run

Agent orchestration coordinates multiple agents by decomposing work into milestones, assigning tasks, running an execution loop, and reconciling state across sessions automatically. It uses a centralized orchestrator to manage dependencies and validate outcomes without manual micromanagement.

A self-healing pipeline automatically rewrites or re-plans tasks when execution results fail acceptance criteria. It validates outcomes during the execution loop and triggers automatic recovery to keep software projects moving without manual intervention.

Workflow orchestration applies deterministic planning to decompose work into milestones and tasks, managing dependencies through a centralized orchestrator. It assigns tasks to agents, validates results, and reconciles state across sessions to ensure execution order is predictable.

Autonomous orchestration is designed for software projects requiring state reconciliation across sessions. It runs an end-to-end execution loop that validates outcomes, manages dependencies, and applies automatic recovery via a self-healing pipeline to maintain project consistency.

Start by invoking the run skill with a plan request to decompose your project into milestones and tasks, then execute to run the autonomous loop. This assigns work to agents, validates results, and monitors progress automatically.

Avoid autonomous execution loops when tasks lack clear acceptance criteria for result validation, as the self-healing pipeline relies on validating outcomes to trigger automatic recovery and re-planning. Projects requiring manual micromanagement or non-deterministic workflows are also poor fits.