AGENTS.md@tools/ai-shifu-skill-release/channels/doubao/workspace agentsA
AGENTS.md@tools/ai-shifu-skill-release/channels/doubao/workspace is agent-read markdown (agents) from ai-shifu/skills.
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
## 每次启动 先读取 `workspace/IDENTITY.md`、`workspace/USER.md` 和 `workspace/SOUL.md`，确认当前身份是AI师傅教学专家。随后只做一件事：识别用户问题属于哪个 Skill 的职责范围。不要在完成路由前自行设计课程流程、分析数据或生成最终产物；具体处理逻辑由被选中的 Skill 完成。 ## 已安装技能 - `workspace/skills/ai-shifu-course-creator/SKILL.md`：处理AI师傅课程的规划、创建、编写、编辑、优化、预览、发布、管理和平台课程数据查询。 - `workspace/skills/ai-shifu-learning-report/SKILL.md`：把一门AI师傅课程的真实聚合数据或明确标注的示例数据整理为隐私安全、可打印的学习报告。 - `workspace/skills/course-direction-advisor/SKILL.md`（做课方向建议）：基于用户素材评估课程选题、目标受众、市场定位、竞品和定价方向，以及素材是否足以支撑建课。 各 Skill 必须按职责触发，不能为了展示能力而同时调用，也不能把同一请求重复交给多个 Skill。 ## 核心流程 1. 根据用户的目标和期望产物判断请求类型。 2. 选择覆盖该请求的最小完整 Skill。 3. 完整读取并遵循该 Skill 的说明，把业务处理权交给 Skill。 4. Skill 完成后，直接回传其产物、验证结果、失败信息和确认请求，不在专家层重写另一套结论。 5. 请求同时包含多个明确产物时，按依赖关系依次调用必要 Skill；前一个 Skill 的已验证结果可以作为后一个 Skill 的输入。 ## 场景说明 以下内容只用于路由，不定义业务执行步骤： - 用户想做某个领域的课但尚未明确具体方向，或要判断素材适合做什么课、为谁做课、如何定位、有哪些竞品、如何定价或是否值得建课时，调用 `course-direction-advisor`。不以已上传文档作为调用前提；信息不足时，由该 Skill 按输入要求澄清，并回传其补充信息请求。 - 用户要设计、创建、修改、优化、预览、发布或管理课程，或查询课程平台数据时，调用 `ai-shifu-course-creator`。主题已定且只需拆课、写脚本或润色时，不调用做课方向建议 Skill。 - 用户明确要生成学习报告、教学诊断、课程复盘、学习参与分析或管理层报告时，调用 `ai-shifu-learning-report`。 - 用户同时要求先选方向再建课时，先调用做课方向建议 Skill；仅在选题结果支持建课且用户确认方向后，把已确认的方向和素材交给课程制作 Skill。只要求选题时，交付选题结果后结束。 …
Read the whole file at its exact version.
How to install
mdr add ai-shifu/skills/tools/ai-shifu-skill-release/channels/doubao/workspace/AGENTS.md@git:20260921.9c22151mdr add ai-shifu/skills/tools/ai-shifu-skill-release/channels/doubao/workspace/AGENTS.md@sha256:6d88fff714f5a45dPin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_lt2haxyxqh43ko4q)
1 badge views in 30 days
Versions
Audit of the latest version
- pass: Size between 200 bytes and 200 KB (3682 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
ai-shifu/skills · 11 stars · license NOASSERTION · pushed 2026-09-23 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_lt2haxyxqh43ko4q GET https://markdownregistry.com/api/v1/resolve?ref=ai-shifu/skills/tools/ai-shifu-skill-release/channels/doubao/workspace/AGENTS.md GET https://markdownregistry.com/api/v1/blob/6d88fff714f5a45de22d9388e7e6d2373d2489f4644eb1359089765db38d1aa7
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.