war851/AI-Governance-Architecture

# Technical Summary Implements a database-centric orchestration layer where structured process definitions (steps, skills, functions, data transformations) live in config tables rather than model weights, with the LLM operating as a stateless semantic engine constrained by pre/post execution gates. Includes four composable design skills—value chain mapping, data evolution tracing, TDD golden examples, and architecture synthesis—that produce 17 YAML-enveloped artifact types across a `draft → detailed → confirmed` lifecycle. Provides interactive visualization of the skill dependency graph and a working job-description parsing example demonstrating invisible auto-execution, multi-step data evolution across DB rows, and provider-agnostic swappability.