Purpose

This document defines the rules and conventions that AI agents must follow when creating databases, tickets, or any structured output within the Next Effect Notion workspace. It serves as the binding contract between human operators and automated agents.

Core Principles

  1. Tickets are the universal intake layer. Every agent action that produces work starts with a ticket in the 🎫 Tickets database.
  2. Agents write to OS Layer databases through tickets. Agents can create and update records in Tasks, Engagements, Initiatives, Internal Projects, and OKRs — but every write must trace back to a source ticket. Agents must never create orphaned work with no ticket lineage.
  3. Agent-generated databases are ephemeral workspaces — they exist to hold structured output for a specific ticket and always trace back to a source ticket.

Ticket Creation Rules

Ticket System — Agent Interface Spec

When an agent creates a ticket, it must provide these required fields:

Field Type Example
Title text "Migrate client data from legacy CRM"
Project Code select NE-OPS
Description rich text Detailed description of the request
Priority select P1

These fields auto-populate or are optional:

Field Behavior
Ticket ID Auto-increments
Serialized ID Formula: Project Code + Ticket ID (e.g., NE-OPS-001)
Source Must be set to "Agent"
Requester Agent name or ID (e.g., "goose-mcp-v1")
Status Defaults to "New"
Created Date Auto-set
Assignee Optional — leave blank for human triage
Due Date Optional
Related Engagement Optional — set if the agent knows the engagement
Related Initiative Optional — set if the agent knows the initiative

Agent-Generated Database Convention

When an agent needs to create a project-specific database to track its own structured work output (migration logs, audit trails, research results, etc.), it must follow these rules:

Naming Convention

{Serialized Ticket ID}-{Purpose}

Examples: