A general best-of-N selection engine wearing a World Cup-style tournament. Any time a task is "mass-produce many candidates and keep the single best" (or the user hands over a field of them), stage 32 teams (8 groups of 4) or 48 teams (12 groups, 2026 format) through a head-to-head round-robin group stage into a single-elimination knockout, judged by taste-calibrated LLM panels, and emit a World Cup-flavored HTML report of the final bracket graph. The "teams" can be anything comparable: essay or copy variants, taglines, names, API or UI designs, prompts, configs, plans, generated code solutions. Essays/prose are one supported profile, not the required shape. This skill should be used when the user wants to bracket a field and crown a winner with real rigor, or whenever a task reduces to "make N variants and pick the best." Triggers: "/worldcup", "run a worldcup", "tournament these", "bracket these 32", "best of N", "mass-produce and pick the best", "stage 48 variants and crown a winner", "group stage then knockout". Requires the ultracode Workflow tool. The judge is the point: domain profiles plug in a fatal-flaw/fabrication gate, a fact-ledger truth anchor (the original competes as one of the N when you want it judged), calibrated pairwise seeding, diverse-lens panels, Elo rating confirmation, and optional cross-model (agentify) jurors for the finals.
Install with
Claude CodeCodexCursorGitHub CopilotGemini CLI
Supported by the Atlas install prompt. Skill-specific requirements vary.
Skill stats
Skills1CategoryGeneralStars0
About this skill
A general best-of-N selection engine wearing a World Cup-style tournament. Any time a task is "mass-produce many candidates and keep the single best" (or the user hands over a field of them), stage 32 teams (8 groups of 4) or 48 teams (12 groups, 2026 format) through a head-to-head round-robin group stage into a single-elimination knockout, judged by taste-calibrated LLM panels, and emit a World Cup-flavored HTML report of the final bracket graph. The "teams" can be anything comparable: essay or copy variants, taglines, names, API or UI designs, prompts, configs, plans, generated code solutions. Essays/prose are one supported profile, not the required shape. This skill should be used when the user wants to bracket a field and crown a winner with real rigor, or whenever a task reduces to "make N variants and pick the best." Triggers: "/worldcup", "run a worldcup", "tournament these", "bracket these 32", "best of N", "mass-produce and pick the best", "stage 48 variants and crown a winner", "group stage then knockout". Requires the ultracode Workflow tool. The judge is the point: domain profiles plug in a fatal-flaw/fabrication gate, a fact-ledger truth anchor (the original competes as one of the N when you want it judged), calibrated pairwise seeding, diverse-lens panels, Elo rating confirmation, and optional cross-model (agentify) jurors for the finals.
Quick start
Copy the install prompt. Use the green button above this README.
Send it to your agent. Paste it into your coding agent’s chat and send.
Review the setup. Your agent checks the skill’s instructions, requirements, and installation scope.
A general best-of-N selection engine wearing a World Cup-style tournament. Any time a task is "mass-produce many candidates and keep the single best" (or the user hands over a field of them), stage 32 teams (8 groups of 4) or 48 teams (12 groups, 2026 format) through a head-to-head round-robin group stage into a single-elimination knockout, judged by taste-calibrated LLM panels, and emit a World Cup-flavored HTML report of the final bracket graph. The "teams" can be anything comparable: essay or copy variants, taglines, names, API or UI designs, prompts, configs, plans, generated code solutions. Essays/prose are one supported profile, not the required shape. This skill should be used when the user wants to bracket a field and crown a winner with real rigor, or whenever a task reduces to "make N variants and pick the best." Triggers: "/worldcup", "run a worldcup", "tournament these", "bracket these 32", "best of N", "mass-produce and pick the best", "stage 48 variants and crown a winner", "group stage then knockout". Requires the ultracode Workflow tool. The judge is the point: domain profiles plug in a fatal-flaw/fabrication gate, a fact-ledger truth anchor (the original competes as one of the N when you want it judged), calibrated pairwise seeding, diverse-lens panels, Elo rating confirmation, and optional cross-model (agentify) jurors for the finals.
About this listing
Category
General
Reported license
MIT
Listing language (estimated)
English
Review status
Not individually reviewed
Agent compatibility
Atlas installation supports Claude Code, Codex, and Cursor. Skill-specific dependencies and features are checked during setup.
Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLI