run-pipeline

A closed-loop marketing initiative orchestrates sequential stages from research to evaluation. Orchestration manages it by threading leaf skills across named stages, pausing at gates for manual approval rather than auto-executing, and storing resumable state in a dedicated loop workspace.

You can scaffold a measurable loop workspace without executing the full pipeline by using the eval-only mode. This mode sets up the six-stage loop structure and transition logs under the designated loop folder, enabling cross-stage evaluation without dispatching stage work.

You can pause a multi-stage content workflow and resume it later without losing state. The orchestrator pauses at every gate and logs transitions, storing all state and artifacts in a durable loop folder to ensure resumable progress across stages.

To start orchestrating a six-stage content loop, run the scaffold pipeline script with your initiative name. This creates the loop structure spanning research, brief, execute, ingest, evaluate, and learn stages, then begins dispatching the workflow.

Yes, this orchestration approach supports warm-start execution for an existing pipeline. It allows you to resume a closed-loop initiative by leveraging the state and artifacts already stored in the loop folder, bypassing initial scaffolding.

The best way to evaluate cross-stage performance is using the orchestrator's transition logs and eval-only mode. By pausing at gates and storing artifacts in the loop workspace, you can perform cross-stage evaluation against the logged state before proceeding.