Skip to content

README.md

Latest commit

 

History

History
35 lines (27 loc) · 2.72 KB

README.md

File metadata and controls

35 lines (27 loc) · 2.72 KB
Raw

Workflow System Documentation

Reference and concept docs for authoring workflows on the Nullplatform workflow engine — a plugin-first orchestration engine with n8n-style data flow, signals as the single pause primitive, and a descriptor-driven canvas.

For the hands-on loop (author → validate → test → publish) start with ../AUTHORING.md. Suite-specific docs (architecture, bring-up order, decisions) live inside each suite's own directory.

Concepts

  • How it works — the mental model in one page: what a workflow is, how it versions, how it runs, what you can observe.
  • Data model — items-based data flow and executeMode (per-item vs whole-batch).
  • Expressions — ${{ … }} syntax, the scope roots (workflow.inputs, variables, steps, $item…), operators, and truthiness.
  • Loops and streaming — the streaming-pagination pattern, engine fan-out, back-edge cycles, split-in-batches, forEach, sub-workflow iteration.
  • Signals and waits — the single pause primitive and the wait plugins built on it.
  • Triggers — how executions start, and the activate-on-alias rule.
  • Config entries — folder/workflow-scoped secrets (${{ secrets.X }}) and shared vars (${{ vars.X }}), with IaC bootstrap examples.

Reference

  • YAML & DSL — the workflow definition formats and field reference.
  • REST API — a copy-paste publish/trigger/inspect quickstart, then the full endpoint reference.
  • Plugin catalog — every shipped plugin with config schema and an example.
  • Built-in plugins — conceptual tour of the plugin families.

Guides

  • nullplatform integration — np-agent-command, np-action-item, and the AIOps patterns.
  • Slack — triggers, messages, button approvals, thread waits, and the conversational-agent loop.
  • Agents — the claude-code-agent step: resumable sessions, and plugins/workflows as tools.
  • Authoring gotchas — symptom → cause → fix for the sharp edges (expression scope, deciders & joins, error routing, deploy-time traps).
  • Scale limits & fan-out observability — the runtime ceilings on org-scale runs and their mitigations, plus fan-out progress and parallelism knobs.
  • AI suggestions that render well in the UI — rules for workflows that create action-item suggestions.