Skip to content
Back to search

mco

CLI-first orchestration for AI coding agents: run selected agents and models in parallel, compare raw answers, and coordinate review or implementation workflows.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/mco-org--mco/badge.svg)](/listings/mco-org--mco)

HTML

<a href="/listings/mco-org--mco"><img src="/api/listings/mco-org--mco/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Install the CLI and its bundled `mco-cli` Skill:

npx @tt-a1i/mco@latest install

Check the agents available on your machine:

mco doctor --json

Run a read-only multi-agent review:

mco review \
  --repo . \
  --prompt "Review this repository for high-risk bugs." \
  --providers claude,codex,pi

Run a coding task with workspace write access:

mco run \
  --repo . \
  --prompt "Implement the requested change and run the relevant tests." \
  --providers codex,pi \
  --execution-mode write

MCO never infers a provider/model team from detected binaries. If neither `--providers` nor `--agent` is supplied, a top-level `providers` config entry is used as the saved default; without one, the team must be selected explicitly. A calling agent should still show and confirm the resolved provider/model team with the user before dispatch.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 200 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <p align="center">
  2. <img src="https://raw.githubusercontent.com/mco-org/mco/main/docs/assets/brand/mco-cover-starry.jpg" alt="MCO — ten agent paths converging through a monumental M beneath a starry sky" width="100%" />
  3. </p>
  4. <h1 align="center">MCO</h1>
  5. <p align="center"><strong>Orchestrate AI coding agents. Compare perspectives. Act with confidence.</strong></p>
  6. <p align="center">
  7. <a href="https://www.npmjs.com/package/@tt-a1i/mco"><img src="https://img.shields.io/npm/v/@tt-a1i/mco?style=flat-square&color=cb3837&logo=npm&logoColor=white" alt="npm version" /></a>
  8. <a href="https://www.npmjs.com/package/@tt-a1i/mco"><img src="https://img.shields.io/npm/dm/@tt-a1i/mco?style=flat-square&color=cb3837" alt="npm downloads" /></a>
  9. <a href="https://github.com/mco-org/mco/stargazers"><img src="https://img.shields.io/github/stars/mco-org/mco?style=flat-square&color=f59e0b" alt="GitHub stars" /></a>
  10. <a href="./LICENSE"><img src="https://img.shields.io/badge/License-MIT-22c55e?style=flat-square" alt="MIT License" /></a>
  11. <img src="https://img.shields.io/badge/Python-3.10%2B-3776AB?style=flat-square&logo=python&logoColor=white" alt="Python 3.10+" />
  12. </p>
  13. <p align="center">English · <a href="./README.zh-CN.md">简体中文</a></p>
  14. MCO is a lightweight, CLI-first orchestration layer for AI coding agents. Give one task to the agents and models you choose, run them in parallel, and compare their raw answers before you act.
  15. Use MCO for code review, implementation, architecture analysis, CI checks, and any workflow where one model's blind spots matter.
  16. It works from a terminal or from another coding agent such as Claude Code, Codex, Cursor, Copilot, Pi, or OpenClaw.
  17. > MCO is actively maintained. For a browser workbench with persistent agent identity and a shared task graph, see [Hive](https://hivehq.dev).
  18. ## Quick start
  19. Install the CLI and its bundled `mco-cli` Skill:
  20. ```bash
  21. npx @tt-a1i/mco@latest install
  22. ```
  23. Check the agents available on your machine:
  24. ```bash
  25. mco doctor --json
  26. ```
  27. Run a read-only multi-agent review:
  28. ```bash
  29. mco review \
  30. --repo . \
  31. --prompt "Review this repository for high-risk bugs." \
  32. --providers claude,codex,pi
  33. ```
  34. Run a coding task with workspace write access:
  35. ```bash
  36. mco run \
  37. --repo . \
  38. --prompt "Implement the requested change and run the relevant tests." \
  39. --providers codex,pi \
  40. --execution-mode write
  41. ```
  42. MCO never infers a provider/model team from detected binaries. If neither `--providers` nor `--agent` is supplied, a top-level `providers` config entry is used as the saved default; without one, the team must be selected explicitly. A calling agent should still show and confirm the resolved provider/model team with the user before dispatch.
  43. ## Why MCO
  44. One agent gives you one perspective. MCO turns selected agents into a review or execution team:
  45. 1. **Choose** — explicitly select the agents for the task.
  46. 2. **Dispatch** — run them in parallel, chain their work, or divide the scope.
  47. 3. **Compare** — retain each invocation's complete raw answer and operational status.
  48. 4. **Decide** — inspect evidence, disagreements, and failures before acting.
  49. MCO keeps answer text opaque. It does not turn natural-language output into findings, severity, confidence, consensus, or an automatic decision.
  50. For explicit review coordination, `--perspectives-json` adds a Provider-specific prompt focus. `--divide files` excludes ignored/local/build directories and round-robins the remaining sorted repository files without overlap, while `--divide dimensions` rotates review lenses in declaration order without changing target paths. These choices are visible in dry-run and arrange only the prompt or scope; the returned invocation answers remain raw.
  51. ## Built-in providers
  52. | Provider | CLI | Provider ID |
  53. |----------|-----|-------------|
  54. | Claude Code | `claude` | `claude` |
  55. | Codex CLI | `codex` | `codex` |
  56. | Gemini CLI | `gemini` | `gemini` |
  57. | OpenCode | `opencode` | `opencode` |
  58. | Qwen Code | `qwen` | `qwen` |
  59. | GitHub Copilot CLI | `copilot` | `copilot` |
  60. | Hermes | `hermes` | `hermes` |
  61. | Pi | `pi` | `pi` |
  62. | [Grok Build](https://docs.x.ai/build/overview) | `grok` | `grok` |
  63. | [Cursor CLI](https://cursor.com/docs/cli/overview) | `cursor` / `agent` | `cursor` |
  64. | Antigravity CLI | `agy` | `agy` |
  65. | Kimi Code CLI | `kimi` | `kimi` |
  66. Each provider CLI remains responsible for its own installation, authentication, model access, and native sandbox behavior.
  67. ## Common workflows
  68. | Goal | Command |
  69. |------|---------|
  70. | General multi-agent task | `mco run --providers claude,codex --prompt "..."` |
  71. | Thin raw-answer code review | `mco review --providers claude,codex --prompt "..."` |
  72. | Compare multiple models | `mco run --agent fast=pi:model-a --agent careful=pi:model-b --prompt "..."` |
  73. | Preview without execution | `mco review --providers claude,pi --dry-run --json` |
  74. | Live terminal progress | `mco review --providers claude,codex --stream live` |
  75. | Machine-readable events | `mco review --providers claude,codex --stream jsonl` |
  76. | File-backed chain | `mco run --agent first=pi:model-a --agent next=pi:model-b --chain --result-mode artifact` |
  77. | Debate and synthesis | `mco review --providers claude,codex --debate --synthesize --result-mode both` |
  78. | Discover provider models | `mco agent models --providers codex,pi --json` |
  79. Pin models for one run without changing provider CLI defaults:
  80. ```bash
  81. mco review \
  82. --providers codex,pi \
  83. --provider-models-json '{"codex":"gpt-5.4","pi":{"provider":"seal","model":"deepseek-v4-pro"}}' \
  84. --prompt "Review this repository for bugs."
  85. ```
  86. ## Permissions and safety
  87. MCO translates one execution profile into each provider's native flags:
  88. | Mode | Intended use | Default |
  89. |------|--------------|---------|
  90. | `read_only` | Inspect and review without workspace mutation | `mco review` |
  91. | `write` | Create and edit workspace files | `mco run` |
  92. | `yolo` | Use the provider's broadest bypass profile | Explicit opt-in only |
  93. Important boundaries:
  94. - `--allow-paths` validates MCO's requested scope; it is not an operating-system sandbox.
  95. - Provider sandbox strength depends on the underlying CLI.
  96. - Hermes oneshot bypasses approvals and therefore requires explicit `--execution-mode yolo`.
  97. - ACP terminal access is a trusted-agent capability. Use isolation for untrusted agents or prompts.
  98. - MCO does not create or manage worktrees. If the user selects parallel writers, partition ownership with non-overlapping `--target-paths` and warn about edit conflicts.
  99. See [Provider and permission reference](./docs/reference/providers.md) for the complete mapping.
  100. ## Use MCO from another agent
  101. MCO's CLI is self-describing. A calling agent can read `mco -h`, resolve saved defaults, confirm the provider/model team with the user, preview the policy, and then execute.
  102. > “Use MCO to run a security review with Claude and Codex, and an architecture review with Pi.”
  103. The installer and runtime use two different selections:
  104. - Installer `--agent` chooses which calling agents receive the MCO Skill.
  105. - Runtime `--providers`, saved `providers` config, or runtime `--agent` declarations choose the task invocations.
  106. ```bash
  107. npx @tt-a1i/mco@latest install --agent codex --agent claude-code --yes
  108. mco doctor --skill-health --json
  109. ```
  110. ## How it works
  111. ```text
  112. You or a calling agent
  113. │
  114. ▼
  115. mco run / review
  116. │
  117. ├── Claude ──┐
  118. ├── Codex │
  119. ├── Gemini ├──► raw answers / file-backed stages ──► output
  120. ├── Pi │
  121. └── ... ───┘
  122. │
  123. text · JSON · JSONL · Markdown artifacts
  124. ```
  125. Provider processes are isolated behind a shared adapter contract: detect, run, poll, cancel, and transport decode. One invocation failure does not discard successful provider answers.
  126. ## Documentation
  127. | Topic | Guide |
  128. |-------|-------|
  129. | Installation, first run, and common workflows | [Workflow guide](./docs/guides/workflows.md) |
  130. | Providers, models, and permission mappings | [Provider reference](./docs/reference/providers.md) |
  131. | CLI flags, outputs, artifacts, and exit codes | [CLI reference](./docs/reference/cli.md) |
  132. | Config files and custom agents | [Configuration reference](./docs/reference/configuration.md) |
  133. | Machine-readable error contract | [Error contract](./docs/contracts/errors-v0.1.x.md) |
  134. | Invocation and artifact contract | [Invocation contract](./docs/contracts/invocation-runtime-v1.md) |
  135. | Provider permission contract | [Permission contract](./docs/contracts/provider-permissions-v0.1.x.md) |
  136. | Release process | [RELEASING.md](./RELEASING.md) |
  137. | Release history | [CHANGELOG.md](./CHANGELOG.md) |
  138. Run `mco <command> --help` for the authoritative option list installed with your version.
  139. ## Development
  140. ```bash
  141. git clone https://github.com/mco-org/mco.git
  142. cd mco
  143. python3 -m pip install -e .
  144. python3 -m unittest discover -s tests -p 'test_*.py'
  145. npm test
  146. ```
  147. ## License
  148. MIT — see [LICENSE](./LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.