The OS Layer is your operational backbone — a connected graph of Notion databases that tracks everything from strategic initiatives to individual tasks. Every database can reach every other database within 1-2 hops, and both humans and AI agents operate within the same system.
How work flows: Initiatives define what to pursue. Engagements define client work. Internal Projects organize delivery. Tickets are the universal intake layer — every request (from a human or agent) enters as a serialized ticket. Tasks are the execution layer where work actually happens.
The ticket system uses serialized IDs in the format <PROJECT_CODE>-<NNN> (e.g., {PREFIX}-OPS-001, CLIENT-014). Project codes map to clients, engagements, and internal workstreams. Tickets link to Tasks, Engagements, Initiatives, Internal Projects, and Clients — so any ticket can trace back to its strategic context.
For agents: AI agents have full read/write access to all databases. They follow a convention-based operating model: every action starts with a ticket, structured output goes in child databases named after the ticket, and all progress is logged via comments. The system prompt and machine-readable config below give agents everything they need to operate.
These explain what the system is, how the codes work, and how everything connects:
Agent Permissions Configuration
Agent Database Creation Convention
Agent Config — Machine Readable
Agent System Prompt — Copy/Paste Block
Agent Patterns — What to Ask Your Agent
💡 Quick start for a new agent: Copy the system prompt from the Copy/Paste Block into your agent's config. It tells the agent to fetch the Machine Readable config page on first action, which contains all database IDs, enum values, and conventions. No hardcoding needed.
‣ → ‣ → ‣