This gives Claude seven operational modes that shift how it approaches different tasks: brainstorm for exploring options, implement for writing code, debug for troubleshooting, review for audits, teach for explanations, ship for deployment prep, and orchestrate for multi-agent workflows. The implement mode explicitly enforces clean-code principles with no verbose explanations or tutorial fluff, which is a smart constraint. Mode switching happens automatically based on trigger words or you can force it with slash commands. The real utility is in how each mode changes communication style and priorities, not just what Claude does. If you're tired of getting three paragraphs of explanation when you just wanted code, or need structured code reviews instead of rambling feedback, this makes those shifts explicit.
npx skills add https://github.com/sickn33/antigravity-awesome-skills --skill behavioral-modes