Workflow tip: use Astra for specs, Luna Max thinking for execution to save quota

dinowxyz · x · 2026-09-17

Sharing a two-chat workflow for tight AI coding quotas: have Astra design the spec and implementation plan (as a .md), explicitly framing the executor as a junior dev who tends to go off-rails so the plan stays precise; then in a separate chat, run Luna at Max thinking to walk through and execute the plan.

Related event: Multi-Agent Workflows Slash AI Coding Quota Costs(3 posts)→

Original post →

More from coding & agent

coding & agent channel →