When you're grinding through a long debugging session or need to cut token usage, this mode strips out articles, filler words, and pleasantries while keeping technical terms and code blocks intact. You trigger it with "caveman mode" or "be brief" and it stays on until you disable it. The skill smartly exits caveman for security warnings or irreversible actions, then resumes after. It's genuinely useful for rapid back and forth where you don't need the polish, though you'll want to turn it off before writing anything user facing. Also claims to reduce false positive content flags on touchy platforms since there's less surface area to trip filters.
npx -y skills add ganyuanran/aegis --skill communicating-concisely --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills