AgentSkillsCN

generate-pathlab-map

为独学、自主学习、以天为单位的 PathLab 探索活动生成严格的 PathLabMapStructure JSON 文件，最终以“继续/暂停/退出”的决策作为终点。适用于用户要求创建 PathLab 地图 JSON、地图草稿，或必须严格遵循精确的字段结构、每日顺序安排、节点与边的完整性，以及仅限 JSON 输出约束的课程地图时使用。

SKILL.md
--- frontmatter
name: generate-pathlab-map
description: Generate strict PathLabMapStructure JSON for solo, self-paced, day-based PathLab explorations that end with a Continue/Pause/Quit decision. Use when a user asks to create PathLab map JSON, map drafts, or curriculum maps that must follow exact schema fields, day sequencing, node/edge integrity, and JSON-only output constraints.

Generate PathLab Map

Generate a single valid JSON object that matches PathLabMapStructure exactly.

Workflow

  1. •Extract required inputs from the user prompt.
  2. •Build a mostly linear day-based map with 5, 6, or 7 days.
  3. •Ensure each node starts with an action, includes low-effort signal checks, and captures proof of action where possible.
  4. •Validate key consistency across nodes, edges, and path.days[].node_keys.
  5. •Output only parseable JSON with no markdown or commentary.

Output Contract

  • •Output one JSON object only.
  • •Include top-level keys: seed, nodes, path; include edges when used.
  • •Do not add extra top-level keys.
  • •Do not invent fields not present in the structure.
  • •Keep all required fields non-empty.
  • •Allow content_url only inside content items.

Structure Rules

Use this exact object structure:

  • •seed: { "title", "description", "slogan" }
  • •nodes: object keyed by node key; each value includes:
    • •title
    • •instructions
    • •node_type: "learning" | "text" | "comment" | "end"
    • •position: { "x": number, "y": number }
    • •content: array of at least 1 item with:
      • •content_type: "text" | "video" | "canva_slide" | "image" | "pdf" | "resource_link"
      • •content_body: non-empty string
      • •content_url: optional
    • •assessments: array of at least 1 item with:
      • •type: "quiz" | "text_answer" | "file_upload" | "image_upload" | "checklist"
      • •prompt: non-empty string
      • •isGraded: boolean
      • •pointsPossible: number
  • •edges (optional): array of { "source_key", "destination_key" }
  • •path:
    • •total_days: exactly 5, 6, or 7
    • •days: array length equals total_days
    • •each day:
      • •day_number: sequential from 1
      • •title
      • •context_text
      • •reflection_prompts: 1-2 short one-sentence prompts
      • •node_keys: 1-2 keys; all keys must exist in nodes

Pedagogy Rules

  • •Every node begins with an action task.
  • •Every node includes at least one low-effort signal: quiz or checklist.
  • •Prefer file_upload or image_upload to capture proof of action.
  • •Allow at most one text_answer per node, one sentence max.
  • •Keep node workload in the 15-45 minute range.
  • •Avoid front-loading explanation; favor immediate action.

Graph And Position Rules

  • •Keep flow mostly linear.
  • •Start near { "x": 100, "y": 100 }.
  • •Increase x by about 200-250 per sequential node.
  • •Keep y consistent for linear flow; vary only for optional branches.
  • •Ensure all edge references exist in nodes.

Final Day Rule

  • •Include synthesis and exact prompt text:
    • •Based on this week, what's your next step? Continue / Pause / Quit
  • •Use node_type: "end" for final-day decision node when appropriate.

Silent Validation Before Output

  • •JSON parses cleanly.
  • •Day count equals path.total_days.
  • •Day sequence is valid and starts from 1.
  • •Reflection prompt count/day is 1 or 2.
  • •Every node_key reference exists.
  • •Final-day decision prompt exists exactly as required.

References

  • •For the exact interface shape used in this repo, see:
    • •references/pathlab-map-structure.md