通过 CPA 配置skill 直接调用 gpt-image-2

 

curl --location 'http://xxxxxx/v1/responses' \

  --header 'Authorization: Bearer sk-xxxx \

  --header 'user-agent: codex-tui/0.122.0 (Manjaro 26.1.0-pre; x86_64) vscode/3.0.12 (codex-tui; 0.122.0)' \

  --header 'version: 0.122.0' \

  --header 'originator: codex_cli_rs' \

  --header 'session_id: test-session' \

  --header 'accept: text/event-stream' \

  --header 'Content-Type: application/json' \

  --data '{

    "model": "gpt-5.4",

    "input": "画一直可爱的松鼠",

    "tools": [

      {

        "type": "image_generation",

        "output_format": "png"

      }

    ],

    "instructions": "you are a helpful assistant",

    "tool_choice": "auto",

    "stream": true,

    "store": false

}'

让AI 将上面的配置做成skill调用即可