What problem does it solve? Business rules are scattered across code, DDL, and logs, making it hard to distinguish stable business truths from incidental implementation details. This Skill bootstraps a project's domain knowledge base by scaffolding draft knowledge entries that human owners review and promote. ## Core Features & Use Cases - Two-phase initialization: Phase A writes the target project's agent entry files (CLAUDE.md / AGENTS.md) with tech stack and layering facts; Phase B generates draft business-truth skeletons (formula, flow, state, rule, concept, term) per module. - Mechanical scaffolding scripts: Provides scan, gaps, new, check, check-paths, and sync-modules commands that list class-name clues, detect coverage gaps, and validate module manifests without auto-extracting content into the knowledge base. - DDL baseline generation: Dumps real database schemas into an impl/ddl-baseline.md so SQL writing never relies on guessed table names. - Use Case: A knowledge-base owner says "initialize the yoooni project's knowledge graph" and the Skill runs gap analysis, drafts module-by-module candidates for human judgment, and produces validated draft Markdown entries ready for owner review. ## Quick Start Ask the AI to initialize the business knowledge graph for your project by providing the target project root path, then confirm the module priority list it proposes.
