🌐 Cross-Agent by DefaultOne install puts the same skills in Claude Code, Gemini CLI, OpenCode, Kimi CLI, and Codex.
📊 Every Skill Is Scoredskill-validator grades clarity, actionability, and token efficiency before a skill merges.
📦 Context-FrugalA lean SKILL.md loads upfront. Detailed references load only when a task needs them, saving tokens on docs the agent isn't using.
✅ Gated Before MergeFormat, lint, security scans, and skill scoring run on every PR. Nothing lands without passing all of them.