AgentSkillsCN

concept-graph

根据教程规范构建概念图（节点 + 先修关系边），并以 `outline/concept_graph.yml` 文件保存。 **触发条件**：概念图、先修关系图、依赖关系图、概念图、先修关系。 **适用场景**：在教程流水线的结构化阶段（C2），你需要将教程知识点拆解为可排序的依赖关系图（在撰写教程散文之前）。 **跳过条件**：若尚未形成教程规范（例如缺少 `output/TUTORIAL_SPEC.md` 文件）。 **网络要求**：无需联网。 **防护措施**：仅进行结构化梳理，避免撰写冗长的散文段落。

SKILL.md
--- frontmatter
name: concept-graph
description: |
  Build a concept graph (nodes + prerequisite edges) from a tutorial spec, saving as `outline/concept_graph.yml`.
  **Trigger**: concept graph, prerequisite graph, dependency graph, 概念图, 先修关系.
  **Use when**: tutorial pipeline 的结构阶段（C2），需要把教程知识点拆成可排序的依赖图（在写教程 prose 前）。
  **Skip if**: 还没有 tutorial spec（例如缺少 `output/TUTORIAL_SPEC.md`）。
  **Network**: none.
  **Guardrail**: 只做结构；避免写长 prose 段落。

Concept Graph (prerequisites)

Goal: represent tutorial concepts as a prerequisite DAG so modules can be planned and ordered.

Inputs

  • •output/TUTORIAL_SPEC.md

Outputs

  • •outline/concept_graph.yml

Output schema (recommended)

A minimal, readable YAML schema:

  • •nodes: list of {id, title, summary}
  • •edges: list of {from, to} meaning from is a prerequisite of to

Constraints:

  • •Graph should be acyclic (DAG).
  • •Prefer 10–30 nodes for a medium tutorial.

Workflow

  1. •Read output/TUTORIAL_SPEC.md and extract the concept list implied by objectives + running example.
  2. •Normalize each concept into a node with a stable id.
  3. •Add prerequisite edges and verify the graph is acyclic.
  4. •Write outline/concept_graph.yml.

Definition of Done

  • • outline/concept_graph.yml exists and is a DAG.
  • • Nodes cover all learning objectives from output/TUTORIAL_SPEC.md.
  • • Node titles are specific (not “misc”).

Troubleshooting

Issue: the graph looks like a linear list

Fix:

  • •Add intermediate prerequisites explicitly (e.g., “data model” before “evaluation protocol”).

Issue: cycles appear (A → B → A)

Fix:

  • •Split concepts or redefine edges so prerequisites flow in one direction.