Set-and-Forget Complex Tasks with the /goal Orchestrator
Define high-level objectives and let the agent manage execution.
Set a high-level coding target using /goal to let the agent orchestrate the entire workflow from start to finish with automated risk assessment.
The Scenario
You need to implement a complete feature or complex bug fix but don't want to micro-manage the individual steps of the logic.
Before & after
You would manually break down a feature request into 10 sub-tasks and manually verify each one as you code them, taking 1–2 hours.
Initialize `/goal` with a specific outcome; an independent judge model tracks progress and auto-approves low-risk steps, completing the task in 2–5 minutes.
The Prompt
/goal [DESCRIBE_YOUR_FINAL_DESIRED_OUTCOME_IN_DETAIL]
The /goal command uses an independent judge model to determine completion and can automatically abandon tasks if it detects they are impossible, saving you from dead-end loops.
Source
Qwen Code Blog: AI Coding Updates and Guides | Qwen Code Docs"/goal + an independent judge model to determine completion, with automatic abandonment when it detects an impossible goal."