codebase-knowledge-graph

What problem does it solve?

This Skill prevents unsafe or redundant edits by forcing an agent to understand a repository’s relationships and contracts rather than treating it as a bag of files.

Core Features & Use Cases

• Whole-system map building: Models modules, files, classes, and functions as typed nodes with semantic edges (calls, imports, persistence, tests) to answer what is load-bearing and what will ripple.
• Foundational vs incidental separation: Distinguishes contract-defining code from leaf or one-off glue to decide what can be extended vs what should not be casually modified.
• Ripple analysis before editing: Inspects upstream callers, downstream dependencies, adjacent implementations, and config/schema/storage boundaries to reduce breaking changes.
• Reuse-first change planning: Guides agents to resolve reuse before creation by locating canonical seams and existing patterns.
• Paired feature integration: Uses graph discovery to identify where variant logic should integrate and what contracts might break when adding feature dimensions.

Quick Start

Invoke codebase-knowledge-graph to map the target area, distinguish foundational from incidental code, and run ripple analysis before you propose any code changes.