LaneOrchestrator
Your task. Astra's plan. The right Codex team.
2 followers
Your task. Astra's plan. The right Codex team.
2 followers
Stop choosing models. Start building. LaneOrchestrator turns your Codex task into an Astra-led workflow: the right specialists, models and thinking levels, chosen automatically from your codebase and the work ahead. With 172 bundled specialists, Astra can tackle complex changes or delegate to Sol, Terra and Luna. See who's doing what and why, with independent review for consequential changes. An open-source skill for Codex desktop and CLI. You set the goal. Astra assembles the team.
GPT-6 Astra is now at the heart of LaneOrchestrator. Describe the work. Let it assemble the team.
I built this Codex skill because choosing the model, thinking level and specialist for every task kept interrupting the actual work. Now Astra reads your prompt and repository context and makes those decisions automatically. You don't need to tell it to use Astra.
For demanding changes, Astra can handle implementation with a fresh, independent reviewer. For other work, it can orchestrate Sol implementation or coordinate Terra specialists. Small, clearly scoped changes can go to Luna or Terra. The model and thinking level follow the task's complexity and risk.
All 172 bundled specialists can use supported models, including Astra. Expertise and model choice are separate, so a backend specialist isn't permanently tied to one model. Routing decisions stay visible, and you can override them.
LaneOrchestrator is open source and works with Codex desktop and CLI. Try it on a real task and tell me where the routing gets it right or misses.
Get started and watch the demo: https://github.com/KarthikRamesh9149/laneorchestrator