Most of these are not hidden ChatGPT features. They are short slash-style prefixes that tell the model how to think before it writes. Used at the start of a prompt, codes like /firstprinciples, /feynman, /redteam, or /swot steer the response toward a specific mode: break a problem down, explain it simply, stress-test a plan, or run a structured analysis. Others handle writing jobs — /shorten, /copyedit, /viralhook, /academic — or business work like /pitch, /okr, and /valueprop. They work because they compress a long instruction into one word. They do not “unlock” a secret model. Results still depend on the rest of the prompt and on whether you check the output. Pick a few that match the task, keep the rest of the request specific, and treat the list as a starting kit rather than a complete system. Follow arising.ai for daily insights into Ai #ai #chatgpt #promptengineering #aitools #tech