Work on one hard mathematics problem in rounds: each round a judge writes a few self-contained questions, independent workers answer them, and the judge keeps a ledger of what is proved, refuted and open, until the judge concludes or the rounds run out, and then a proof.md says plainly what is and is not proved. A run takes hours and dozens of worker runs.
Install
npx skills add anthropics/claude-plugins-official --skill siegenpx skills add anthropics/claude-plugins-official --skill siege -a claude-code -gnpx skills add anthropics/claude-plugins-official --skill siege -a codex -ggemini skills install https://github.com/anthropics/claude-plugins-official.git --path plugins/math-proof/skills/siegegh skill install anthropics/claude-plugins-official siegegit clone --depth 1 https://github.com/anthropics/claude-plugins-official /tmp/anthropics__claude-plugins-official
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/anthropics__claude-plugins-official/plugins/math-proof/skills/siege ~/.claude/skills/siege # Claude Code
cp -R /tmp/anthropics__claude-plugins-official/plugins/math-proof/skills/siege ~/.agents/skills/siege # Codex, Cursor, Gemini CLI, Copilot, OpenCode