Codex orchestrator/subagent skill: Astra or Sol leads, Luna subagents
Melvin Vivas · X post · 2026-09-15 · Open on X
Topics: AI Agents, Tool Use & MCP, AI Dev Tools & Productivity · Level: intermediate
Summary
The creator shares a Codex skill for multi-agent setups. Astra or Sol acts as the orchestrator and Luna models run as subagents. He says that with this skill, subagents always report back to the orchestrator.
Key points
- Orchestrator pattern: a stronger model (Astra or Sol) coordinates and cheaper Luna models do the subtasks.
- The skill is meant to fix subagents failing to report results back to the orchestrator.
- It's packaged as a Codex skill in a public GitHub repo.
Resources mentioned
- donvito/codex-astra-luna-orchestrator · repo · github.com · free
The creator's Codex skill for an orchestrator-plus-subagents setup (Astra orchestrator, Luna subagents), configured through config.toml.
Also in: Codex Orchestrator Pattern: Astra/Sol Orchestrator with Luna Subagents (Melvin Vivas on X · notes), Open-source Codex skill: Astra/Sol orchestrator with Luna subagents (Melvin Vivas on X · notes), Cost-Efficient Codex Subagents: Astra/Sol Orchestrator + Luna Workers (Melvin Vivas on X · notes), Codex Orchestrator v0.2.1: GPT-6 Sol orchestrator with Luna subagents (Melvin Vivas on X · notes) and 27 more
Try this
- Try the codex-astra-luna-orchestrator skill if your Codex subagents don't report back to the orchestrator.
More in AI Agents, Tool Use & MCP
- Cut agent costs by giving chore subagents a cheaper model
- WebMCP: Exposing Existing App Features as Agent Tools
- OpenResearch: turn a coding agent into a research agent
- Codex traces: picker model is the orchestrator, subagents stay Luna
- Simpler Codex orchestration: drop the project .codex/config.toml
- Portable Computer Now Runs AI Agents and Models Locally on NVIDIA RTX PCs