schema Command
The schema command analyzes your workflow configuration to show what fields each action expects and produces - without making any API calls.
bash
BASH
agac schema -a <workflow-name> [options]
This catches field reference errors upfront. For example, if you typed extract_facts.fact instead of extract_facts.facts, you'd discover this at runtime after processing hundreds of records. The schema command validates these references statically.
:::tip Run from Anywhere You can run this command from any subdirectory within your project. :::
Options
| Option | Description |
|---|---|
-a, --agent TEXT | Workflow name (required) |
-u, --user-code DIRECTORY | Path to user code directory containing tools |
--json | Output as JSON for programmatic use |
-v, --verbose | Show detailed data flow visualization |
Examples
bash
BASH
# Show schemas in table format
agac schema -a my_workflow
# Output as JSON for programmatic use
agac schema -a my_workflow --json
# Show detailed data flow tree
agac schema -a my_workflow --verbose
# Include tool schemas from user code
agac schema -a my_workflow -u ./tools
Output Reference
The output shows:
- Input: Fields the action requires from upstream actions or source data
- Output: Fields the action produces for downstream actions
- (none): No input fields required
- (schemaless): Output schema not defined (tool without a YAML
schema:field) - (dynamic): Schema determined at runtime
Schema Sources by Action Type:
| Action Type | Input Schema Source | Output Schema Source |
|---|---|---|
| LLM | Template references and context_scope | schema field |
| Tool | context_scope in workflow YAML | schema field in workflow YAML |
See Also
- run Command - Execute agentic workflows
- Troubleshooting - Debug workflow issues