swamp-getting-started

Start by stating your automation goal in plain words. The walkthrough searches local model types and the extension registry for a match, creates a model with swamp model create, validates it, runs a read-only method, and shows the output step by step.

First find a model type with swamp model type search or swamp extension search, pulling an extension if needed. Then run swamp model create <type> <name>, edit the generated YAML to match your goal, and confirm it passes swamp model validate.

The walkthrough runs swamp model search first; if models already exist, it tells you that you are past onboarding and stops. You can then describe what you want to work on and be routed to the appropriate swamp guide.

Validation usually fails due to missing required arguments, invalid argument values, or a wrong file path. Check the type schema with swamp model type describe, fix the model YAML, then re-run swamp model validate before continuing.

Use command/shell only for genuine one-off ad-hoc commands like checking disk space. For wrapping CLI tools or building integrations, create a typed model or custom extension with proper schemas instead.

Create a vault with swamp vault create local_encryption, store secrets with swamp vault put, and reference them in model YAML using the vault.get expression. This keeps API keys out of plain configuration files.