claudecodex

agent-orchestration

Cross-provider agent orchestration — spawn, steer, observe, cancel, and collect Claude Code, Codex, Grok Build, and Kimi agents through one durable MCP control plane, or launch any installed CLIs as a visible tmux team from a natural-language or saved workflow spec.

0 on GitHubby ByteDesk AIMITupdated 2 days agoEngineering
/

Plugin agent-orchestration from marketplace bytedesk in ByteDeskAI/bytedesk-marketplace. Commands read from the repo's manifest — 2 days ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use and it starts local MCP servers. The exact commands are listed below; read them before installing.

What it installs

Skills (11)

  • agent-orchestrate

    Delegate bounded work to Claude Code, Codex, Grok Build, or Kimi through the agent-orchestration MCP server. Use when this host (Claude, Codex, Grok, or Kimi) should spawn another catalog CLI, run competing model reviews, or coordinate parallel provider work.

  • agent-orchestration-doctor

    Diagnose the agent-orchestration plugin, provider CLIs, ACP adapters, state paths, and routing catalog without exposing secrets or sending provider prompts. Use when orchestration cannot spawn, a provider is missing, an installed plugin fails to start, or the user asks whether Claude, Codex, Grok, o

  • goal-feedback-loop

    Coordinate an admitted Task Management goal through PM, governed implementation, independent validation, approved test deployment and dogfood until its accepted criteria are proven. Use when the user says "/goal-feedback-loop", "run the goal loop", "drive EP-n to done", "take this goal end to end",

  • install-codex-orchestration-agent

    Install the bundled cross-provider Codex custom-agent template into personal or project scope with preview and overwrite protection. Use when the user asks to add, configure, or install the Codex orchestration agent after installing this plugin.

  • orchestrate

    Pick the right way to get work done by another agent, across agent-orchestration and task-management — dispatch a board task, drain the ready queue, ticket another repo, message one lead or all leads, wait for a reply, launch a team, ask another model, run a goal, check status or health, or mine for

  • orchestration-compose

    Turn a natural-language description of a multi-agent collaboration ("a Fable conductor, two designers on Codex and Grok, an Opus judge, human sign-off at the end") into a validated tmux orchestration spec, then save it as a reusable workflow. Use when the user describes agents working together, asks

  • orchestration-conduct

    The conductor's protocol inside a launched tmux orchestration — write briefs, send them through the mailbox, wait for replies, route to judges, run revision rounds, stop at human gates, and write the final report. Loaded automatically by the orchestrator agent from its BOOTSTRAP.md; also useful when

  • orchestration-launch

    Launch a saved orchestration workflow (or a spec file) as a tmux session with one pane per agent, each CLI started with its role, skills, and mailbox. Use when the user says "launch", "start the tournament", "spin up the team", "run the <name> orchestration", or names a workflow.

  • orchestration-status

    Inspect, troubleshoot, or stop a running tmux orchestration — which agents are alive, which replies are pending, what the journal shows, what an agent's screen says. Use when the user asks how a run is going, whether an agent is stuck, or wants to stop or clean up runs.

  • roadmap-governance

    Govern and evolve ROADMAP.md with the repository roadmap validator (formerly roadmap-orchestrator; $roadmap-orchestrator and /roadmap-orchestrator mean this skill). Use when asked to enhance the roadmap, extend the roadmap, select the next roadmap work, or act on task, unlock, trajectory, gap, or go

  • setup-agent-orchestration

    Prepare a machine for tmux-hosted multi-agent orchestrations — verify or install tmux for the OS, inventory installed agent CLIs against the provider adapters, create the user workflow and adapter folders, and register any extra CLI as an adapter, and wire Claude Code, Codex, Grok Build or Kimi as a

Subagents (2)

  • cross-provider-orchestrator

    Coordinates bounded Claude, Codex, Grok, and Kimi work through the agent-orchestration MCP server.

  • orchestration-observer

    Attaches to one existing orchestration, monitors it read-only, and reports evidence-backed findings to conductors.

MCP servers (1)

  • agent-orchestration

    runs: node ${CLAUDE_PLUGIN_ROOT}/dist/host-launcher.cjs

Hooks (6)

  • SessionStart

    node "${CLAUDE_PLUGIN_ROOT}/dist/cli.cjs" services ensure --consumer-cwd . --detach

  • SessionStart

    node "${CLAUDE_PLUGIN_ROOT}/topology/session-hook.mjs" SessionStart

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/autonomy-allow.mjs"

  • UserPromptSubmit

    node "${CLAUDE_PLUGIN_ROOT}/topology/session-hook.mjs" UserPromptSubmit

  • PostToolUse

    node "${CLAUDE_PLUGIN_ROOT}/topology/session-hook.mjs" PostToolUse

  • Stop

    node "${CLAUDE_PLUGIN_ROOT}/topology/session-hook.mjs" Stop

View source on GitHub

More from ByteDeskAI/bytedesk-marketplace

All 21 plugins →

structurizr

claudecodex

Enterprise C4 modeling with Structurizr DSL — indexed language reference, CLI/MCP tooling, 32 skills, validation gates, pattern catalog, and cookbook recipes.

32 skills · 12 commands · 1 agent · 1 MCP server

0

task-management

claudecodex

Takes over Claude Code's task system — native task calls, approved plans, decisions and commits mirrored into a git-tracked markdown store at .bytedesk/task-management/. One store per project shared across git worktrees, claim-based parallel work with node_modules sharing, an MCP server, and a live dashboard.

27 skills · 1 MCP server · hooks

0

design-skills

claudecodex

Design craft skills: motion and animation, Apple platform design, awwwards-grade frontend, prototyping, composition, and image generation. Split out of the old design-system plugin; carries no ByteDesk design content.

20 skills · 4 agents

0

design-patterns

claudecodex

Source-neutral design-pattern advisor with Markdown catalog, MCP tooling, smell scans, and architecture decision support.

6 skills · 11 commands · 1 agent · 1 MCP server · hooks

0

platform-dev

claudecodex

ByteDesk core implementation lifecycle — TDD engineer, feature start, PR-ready, worktree operator, architect, grounding, integration branches.

9 skills · 9 agents

0

remote-gateway

claudecodex

Operator skills for the ByteDesk remote gateway, including /remote-gateway-login (local yaml or --url/--method/--username/--password) and the gateway-logs and gateway-profiling diagnosis skills.

17 skills

0

Similar plugins

Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).

39 skills

mattpocock/skills 323,108

claude-mem

claudecodex

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

mem0

claudecodex

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks

mem0ai/mem0 68,893

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Backend API design, GraphQL architecture, workflow orchestration with Temporal, and test-driven backend development

9 skills · 1 command · 8 agents

wshobson/agents 42,461

Maintain agent-orchestration?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins21 plugins