Pipeline YAML schema

Canonical shape is pipeline.EditorDefinition (DB published YAML). Load this file before writing or editing a definition.
Step details: steps.md. Capability params: capabilities.md.

name: example_pipeline
description: "Human-readable description"
enabled: true
resumable: false
triggers:
  - type: event
    enabled: true
    event: demo.item.created
  - type: cron
    enabled: false
    cron: "0 3 * * *"
    cron_timeout: "30m"
steps:
  - name: notify
    capability: core
    operation: notify_send
    params:
      template_id: "demo.new_item"
      channels: ["slack"]
      payload:
        title: '{{event "title"}}'

Top-level fields

Field	Required	Notes
`name`	yes	Unique; validated by `pipeline.ValidateName`
`description`	no	Human-readable
`enabled`	no	Published disabled pipelines appear in `list` but cannot `run` / load into the engine
`resumable`	no	Checkpoint + restart recovery
`triggers`	yes	Array of event / cron / webhook triggers
`steps`	yes	Ordered capability invocations

Triggers

Type	Fields
`event`	`event` (DataEvent.EventType)
`cron`	`cron`, optional `cron_timeout`
`webhook`	`webhook.path`, auth token and/or hmac_secret

Authoring checklist

Copy the skeleton; set name and enabled: true.
Add at least one enabled trigger.
For each step, open capabilities/<type>.md and fill required params.
Use event / step helpers only from steps.md.
flowbot pipeline apply --file ... then get / run.
apply rejects kind: pipeline_blueprint. Instantiate via flowbot blueprint instantiate (catalog: pkg/pipeline/blueprints/).