boluopai.local-teacher
Teach sources, review concepts and develop learning through authorized local
work. This Skill emphasizes sustained explanation, source preparation, design,
experiments and project adjustment. It shares the full Course method with
web-teacher; it is not a simplified practical clone. The exact identifier is
local-teacher.
Examples: “Explain this chapter coherently without quizzes”; “Recover the conceptual structure of this argument”; “Build this small experiment, carry the formal detail, and let me predict or interpret its result”; “Prepare a glossary, guide and applicable training for this source.”
Install the linked group
Download complete sibling directories, preserving their names and resources:
| Companion | Relationship and download |
|---|---|
| web-teacher | Required shared teaching/Course method and all branches; download. Local Teacher reads this owner directly. |
| yinyang | Shared modeling, epistemic and participant-relationship account; download. Follow its public scope; do not infer a recipient portrait from this example. |
| task-inquiry | Useful support for consequential ambiguity and competing interpretations; download. Teacher retains learning/practice judgment. |
| research | Source investigation, factual uncertainty and evidence comparison; download. It does not install a search provider. |
Install in your host's supported Skill directory and load SKILL.md.
The interface manifest in agents/openai.yaml is for compatible Codex hosts;
other hosts can read the same entrypoint and linked resources. No CLI, daemon,
hidden coordination layer or account is required by Teacher itself. Text lessons
need the actual source and conversation. Saved documents/State need authorized
file access; local experiments need the project's real tools and authority.
Visuals and source readers use capabilities actually available to the host.
Before Course work, configure recipient-owned sources, paths and State through the shared configuration guide. Do not put personal records in installed/public packages. A known-source isolated question can be answered without inventing a course or progress store. Teacher carries articulable reasoning, source recovery, calculation and execution details; the learner's selected intuitive judgment and revision remain attributable. An assisted contribution can be evidence; a model answer or passing run alone does not establish understanding. There is no mastery gate for onward explanation.
This public adaptation preserves the local method and consumes the shared Web resources through public relative links. Private host paths, learner/corpus records and current research history are excluded. Original package material is MIT licensed; companion packages and supplied sources retain their own licenses. There is no third-party implementation bundled here.
中文说明
本 Skill 在本地讲解原始材料、复习概念，并通过已授权的实际工作支持学习。侧重持续讲解、材料准备、设计、实验和项目调整，与 web-teacher 共用完整
Course 方法，不是弱化的实践副本。准确标识为 local-teacher。
示例：“连贯讲解这一章，不做测验”；“恢复这个论证的概念结构”； “搭建这个小实验，你承担形式细节，让我预测或解释结果”； “为这份材料准备术语表、指南及适用训练”。
| 配套 Skill | 职责与取得方式 |
|---|---|
| web-teacher | 必需的完整共享教学／Course 方法及分支；下载完整包，本地 Teacher 直接读取这个方法负责人。 |
| yinyang | 共享建模、认识论及参与者关系；下载。遵守公开范围，不把示例视为接收方画像。 |
| task-inquiry | 重要歧义与竞争解释的支持；下载。Teacher 保留学习和练习判断。 |
| research | 来源调查、事实不确定性和证据比较；下载。它不安装搜索服务。 |
将完整目录作为同级目录安装到宿主支持的 Skill 位置，加载 SKILL.md。兼容 Codex 的宿主可使用 agents/openai.yaml；其他宿主读取相同入口及资源。
Teacher 本身不需要 CLI、守护进程、隐藏协调层或账号。文本教学需要实际来源与对话，保存材料／State 需要授权文件访问，实验需要项目真实工具和权限。图示和材料阅读器使用宿主实际可用能力。
Course 工作前按共享配置指南绑定接收方自己的材料、路径和 State，不将私人记录写入安装包或公开包。孤立的已知来源问题无需虚构课程或进度库。Teacher 承担可表达的推理、材料恢复、计算与执行细节；学习者选定的直觉判断和修正保留可归属性。受辅助的贡献可作证据，模型答案或程序运行通过不能独自证明理解，也不以掌握证明阻挡后续讲解。
此公开改写保留本地方法，通过公开相对链接消费共享 Web 资源；排除私人主机路径、学习者／材料库记录及当前研究历史。原创内容采用 MIT 许可，配套包与提供的来源保留各自许可，本包不附第三方实现。