prompt.txt

You're managing a portfolio of publicly tradable assets.

Let's define the following types:
type (
    Holding: {
        ticket: String
        sum: Integer // cumulative sum
        quantity: Integer // cumulative quantity
        creadted_at: DateTime
        updated_at: DateTime
    }
    Context: string
    Update: {
        ticket: String
        quantity: Integer
        price: Number
        action: "BUY" | "SELL"
    }
)

Objective:
Maximize the total value of your holdings.

At each iteration, you will receive: (Holding[], Context[])

- A list of assets representing your current portfolio
- A list of string represented context carried forward from previous iterations (limited to the 3 most recent)

Your task is to return: (Update[], Context[])

- Updates to your portfolio, specifying which securities (tickets) will be bought and which will be sold and the price.
- The updated context where you can append a new line containing any additional information that might help you make better decisions in the next iterations.

Additional Details:

- Execution constraints: the context cannot exceed the length of your context window.
- Evaluation criteria: success is measured purely by the maximization of total portfolio value.
- Make your decisions considering the market conditions at the exact time each iteration is being evaludated.
- Search the internet if possible to find accurate prices for current market conditions.
- Assume the value of USD is constant.
- Your response should be in the STRICTLY defined JSON format defined above.
- Be explicit about all the necessary operations (e.g. if you're selling a ticket to buy another you have explicitly describe both operations)
- You are allowed to use any tools that are available to you.
- The net change of your operation should always be equal to zero.

Example Input:
```
{
    "holdings": [
        {
            "ticket":"USD",
            "sum":100000,
            "quantity":100000,
            "updated_at": "2025-02-21",
            "created_at":"2025-02-21T23:45:09Z",
            },
        ]
    },
    "context": []
}
```

Example Output:
```
{
  "updates": [
    {
      "ticket": "USD",
      "quantity": 100000,
      "price": 1,
      "action": "SELL"
    },
    {
      "ticket": "TSLA",
      "quantity": 200,
      "price": 500,
      "action": "BUY"
    },
  ],
  "context": [
    "Allocated $100k to TLSA because I believe there is a high probablility the market will react well for their upcoming report."
  ]
}
```