Building Modules
How platform teams build versioned modules (golden paths) with Module Studio: schemas, actions, bindings, simulation, publishing.
This page describes the deck-aligned model of authoring modules in Argy.
Modules are built in Module Studio as workflow graphs and published into the catalog as versioned golden paths.
What a module is
A module is a versioned workflow composed of actions. Technically it is a graph:
- Inputs schema (JSON Schema)
- Outputs schema (JSON Schema)
- Workflow: nodes + edges (DAG)
- Bindings between nodes, e.g.
{{nodeLabel.outputs.key}}
Modules are versioned with SemVer and can be labeled with status (stable / experimental / deprecated / draft).
Build modules in Module Studio
1) Design the schema
Define what teams can configure. Keep inputs minimal, validate constraints, and use safe defaults.
2) Compose actions (50+)
Use the palette to build workflows: Cloud & IaC, CI/CD, Security, Observability, Notifications, Flow control (parallel/conditional/wait), and more.
3) Bind inputs/outputs
Connect nodes and use bindings so data flows through the module.
4) Add governance
Add approval gates and governance checks where required (publish/deploy actions, environment-level rules, auditability).
5) Simulate before publishing
Run a sandbox simulation: follow logs in real time (WebSocket), inspect outputs, validate the graph (cycles, schema checks).
6) Publish and version
Publish with SemVer, changelog, visibility (private / org / public), and status badges (stable / experimental / deprecated / draft).
The Argy AI action (module-defined subagent)
Among the available actions, Argy AI embeds an AI-driven step defined at module level (custom prompts + tools). It can orchestrate sub-agents, which is how enterprises build their own governed AI agents inside modules.
Next steps
- Core concepts: Modules & Golden Paths
- Governance controls: Policies & Guardrails
- Versioning: Versioning & Lifecycle