Lessons, feedback, references, and progress stay organized in an inspectable HTML workspace.
Open-source · local-first
Mastery Tutor turns a compatible AI agent into a teaching system with guided HTML lessons, practice, transfer checks, spaced review, and progress that stays in your own workspace.
Lessons, feedback, references, and progress stay organized in an inspectable HTML workspace.
Understanding is checked through independent retrieval, application, transfer, and delayed review.
Retries keep the original task and earliest error while revealing only the next useful hint.
The learner owns the `.mastery/` record; no hosted account or remote database is required.
A second cooperating Skill creates code exercises, simulations, visual explanations, and accessible fallbacks.
Packaging, engine behavior, live-agent behavior, usability, and learning outcomes are reported as separate evidence levels.
One-prompt install
Install Mastery Tutor from https://github.com/fanfanfanfanfan626/mastery-tutor. Read AI_INSTALL.md first, detect the current host, install both required Skills, preserve every .mastery learner workspace, and do not claim success until that host's verification steps pass.
Codex uses the complete plugin adapter. Other compatible Agent Skills hosts install `mastery-coach` and `mastery-tool-creator` together.
本地优先掌握式学习导师
Mastery Tutor 把引导式 HTML 课程、动手练习、迁移检验、间隔复习和可检查的学习记录放进同一套工作流。学习数据留在你选择的本地工作区,不需要注册远程账号。
阅读中文说明