During the use of Qwen Code CLI, in the command line conversation process, when encountering file creation or editing, it always requires manual confirmation:
I want to save trouble - create files and edit files automatically, auto-accept new files, and auto-allow file editing.
I checked the documentation, and this is how you modify the configuration file:
"tools": {
"approvalMode": "auto-edit"
}
My complete configuration file: C:\Users\admin\.qwen\settings.json
{
"security": {
"auth": {
"selectedType": "qwen-oauth"
}
},
"model": {
"name": "coder-model"
},
"ui": {
"accessibility": {
"enableLoadingPhrases": false
}
},
"$version": 3,
"permissions": {
"allow": [
"Bash(npm install *)"
]
},
"tools": {
"approvalMode": "auto-edit"
}
}
After configuration, during each conversation, Qwen Code CLI will automatically create files and modify code files without manual confirmation each time. The configuration file is persistently stored, and each conversation will apply this configuration.
Before configuration, there was a prompt like this every time:
Now I will create the es.html file Apply this change? │ │ › 1. Yes, once │ 2. Yes, always │ 3. No, suggest changes (esc)
Very annoying.
After modifying the configuration file, it's worry-free - completely automatic programming, no need to wait for me to select options every time to create or edit files.
There's also a more automated configuration, besides auto-editing files, it also auto-executes in the command line every time (with risks)
"tools": {
"approvalMode": "yolo"
}
Currently supported modes in QwenCode:
Approval Mode │ 1. plan - Analyze only, don't modify files or execute commands │ 2. default - Require approval for file edits or shell commands │ › 3. auto-edit - Auto-approve file edits │ 4. yolo - Auto-approve all tools
Recently, Alibaba's newly released large model Qwen 3.6 plus is much better than 3.5 - things that couldn't be programmed before can now be done.