Multi-Model-Skill-Sync

create and maintain a canonical project ai control layer for software repos, with one source-of-truth project spec plus per-model skill and agent adapters for chatgpt, claude, gemini, and similar systems. use when chatgpt needs to reduce ai development drift, generate or update repo-resident ai instructions under .ai/skills and .agents, design a manual-trigger github action for skill sync, add approval gates for destructive or high-risk work, or account for context drift across models and over time.