project-supervisor-orchestrator — Claude Code subagent for agent
A subagent that runs a chain of other agents step by step and returns the result as JSON. Useful when one request has to pass through several specialized agents.
npx claude-code-templates@latest --agent podcast-creator-team/project-supervisor-orchestratorWhat project-supervisor-orchestrator does
The subagent inspects an incoming request and decides whether it carries enough data to start. When the full set of fields is there, it invokes agents in order through the call_agent function and merges their output. When something is missing, it asks exactly one clarifying question before routing the work onward.
Inside are instructions for intent detection, passing data between sequential agents, and error handling. Every response, including a failure or a clarification request, comes back as JSON with a status of success, clarification_needed or error. Only Read and Write are listed as tools, so the agent itself writes no code and runs no commands.
The completeness check keys off specific episode fields — title, guest, topics, duration. Point it at a different input shape and you will need to edit that detection logic, or the agent keeps asking for clarification. Install with npx claude-code-templates@latest --agent podcast-creator-team/project-supervisor-orchestrator; the file lands at .claude/agents/project-supervisor-orchestrator.md.
What it helps with
- run a request through a chain of agents in order
- decide whether the data is complete or ask for the missing part
- merge output from several agents into one response
- return a failed step as parseable output
- keep data intact across agent handoffs
Example prompts for the agent
- run project-supervisor-orchestrator on this episode
- check whether the fields are complete, and ask if they are not
- push the data through the agent chain and return JSON
How to install project-supervisor-orchestrator
1claude-code-templates
npx claude-code-templates@latest --agent podcast-creator-team/project-supervisor-orchestrator2Manually: put the files in place
.claude/agents/project-supervisor-orchestrator.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/project-supervisor-orchestrator.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/project-supervisor-orchestrator.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about project-supervisor-orchestrator
How do I install project-supervisor-orchestrator in Claude?
Run npx claude-code-templates@latest --agent podcast-creator-team/project-supervisor-orchestrator, then check that the file appears at .claude/agents/project-supervisor-orchestrator.md.
What does project-supervisor-orchestrator do?
It coordinates several specialized agents: it checks whether the incoming request is complete, invokes agents in sequence, and combines their output into JSON.
Which tools does this subagent have?
The description lists only Read and Write. Other agents are invoked through the call_agent function rather than platform tools.
