Tana Systems Lab: Self-Driving Tana
Learn how to build AGENTS.md and SKILLS.md files that teach your AI agent to work with your specific Tana setup reliably and repeatably.
In the last Systems Lab, we showed how MCP can be used to help design your Tana schema and ontology. In this next session, we’ll tackle the main challenge that comes next: teaching your agent to work with your specific Tana setup reliably and repeatably.
We’ll cover building an AGENTS.md that reflects your unique tag and field conventions, personal lingo, and operating rules, creating SKILLS.md files that let your agent query, capture, and update your system correctly, and iterating on this process so the behavior gets tighter over time.
