Docyrus Automation Design

Build an automation with docyrus automation, then validate its trigger/node graph and test that it actually fires. An automation = one trigger (the event) + an ordered graph of action nodes (what runs). This skill is the design workflow; the platform's conceptual catalog of every trigger and node type lives in the docyrus-platform skill (references/automation-and-workflows.md), and the exhaustive CLI flags are available via docyrus automation … --help (command index: docyrus-cli-app). This skill ties them together and records the gotchas that only surface when you actually run the commands.

Workflow

Follow in order. An automation that hasn't been validated and test-run is not done.

1. Confirm app + auth. Every automation belongs to an app.

   docyrus auth who --json          # confirm session + tenant
   docyrus apps list --json         # find the target appSlug / appId
   

   No session → stop and ask the user to run docyrus auth login.

2. Design before issuing commands. Decide: the trigger (which event; which data source it watches), then the action nodes in order, each node's type, its condition (does it run?), and its field mappings (what data it writes/sends). Sketch the graph for the user and confirm. See references/trigger-and-node-catalog.md to choose types.

3. Create the automation + its first trigger (one atomic call — see Create). You must pass a --triggerType; the automation cannot exist without one.