DRAFT Protocol FAQs

Question 1

How does DRAFT Protocol prevent AI errors and scope creep?

Accepted Answer

It structures intent elicitation across Define, Rules, Artifacts, Flex, and Test dimensions. By identifying ambiguous or missing information and requiring human confirmation before AI execution, DRAFT prevents 'assumed intent' issues, ensuring AI acts precisely on what you mean.

Question 2

What are the DRAFT Protocol's adaptive scrutiny tiers?

Accepted Answer

DRAFT offers Casual, Standard, and Consequential tiers. It automatically classifies requests based on potential impact, applying appropriate levels of rigor – from internal mapping for simple tasks to maximum scrutiny with mandatory dimensions and Devil's Advocate checks for high-impact actions.

Question 3

What is DRAFT Protocol?

Accepted Answer

DRAFT Protocol is an intake governance system for AI tool calls. It ensures explicit understanding of human intent by mapping requests across five critical dimensions and preventing AI execution until that intent is confirmed, thereby preventing errors and misunderstandings.

Question 4

Is DRAFT Protocol compatible with my existing AI clients and tools?

Accepted Answer

Yes, DRAFT Protocol boasts broad compatibility. It integrates seamlessly with AI clients supporting Model Context Protocol (MCP) transports like stdio, SSE, and Streamable HTTP. A universal Chrome extension also extends its governance to popular web-based AI chat platforms like ChatGPT, Claude, and Copilot.

Question 5

Does DRAFT Protocol require a Large Language Model (LLM) to function?

Accepted Answer

No, DRAFT Protocol works out-of-the-box using keyword matching and heuristics for full governance. However, it offers optional LLM integration with providers like Ollama, OpenAI, and Anthropic for enhanced semantic classification, embedding-based field assessment, and context-aware suggestions, improving accuracy and intelligence.

DRAFT Protocol

DRAFT Protocol

Key Features

Use Cases

Key Features

Use Cases