Darwinomics Engine

The deterministic execution layer. Takes approved execution plans from the agent and runs them step by step. Does NOT decide what to run — the agent decides. The engine only executes. Same research data artifact input + same tool + same parameters = same research data artifact output, guaranteed. Content-addressed caching means identical runs are zero-cost. If a step fails, the engine reports the error to the agent for recovery.

Use in DarwinOmics

How to use it

Use this workflow element when the submitted analysis needs a multi-step execution path, imported workflow, Darwinomics wrapper, or workflow-scale comparison rather than a single module.