4. 🛠️ Instruction Generation (Intermediate Execution Plan)

After understanding and validating the user input, the LLM generates a structured execution plan. This plan is used by the system’s backend modules (compiler, deployer, signer) to carry out the command.

Example Output (Internal Representation):

jsonCopyEdit{
  "intent": "deploy_erc20",
  "params": {
    "name": "MyToken",
    "symbol": "MYT",
    "supply": 1000000,
    "decimals": 18
  },
  "network": "goerli",
  "action_required": "wallet_connect"
}

This JSON-like structure becomes the “instruction packet” for the deployment pipeline, enabling modular execution.