Workflow Compose

Workflow Compose chains multiple skills sequentially using YAML declarations, automating complex pipelines like API definition, validation, and model generation in a single repeatable workflow. Define steps in YAML, specify dependencies, and execute with deterministic error handling and audit logging built in.

Yes. Mark steps as `required` to stop on failure or omit the flag to continue on errors. This enables robust execution where non-critical steps don't block the pipeline, while critical ones halt processing and log the failure to audit records.

Execution history automatically persists to `/registry/workflow_history.json` with full provenance tracking, and integrates with `audit.log` for traceability. Every step execution is recorded with outcomes and timing for compliance and debugging.

Write a YAML file listing skills as sequential steps with their inputs and error handling preferences. Pass the file path to `workflow_compose.py` to execute the entire chain in order, with timeouts and schema validation enforced at each step.

No. Workflow Compose uses declarative YAML syntax to define orchestration logic without custom scripting. The framework handles in-process skill execution, validation, timeout management, and audit integration automatically.

Schema validation ensures YAML structure matches workflow requirements before execution. The engine validates skill references, parameter types, and dependency resolution, preventing misconfigured workflows from running and catching errors early.