Achieve Complex Coding Objectives with Autonomous /goal Sessions
Use /goal to let AI finish complex coding tasks autonomously.
Use the /goal command to initiate autonomous coding sessions where the agent self-corrects and pursues a final objective until completion.
The Scenario
You have a complex, multi-step coding task like refactoring a module or implementing a persistent feature, and you don't want to babysit the AI through every file change.
Before & after
A developer manually tracks state, runs tests, and applies fixes step-by-step for a complex bug, often taking 45–60 minutes of focused attention.
Run the /goal command, and the agent will use an independent judge model to detect impossible goals and execute sub-tasks autonomously, finishing in 2–4 minutes.
The Prompt
/goal [DESCRIBE_YOUR_CODING_OBJECTIVE_E.G._REFACTOR_AUTH_LOGIC_TO_USE_JWT] --auto-approve
Qwen Code v0.16.0 introduced the /goal command which uses an independent judge model to ensure the agent reaches the defined objective without human hand-holding. Change the way you interact with coding agents by defining the end result rather than the specific steps.
Source
Qwen Code Blog: AI Coding Updates and Guides | Qwen Code Docs"/goal autonomous coding... Qwen Code v0.16.0 follows suit—plus an independent judge model and impossible-goal detection."