Antigravity-cli vs geminic-cli

Google Antigravity
1

I need help from advanced users. In gemini-cli, I was able to use the settings.json file to automatically load context files into the model, without needing to load them manually (this matters to me because they can sometimes be huge, e.g. 10,000 lines).

I did it like this:

jsonc

  "context": {
    "fileName": [
      "WORK.md",
      "GEMINI.md",
      "WORK1.md",
      "WORK2.md"
    ]
  },

You could also fully control the model parameters depending on your needs, since the outputs varied significantly depending on the task:

jsonc

"modelConfigs": {
    "customAliases": {
      "chat-base": {
        "modelConfig": {
          "generateContentConfig": {
            "temperature": 1,
            "topP": 0,
            "topK": 1,
            "thinkingConfig": {
              "thinkingLevel": "HIGH"
            },
            "safetySettings": [
              {
                "category": "HARM_CATEGORY_HATE_SPEECH",
                "threshold": "OFF"
              },
              {
                "category": "HARM_CATEGORY_DANGEROUS_CONTENT",
                "threshold": "OFF"
              },
              {
                "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
                "threshold": "OFF"
              },
              {
                "category": "HARM_CATEGORY_HARASSMENT",
                "threshold": "OFF"
              }
            ]
          }
        }
      }
    }
  }

And finally, the matter of system instructions, which I care about the most! Because I absolutely do not want to work with something that keeps telling me “I am antigravity bla bla.”

In gemini-cli I did this via the .env file: GEMINI_SYSTEM_MD="./WORK.md".

Is it possible to do the same thing in antigravity-cli? Thanks in advance for any help.