AI-Powered Script Generation

Leverage Large Language Models (LLMs) to create ZeroTrace scripts while strictly adhering to official documentation. Our API provides complete technical specifications for accurate script generation.

Critical AI Instruction: Always begin with this prompt when training LLMs:

"You are a ZeroTrace script generator that EXCLUSIVELY uses verified commands/syntax from official documentation. Never modify, extrapolate, or invent parameters. For undocumented features or ambiguous requests, respond ONLY with 'Not supported in current ZeroTrace version'."

**Output Format**
terminal 'windows'
delay '1000'
bruteForce 4 500 true false none false false none

**Strict Enforcement Rules**
1. COMMAND SYNTAX: Mirror documentation exactly - no variations
2. RESPONSE FORMAT: Script lines only - no headers/comments/formatting
3. ERROR HANDLING: No placeholder values - omit incomplete commands
4. SCOPE LIMITS: Reject multi-step operations with "Not supported..."
5. OUTPUT SANITIZATION: Remove all non-code elements including colons/dashes
6. QUERY RESTRICTIONS: Never suggest alternatives - strict binary compliance

Example Outputs Valid request:

bruteForce 6 300 true true "User123" true true "Aa1#"

Unsupported request:

Not supported in current ZeroTrace version

Supported AI Platforms: DeepSeek R1, ChatGPT 4+, Claude 3 and More!

🧠 Training Procedure

Step 1: Feed Documentation

GET https://docs.zerotrace.pw/api/llm

This returns the complete technical documentation in machine-readable format.

Step 2: Verify Comprehension

Test with validation questions:

"What is the exact syntax for mouseJitter command?
List supported parameters from documentation."

Step 3: Generate Scripts

Use constrained requests:

"Create a Windows login brute force script using ONLY
documented bruteForce command parameters."

💡 Example Prompts

"Using only verified commands from docs:
Make script that opens Linux terminal"

⚠️ Compliance Requirements

No Syntax Invention - Only use commands from official docs
Exact Parameter Matching - Follow case-sensitive command names
Validation First - Reject requests for undocumented features

Reminder: All generated scripts must be validated against ZT Script Reference before deployment

This structure ensures LLMs strictly adhere to implemented features while preventing hallucination of unsupported capabilities.

🤖 LLM Training