Instead of asking Claude "what do you think," this skill reframes your question as "who in the world knows this best, and what would they say?" It's based on Andrej Karpathy's idea of treating LLMs as simulators rather than entities. Trigger it with phrases like "顶级专家" or "best minds" and it identifies the actual expert who should answer your question, then simulates their perspective based on their real work and public statements. The difference from similar skills is simple: it doesn't pick from a preset list of thirteen coaches, it finds whoever genuinely knows your specific problem best. Works as a thinking methodology layer on top of your regular prompts.
npx -y skills add ceeon/best-minds --skill best-minds --agent claude-codeInstalls to .claude/skills
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills