Execute End-to-End Coding Tasks Automatically with /goal
Use the /goal command to let AI autonomously execute complex tasks.
Initiate autonomous coding sessions with the /goal command to allow the AI to handle multi-step tasks, complete with independent validation and error detection.
The Scenario
You have a clear coding task, such as fixing a specific bug or refactoring a module, and you want the AI to handle the end-to-end execution without constant manual oversight.
Before & after
Manually investigating a bug, writing a fix, and running tests typically requires your full attention for 30–60 minutes depending on complexity.
Using Qwen Code's /goal command, the agent autonomously plans, executes, and validates the fix with an independent judge model. This takes roughly 2–4 minutes while you focus on other work.
The Prompt
/goal [DESCRIBE_YOUR_CODING_OBJECTIVE_E.G._REFACTOR_AUTH_FLOW_TO_USE_JWT]
The /goal command allows the AI to operate agentically, meaning it doesn't just suggest code but attempts to complete the entire requested objective by navigating files and testing its own work.
Source
Qwen Code Blog: AI Coding Updates and Guides | Qwen Code Docs"Qwen Code v0.16.0 follows suit—plus an independent judge model and impossible-goal detection."