workflow

You can orchestrate multi-step engineering workflows without writing custom code by running a workflow orchestrator that sequences steps across skills, announces each stage, and waits for user confirmation before proceeding. It supports feature, bugfix, and hotfix scenarios.

Yes, you can define a custom project workflow in your repository and run it through the orchestrator to guide the sequence. This allows you to apply step-by-step orchestration to your specific engineering processes while maintaining guardrails.

A workflow orchestrator maintains handoff integrity by preserving plan-ceo and plan-eng outputs for downstream steps. This ensures that data and context from the planning phase are consistently passed to the engineering phase without loss.

The best way to ensure safe, incremental progress is to use an orchestrator that implements guardrails by announcing each step and waiting for user confirmation. This provides an auditable trail from planning to QA.

No specific dependencies are required to run a multi-skill workflow in the cursor-stack. The orchestrator operates independently without external dependencies, allowing you to sequence tasks immediately.

You should avoid using an automated workflow orchestrator for feature development when your project requires bypassing step-by-step announcements and user confirmation guardrails, as the orchestrator enforces these incremental checkpoints for safe progress.