Turn ChatGPT Into a Cross-Disciplinary Strategist With the God Mode V2 Custom Instructions

Custom instructions that make ChatGPT think across disciplines

God Mode V2 Prompt

By the end of this guide, you'll have a 5-principle configuration pasted into ChatGPT's Custom Instructions turning every chat into a structured strategy session with diagnosis, framing, red-team stress-testing, and a force multiplier recommendation.

Who This Is For

  • Founders and operators who get generic, cautious output from ChatGPT and want something sharper

  • Consultants or strategists who want AI to engage with problems structurally, not conversationally

  • Anyone tired of AI that hedges, restates the question, and buries the answer in filler

What "God Mode" actually is: A community-developed custom instructions configuration — not an official ChatGPT feature. It works by changing the system-level behavior of every conversation, forcing a structured 5-step output pattern instead of the default conversational response style.

What You'll Build

A persistent custom instructions setup where every ChatGPT response follows this structure:

  1. Diagnose — root cause analysis before any answer

  2. Frame — reframe the problem through a strategic lens

  3. Advance — specific, actionable directives with no filler

  4. Stress-Test — red team section identifying risks and failure points

  5. Elevate — a "Force Multiplier" to increase impact beyond the immediate solution

What You'll Need

  • A ChatGPT account (free or paid — custom instructions work on both)

  • 5 minutes to paste and configure

For Claude Users

Claude doesn't have a "Custom Instructions" field in the same location, but you get the same effect two ways:

  • Claude Projects — open a Project → set the system prompt in the Project instructions field. Paste the same configuration above.

  • Start of every chat — paste the configuration as the first message in a new conversation before your actual question.

Going Further

  • Domain-specific versions — once the base configuration is working, add a domain layer: "All strategic framing should prioritize SaaS growth metrics" or "All risk assessment should weight regulatory compliance first"

  • Reset when needed — if a conversation drifts back to generic output, type: "Return to God Mode protocol. Apply all 5 principles to your next response."

  • Limitation to know — structured output quality scales with input quality. A vague question still gets a structured response — but it structures a vague answer. Be specific about what you're actually trying to solve.