Module 6: Prompt Reverse-Engineering (Safely)
Module 6 of 9
What This Teaches
How to infer questions without prompt injection.
Correct Framing: You are not manipulating prompts. You are modeling question distributions. Related: Precog Modeling
Steps
- Identify primary user question: What is the main query users will ask?
- Identify follow-up questions: What questions naturally follow the primary question?
- Identify trust questions: What questions test credibility or safety?
- Identify safety constraints: What boundaries must answers respect?
Prechunking Rule
If a question can be asked, its answer must already exist as a chunk.
Don't wait for questions. Anticipate them and publish answers in advance.
Optional Operator Task
Task: Identify one primary question users ask about your service. List the next five follow-up questions they would naturally ask. For each question, identify what chunk must exist to answer it.
Constraint: Each follow-up question must be logically necessary, not speculative. Each required chunk must be a single atomic fact.
What success looks like: You produce a question tree where every question has a corresponding chunk that answers it. You've modeled the question distribution and ensured answers exist before questions are asked.
This task is optional. No submission required. No validation. Use it to convert theory into applied thinking.