Divergent ideation across model families — the same brainstorm, but each cognitive frame is handed to a different model (Claude sub-agent, OpenAI Codex, a cheap third via OpenCode) so the ideas diverge by frame and by model priors at once.
安装
npx skills add szarkans/multi --skill adhdnpx skills add szarkans/multi --skill adhd -a claude-code -gnpx skills add szarkans/multi --skill adhd -a codex -ggemini skills install https://github.com/szarkans/multi.git --path skills/adhdgh skill install szarkans/multi adhdgit clone --depth 1 https://github.com/szarkans/multi /tmp/szarkans__multi
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/szarkans__multi/skills/adhd ~/.claude/skills/adhd # Claude Code
cp -R /tmp/szarkans__multi/skills/adhd ~/.agents/skills/adhd # Codex, Cursor, Gemini CLI, Copilot, OpenCode