Jarrodwatts Claude Delegator
jarrodwatts-claude-delegator is a Claude Code plugin for the Build phase that delegates scoped tasks to Codex and Gemini through MCP-style orchestration and specialized expert subagents.
by jarrodwatts · github.com/jarrodwatts/claude-delegator
Install Claude Delegator when you want to hand scoped work from Claude Code to Codex or Gemini via MCP-style orchestration and expert subagents.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install jarrodwatts-claude-delegator@jarrodwatts/claude-delegatorBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:jarrodwatts/claude-delegator") and the paste-ready config with skillselion.get_install_config("plugin:jarrodwatts/claude-delegator").
What it does
jarrodwatts-claude-delegator is a Claude Code plugin that lets solo builders delegate bounded tasks to Codex and Gemini while Claude remains the orchestrator. Catalog copy and keywords emphasize MCP, OpenAI GPT/Codex paths, subagents, and expert modes such as analyst, architect, reviewer, and security—useful when one model is great at planning but another is stronger at a narrow audit or implementation slice. You install it when parallel model access is part of your stack but context switching between terminals and browsers is killing momentum. Typical flows: scope a change in Claude, delegate a test-heavy or review-heavy slice, merge results back into your main thread. It is an orchestration layer, not a replacement for your own acceptance criteria or production secrets handling. Expect intermediate setup (API access, delegate configuration) and clear task boundaries so delegation stays safe on a one-person codebase.
Highlights
- Delegates tasks from Claude Code to Codex and Gemini without leaving the primary session
- MCP-oriented orchestration with scoped delegation rather than blind full-repo handoffs
- Specialized expert personas including analyst, architect, reviewer, and security-oriented roles
- Single-plugin bundle (pluginCount 1) from jarrodwatts/claude-delegator (~933 catalog stars)
Why builders use it
Solo builders waste time context-switching between Claude Code, Codex, and Gemini when each model fits a different slice of the same feature.
After installation, you route analyst, architect, reviewer, or security-style work to the right model from inside Claude Code and fold answers back into one workflow.
At a glance
- Type - Plugin in AI Agents.
- Adoption - 0 installs, 933 stars, 0 votes.
FAQ
Who is jarrodwatts-claude-delegator for?
It is for Claude Code users who want Codex and Gemini in the loop via delegation, especially when expert-style roles improve quality on a small team of one.
When should I use jarrodwatts-claude-delegator?
Use it during implementation and review when a subtask fits another model better—architecture critique, security pass, or Codex-heavy coding—while Claude coordinates the session.
How do I add jarrodwatts-claude-delegator to my agent?
Add the jarrodwatts/claude-delegator plugin in Claude Code, configure MCP and provider credentials per the repo README, then invoke delegator commands with explicit scope for each handoff.
Comments
Share how you use jarrodwatts-claude-delegator, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.