How AI Verifies Unfamiliar Commands Before Use

One of the smartest things Vex does is check how a command works before using it. This pre-command knowledge check prevents frustrating trial-and-error loops where AI assistants guess at parameters and fail repeatedly.

How it works:

1. Vex encounters an unfamiliar command
2. Calls get_tool_info("command.name") or what_is("command.name")
3. Reads the full documentation: parameters, types, descriptions, examples
4. Uses the command correctly on the first try

When checks happen:

• First use of any command in a session
• Commands with complex parameters (like unity.batch_operation)
• When Vex is uncertain about required vs. optional fields

Example: You ask Vex to validate a path. Vex has not used workspace.diagnose yet. Instead of guessing, Vex calls what_is("workspace.diagnose"), learns the required parameters, then calls workspace.diagnose action="check" path="Assets/Scripts/Player.cs" correctly the first time. No errors. No retries.

Why this matters: Without checks, Vex would guess, fail, read the error, adjust, and retry — wasting time and cluttering your conversation. With checks, commands work the first time. This is enforced by Law 6 in GameCatalyst’s behavioral system. It is built into how Vex operates at the deepest level, ensuring reliable results.

What Vex learns from a knowledge check:

When Vex calls what_is("workspace.diagnose"), it receives the complete command specification: the command name, required and optional parameters with their types, a description of what the command does, aliases (workspace.diagnose can also be called as diagnose.workspace or path.check), and working examples showing correct usage. This is everything Vex needs to call the command correctly.

Built into the behavioral laws:

Pre-command knowledge checks are not just a good habit — they are required by Law 6 in GameCatalyst’s behavioral system. This law is classified as high priority and applies to all command categories. Vex is instructed to verify unfamiliar commands before calling them, and the periodic five-minute reminders reinforce this behavior throughout your session.

The result for you:

You see clean, successful command results instead of error messages and retries. Vex works confidently and efficiently because it verifies before acting. This is one of the many ways GameCatalyst makes AI-assisted development more productive than using unguided AI tools.

Knowledge checks are a small upfront investment that pays off with reliable, error-free command execution throughout your session.