3. ✅ Command Validation & Error Detection

Before proceeding with execution, the bot uses the LLM to perform natural-language-based input validation.

Examples of validations:

• Check format: Is the address a valid Ethereum address?

• Check completeness: Are all required parameters provided?

• Check semantic correctness:

  • “Deploy a token with 0 supply” → flagged as invalid

  • “Send 200 tokens” → flagged if token name or recipient missing

• Check token limits: Warns if supply exceeds practical values (e.g., >1e+30)

The LLM can also rewrite or normalize malformed prompts to make them executable:

“Can u make me a coin, like 1 bil DogeCoin?” → Interpreted as name = "DogeCoin", symbol = "DOGE", supply = 1000000000