📄 中文摘要
在使用 Claude Code 或 OpenCode 的过程中,开发者常常会面临技能和命令数量激增的问题。最初,开发者可能只拥有少量技能,这些技能在项目中表现良好。然而,随着项目的增加和团队成员共享技能,文件数量迅速增加,导致文件分散在不同目录中,难以找到所需的文件。此时,AI 代理可能缺乏必要的上下文,或者被不必要的信息淹没。Agent-i-Kit 旨在解决这一问题,专注于技能和工具的发现,而非单纯的存储。通过优化技能管理,开发者能够更高效地使用 AI 工具。
📄 English Summary
Your AI Agent's Skill List Is Getting Out of Hand
Developers using Claude Code or OpenCode often encounter the issue of an overwhelming number of skills and commands. Initially, they may start with a few skills that work well for their projects. However, as they add more skills for new projects and incorporate those shared by teammates, the number of files increases rapidly. This leads to files scattered across directories, making it difficult to locate the necessary ones. Consequently, the AI agent may lack essential context or be overloaded with irrelevant information. Agent-i-Kit is designed to address this problem by focusing on the discovery of skills and tools rather than mere storage. By optimizing skill management, developers can utilize AI tools more efficiently.
Powered by Cloudflare Workers + Payload CMS + Claude 3.5
数据源: OpenAI, Google AI, DeepMind, AWS ML Blog, HuggingFace 等