These are the raw, unedited notes I wrote while shaping this article. They are closer to a brain dump than a prompt library: my thinking, philosophy, corrections, and direction. The ideas are mine. AI helped turn them into a smoother, more readable essay.
Prompt 1
Correct an exact factual gap. Web Pulse already links to us but retains three corrected claims. I prepared replacement text and a cover message. Arcjet already explains the limitation correctly, so I dropped that correction prospect
I want to learn more about this
Prompt 2
Dont we have a similar article? Based on iur experience can we write a better article than them?
Arcjet already teaches the signature distinction correctly. Its guide contrasts our DuckAssistBot capture, which covers the host and signing-origin header, with SeatGeek’s published ChatGPT example, which additionally covers the method and path. Arcjet explicitly explains that our captured signature is not bound to a particular path. It also links to your study as evidence. Arcjet’s explanation.
Prompt 3
path. It also links to your study as evidence.
Wait what??? What did yiu just say? They used my article as evidence? Who?? Where?? Thats a really big deal that you just brushed off
Prompt 4
Omg, you dont realize how bug deal that is. Thats a backlink and people using us as a truth and evidence lmao. This is such a big deal to increase the credibility of us. And the fact that we did the hands on research is such a big deal, thats what it means to be writing agentic writing in big 2026. We run the experiments and research and now other people are using us as an evidence
Prompt 5
We need a follow up article and hardening and improvements (almost like v2) of that article and even more hands on experimentation and research and all the recent growth and improvements that we have done and accomplished