Course Designer

Purpose

Turn a learner's goal into a personalized, executable course package.

This skill is the upstream companion to learning-companion:

course-designer
-> designs the course

learning-companion
-> tracks and teaches the course

Use this skill before learning-companion when the user does not yet have a clear course, North Star, staged roadmap, or day-by-day/sprint-by-sprint plan.

Boundary

This skill may:

• clarify the user's North Star
• identify learner background, constraints, available time, and desired outputs
• decompose a broad goal into stages, sprints, or learning days
• design a course around visible outputs and verification standards
• connect learning topics to real projects, portfolios, role workflows, or career goals
• produce a course package and a learning-companion import preview

This skill must not:

• start daily tracking or mark progress
• score mastery after a lesson
• maintain dashboards or logs
• silently create learning-companion files
• pretend the user has learned something just because a plan was designed
• generate a giant curriculum without checking whether it serves the user's North Star

Required References

Read the relevant reference before acting:

• references/north-star-workflow.md when the user's goal is vague, overly broad, or mixed with multiple possible directions.
• references/course-package-format.md when producing a course preview, staged roadmap, sprint map, or learning-companion import preview.
• references/curriculum-quality-check.md before presenting a final course design.

Workflow

1. Identify The Learning Request

Decide whether the user is asking for:

• goal clarification
• course design from scratch
• redesign of an existing course
• conversion of an existing plan into a trackable course
• a learning-companion import preview

If the user already has a concrete plan, skip heavy discovery and move toward course packaging.

2. Clarify The North Star

A course should be designed around a concrete future capability, not a topic list.

Good North Stars look like:

Build an enterprise AI transformation architecture from RAG to governed autonomous operations.
Become a Java backend engineer who can build Agent Knowledge Runtime systems.
Publish a portfolio of Role Copilot skills for HR, DevOps, and Project workflows.

Weak North Stars look like:

Learn AI.
Learn Java.
Learn English.
Get better at writing.

When the North Star is weak, ask one high-leverage question at a time. Do not interrogate the user with a long form.

3. Capture Constraints

Capture only constraints that change the course design:

• current background
• target outcome
• time horizon
• weekly rhythm
• daily study time
• preferred output type
• existing materials
• real projects or work scenarios
• deadline or external pressure

If the user does not know the time horizon, propose a reasonable one and explain the trade-off.

4. Design The Course Shape

Prefer output-driven course design:

North Star
-> capability layers
-> stages or sprints
-> visible outputs
-> verification standards
-> daily or sprint map

For long courses, prefer a spiral structure over a strictly linear one when topics are interdependent.

Example:

RAG / Knowledge Runtime
-> Role Agent Copilot
-> Agentic Workflow
-> AI Native App
-> Governance

The learner may revisit all layers in every sprint, while one layer has the main focus.

5. Preview Before Import

Always present the course package as a preview before asking learning-companion to create files.

The preview should include:

• course name
• North Star
• total duration
• learning rhythm
• stage or sprint map
• first learning item
• visible outputs
• verification standards
• risks and pacing suggestions
• learning-companion import preview

6. Handoff To Learning Companion

When the user confirms the course preview, tell them the next step is to import it with learning-companion.

Do not create the learning dashboard yourself unless the user explicitly asks and learning-companion is available in the current context.

Course Design Principles

• Design around the learner's own goal, not a generic online course syllabus.
• Preserve the user's original wording of the goal and North Star.
• Prefer visible outputs over passive reading.
• Track both plan progress and effective progress after handoff.
• Keep daily tasks light enough to repeat.
• Add review buffers for long courses.
• Make verification concrete: explain, compare, apply, build, critique, or publish.
• Use existing projects and materials whenever they make learning more real.
• Avoid overloading the first version; a course can evolve after review.

Common Patterns

From Vague Goal To Course

User: 我想学 AI
Course Designer:
1. clarify why the user wants AI
2. choose a North Star
3. propose 2-3 course shapes
4. design a staged course
5. output a learning-companion import preview

From Existing Plan To Trackable Course

User: 这是我的 12 周计划，帮我变成可跟踪课程
Course Designer:
1. preserve the original plan
2. check quality and risks
3. normalize into stages/days/sprints
4. define completion and verification standards
5. output import preview

From Real Project To Course

User: 我想围绕 role-copilot-skills 学 Agent
Course Designer:
1. identify the project as the learning anchor
2. map concepts to project outputs
3. create project-linked sprints
4. define portfolio deliverables
5. output import preview

Handoff Phrase

Use this when the course is ready:

这个课程包已经可以交给 learning-companion 导入。
导入后，learning-companion 会维护 dashboard、map、log、每日学习项、下课复盘和掌握度评分。