Concept Map -- Domain Knowledge

Concept Map -- Domain Knowledge

Reference knowledge for generating concept maps that visualize how all concepts within a given argument relate to each other and to the training plan's main concepts (Core Principles and Key Concepts).

What This Skill Does

Takes an argument (a topic, domain, sub-goal, or free-form subject) and produces a structured concept map that:

1. Maps all concepts within the argument -- principles, innovations, applications, prerequisites
2. Cross-references the training plan -- every node that corresponds to a training plan Core Principle or Key Concept is annotated with its SG number
3. Shows relationships between concepts using the standard relationship vocabulary
4. Highlights coverage gaps -- concepts in the map that are NOT covered by any sub-goal in the training plan

---

Inputs

The skill requires two things:

1. Argument: the topic or domain to map (provided by the user)
2. Training plan context: the active training plan file in the current directory (read automatically)

If no training plan file is found, produce the concept map standalone (without SG annotations) and note that no plan was found.

---

Step-by-Step Process

Step 1: Read the Training Plan

1. Use Glob to find training-plan-*.md files in the current directory
2. Read the plan to extract:
   • Core Principles table (principle name, statement)
   • Sub-Goals table (SG number, axis, sub-goal statement, principle mapping)
   • Prerequisite Graph
3. Build an index: {concept_name -> SG-N} for every Core Principle and Key Concept sub-goal

Step 2: Analyze the Argument

Decompose the argument into its constituent concepts:

1. Identify the core concept (the argument itself or its central idea)
2. Extract principles -- foundational truths the argument rests on
3. Extract innovations -- key breakthroughs or contributions within the argument
4. Extract applications -- practical uses or derived outcomes
5. Extract prerequisites -- what must be known before understanding the argument
6. Identify cross-links -- how concepts within the argument relate to each other

Step 3: Cross-Reference with Training Plan

For each concept extracted in Step 2:

1. Check if it matches (exact or semantic) a Core Principle or Key Concept in the plan
2. If matched, annotate the node with [SG-N]
3. If NOT matched, mark as (unmapped) -- this concept exists in the domain but is not covered by the plan

Step 4: Build the Concept Map

Use the standard concept map format (see Format section below) with these additions:

• SG annotations on mapped nodes: **[Concept Name]** (principle) [SG-3]: description
• Unmapped markers on uncovered nodes: **[Concept Name]** (principle) [unmapped]: description
• Plan coverage summary after the map

Step 5: Generate Coverage Analysis

After the map, produce:

## Plan Coverage Analysis

**Mapped concepts**: N of M total concepts are covered by sub-goals
**Unmapped concepts**: list of concepts not in the training plan
**Suggestion**: whether unmapped concepts warrant new sub-goals or are acceptable gaps

---

Concept Map Format

Maps are markdown nested lists. Each node has a bold name, a type in parentheses, an optional SG annotation in brackets, and a one-line description after a colon.

Node Types

Type	Meaning	Usage
(core)	Central concept being mapped	Exactly one per map
(principle)	Foundational truth or constraint	The 3-5 load-bearing ideas
(innovation)	Conceptual breakthrough or key contribution	Creative leaps that built the concept
(application)	Practical use or derived concept	Where theory meets reality
(prerequisite)	Required prior knowledge	What you need before this concept

Relationship Labels

Label	Meaning	Direction
rests on	Built on these principles	core -> principle
requires	Needs these prerequisites	core -> prerequisite
led to	Produced these innovations	principle -> innovation
enables	Enables these applications	innovation -> application
generalizes	Generalization of the child	parent -> child
is a special case of	Specialization	child -> parent
contrasts with	In tension or opposition	sibling <-> sibling
equivalent to	Different formulations of same idea	sibling <-> sibling

Structure Rules

1. Start from the core concept (the argument) at the top level
2. Group children by relationship type: requires first, then rests on, then enables
3. Nest innovations under their parent principle using led to
4. Add applications showing practical impact
5. Add prerequisites for context
6. Keep descriptions to one line
7. Annotate every node that maps to a training plan sub-goal with [SG-N]
8. Mark nodes not covered by the plan with [unmapped]

Example (with Training Plan Cross-Reference)

## Concept Map: Cost Estimation Methods

- **Cost Estimation Methods** (core): Techniques for predicting project costs at various accuracy levels
  - requires:
    - **Work Breakdown Structure** (prerequisite) [SG-2]: Hierarchical decomposition of project scope into deliverables
    - **Basic Accounting** (prerequisite) [unmapped]: Understanding of cost categories and financial statements
  - rests on:
    - **Cost traceability** (principle) [SG-3]: Every cost is either direct (traceable to scope) or indirect (project-wide)
      - led to:
        - **Unit-rate estimation** (innovation) [SG-6]: Quantity x unit rate = direct cost for each line item
          - enables:
            - **Definitive estimates** (application) [SG-9]: AACE Class 1-2 estimates with -5%/+10% accuracy
        - **Factored estimation** (innovation) [SG-7]: Scale from known costs using capacity or complexity factors
          - enables:
            - **Screening estimates** (application) [unmapped]: Quick order-of-magnitude for early feasibility
    - **Accuracy improves with definition** (principle) [SG-4]: Estimate accuracy is a function of project definition level
      - led to:
        - **AACE classification system** (innovation) [SG-8]: 5-class system mapping definition % to accuracy range
  - enables:
    - **Bid evaluation** (application) [unmapped]: Comparing contractor bids against independent estimates
    - **Budget authorization** (application) [SG-10]: Securing funding based on estimate confidence level

## Plan Coverage Analysis

**Mapped concepts**: 8 of 11 total concepts are covered by sub-goals
**Unmapped concepts**: Basic Accounting (prerequisite), Screening estimates (application), Bid evaluation (application)
**Suggestion**: Basic Accounting is a prerequisite gap -- verify learner has this knowledge. Screening estimates and Bid evaluation are applications outside the plan's scope -- acceptable gaps unless the learner's role requires them.

---

Depth Scaling

Depth	Total Nodes	Levels	When to Use
shallow	5-8	2	Quick overview, single sub-goal scope
standard	10-15	3	Default -- one topic or related group of sub-goals
deep	15-25	4	Full training plan scope, complex multi-principle topic

Default: standard (10-15 nodes).

If the argument spans the entire training plan, use deep depth automatically.

---

Whole-Plan Map Mode

When the argument is the training plan itself (e.g., "map the whole plan", "show all concepts"), produce a map that:

1. Uses the plan's main goal as the (core) node
2. Places every Core Principle as first-level (principle) nodes with [SG-N]
3. Nests every Key Concept under its parent Core Principle with [SG-N]
4. Shows the prerequisite graph from the plan as (prerequisite) nodes
5. Adds Verification sub-goals as (application) nodes
6. Uses deep depth (15-25 nodes)

This gives a complete structural overview of the training plan's concept architecture.

---

Integration with Other Skills

• concept-explainer: If the user wants a full explanation of a node in the map, invoke concept-explainer for that node
• goal-decomposer: The prerequisite graph from goal-decomposer feeds into this skill's cross-reference
• ttt-session: Concept maps generated during sessions (Phase 3.5) use the same format but are scoped to the session's teaching content
• concept-library: Concept files exported from sessions can be referenced to enrich the map with previously taught content