I will read the paper, but I am curious if the methods promoted by eng/researchers at openai for models like codex 5.2/5.3 work? ie, is having a separate agent look at recent agent sessions and deduce problems the agents ran into and update agents.md (or more likely, the indexed docs referenced in an agents.md) actually helpful? A priori that seems like the main kind of meta prompting/harness you might expect to work more robustly.
imho, your best bet is to align your company on using the enterprise offerings of one of the major providers (OpenAI, for example). Once you make that purchase, you can integrate your Dremio lakehouse with their tooling using the MCP approach: https://platform.openai.com/docs/guides/tools-remote-mcp
