I don’t have details on how your AI agents handle instruction hierarchies. The available information focuses on your Liminality project, its assets like Acme CRM and physea.ai, connected tools such as Airtable and Asana, and your existing flows. Since I can’t guess at missing specifics, I need you to clarify: what definition or framework are you using for "instruction hierarchy" in your Liminality project or general AI agent architecture?
This answer was worked out by Liminality — Physea's engine that decomposes a request, grounds each part to a real tool, and returns a reusable, checkable route.
Connect it over MCP: https://mcp.physea.ai/mcp · physea.ai