Agent Skills: Module Planner

Module Planner

Goal: turn a concept DAG into a teachable module sequence with clear objectives and outputs.

Inputs

• outline/concept_graph.yml

Outputs

• outline/module_plan.yml

Output schema (recommended)

• modules: ordered list of modules
  • id, title
  • objectives (3–6 measurable bullets)
  • concepts (node ids from outline/concept_graph.yml)
  • outputs (what the learner produces)
  • running_example_steps (optional but recommended)

Workflow

1. Read outline/concept_graph.yml and topologically sort concepts.
2. Cluster concepts into modules (keep module scope coherent; avoid “misc”).
3. For each module:
   • write measurable objectives
   • define concrete outputs (code/artifact)
   • specify how the running example advances
4. Write outline/module_plan.yml.

Definition of Done

• [ ] outline/module_plan.yml exists and modules are ordered by prerequisites.
• [ ] Every module has objectives + outputs.
• [ ] Every concept node from outline/concept_graph.yml is covered by at least one module.

Troubleshooting

Issue: modules are too many / too granular

Fix:

• Merge adjacent modules with shared prerequisites; target ~5–12 modules for most tutorials.

Issue: objectives are not verifiable

Fix:

• Rewrite objectives so exercise-builder can attach a concrete exercise for each module.