Build Knowledge Graph

build-knowledge-graph

Translate the raw legacy code inventory into a structured, queryable knowledge graph that captures domain concepts, business rules, and business processes — stripped of all implementation artifacts.

When to use this skill

After read-legacy-code is complete and the legacy inventory files exist in _superml/legacy-inventory/.

Can also be triggered standalone if the user has:

• External documents (specifications, user manuals, requirement docs)
• Domain expert willing to describe the business in conversation
• A mix of code + documents

Ask user on activation:

"I'll build the knowledge graph from the legacy code inventory. Do you also have any additional sources to include?

1. Functional specification documents (Word/PDF/text)
2. Data dictionaries provided by business
3. Process flow diagrams or swimlane diagrams
4. A domain expert you can relay information from
5. None — proceed from code only"

What is a knowledge graph (in this context)?

Not a graph database. A structured collection of markdown files that captures:

• Domain entities — the business objects (Customer, Order, Policy, Account...)
• Business rules — the constraints, validations, calculations, and decisions
• Business processes — end-to-end flows, triggers, steps, outcomes
• Domain vocabulary — what things are called in THIS business

It is the bridge between "what the legacy code does" and "what the new system must do."

Outputs

Written to {project-root}/_superml/knowledge-graph/:

• graph.md — master knowledge graph (entities + relationships + rules index)
• entities/ — one file per domain entity
• business-rules.md — numbered rule registry
• process-flows.md — business processes
• domain-glossary.md — business vocabulary

Instructions

Read ./workflow.md and execute each step in order.