-judge.
Open benchmark for AI coding agents on full-stack feature shipping (React+Vite+Tailwind/FastAPI+SQLModel). 71 PR-mined tasks · 32k tests · execution-based scoring (pytest+Playwright) · no LLM-as-judge.
npx skills add prototypebench/prototypebenchOpen benchmark for AI coding agents on full-stack feature shipping (React+Vite+Tailwind/FastAPI+SQLModel). 71 PR-mined tasks · 32k tests · execution-based scoring (pytest+Playwright) · no LLM-as-judge.
在 任意Agent（如Claude Code） 中使用 GPT Image 2生成图像，复用你现有的 ChatGPT Plus 或 Pro 订阅；无需单独的 OpenAI API密钥，也没有按图计费。通过本地 Codex CLI 支持文生图、图生图编辑、风格迁移和多参考图组合。
your archive + Muse, the AI that actually knows you. Self-hosted, no npm, no bundler, no build step. Powered by Claude Agent SDK.
Reusable skills for document parsing and agent workflows, turning PDFs, DOCX, PPTs, and images into LLM-ready Markdown.
🧠 Enable AI coding agents to adopt product management skills and build user-focused software efficiently.
Automate Reddit tasks using your actual browser and account with support for AI agents and the SKILL.md format.
Production Go toolkit for MCP servers — 72 packages, middleware chains, RBAC, FinOps, circuit breakers, 85%+ coverage