Compare coding-agent answers and coordinate reviews with MCO
Dispatch one repository task to explicit Claude, Codex, Gemini, Cursor, OpenCode, Qwen, Copilot, or custom provider teams and compare the raw answers before acting.
npx skills add agentskillexchange/skills --skill compare-coding-agent-answers-and-coordinate-reviews-with-mco
Use MCO when a user or calling agent needs several coding agents or models to review, analyze, or implement against the same repository task and return separate raw answers for comparison. Invoke it instead of asking one product normally when the risk is single-model blind spots, review disagreement, or parallel specialist perspectives that should remain visible before a human or lead agent acts. The scope boundary is explicit CLI-first dispatch, provider selection, permissions, streaming, and raw-answer comparison for coding-agent workflows; it is not a general multi-agent framework, worktree manager, autonomous project manager, or SDK listing.
What this skill actually does
Inputs and prerequisites: Node.js 18+, Python 3.10+, MCO CLI, and the provider CLIs selected for the run such as Codex, Claude Code, Gemini, Cursor, OpenCode, or Qwen.
Setup notes: Install with npx @tt-a1i/mco@latest install, run mco doctor –json to inspect available providers, then run commands such as mco review –repo . –prompt ‘Review this repository for high-risk bugs.’ –providers claude,codex or mco run –repo . –prompt ‘Implement the requested change and run tests.’ –providers codex,pi –execution-mode write.
Source and verification boundary: use https://github.com/mco-org/mco as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Codex workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.