GitHub API Cache Overview Cache GitHub project metadata ONCE at session start. All subsequent operations use cached data. Core principle: Fetch once, extract many. Never repeat API calls for the same data. This skill is called by and provides cached data to all other skills. The Problem GitHub GraphQL API has a 5,000 point/hour limit. Without caching: - = 1 call per invocation - = 1 call per invocation - = 1 call per invocation A typical session startup was consuming 3,500+ of 5,000 points through repeated calls. The Solution Fetch project metadata ONCE and cache in environment variables. All…