Love this! I'm literally taking on a new project on Monday and will need to get up to speed fast, I'll definitely steal some of the prompts.
One of the things I did before is that I told Claude code to build me an interactive tutorial through the codebase that teaches me both about the repo and the language used (since I have never used it before). It created a RAG application that walks me through the key parts but also lets me ask questions as I go. Highly recommend that approach as well!
Great prompts, thanks for sharing. One adaption I like is to ask it to write to HTML and use SVG to create diagrams. That allows a more visual and interactive output, at the cost of some tokens.
The onboarding use case is underrated as an entry point for AI adoption in engineering teams. From an IT budget perspective, reducing time-to-productivity for new engineers has a clear, measurable ROI — which makes it easier to justify the AI tool spend internally. The pattern we see is that companies who start with clear, bounded use cases like this build the internal confidence to expand AI tooling more broadly. The cognitive load problem you're describing (turning tangled complexity into structured understanding) is exactly what AI excels at when given the right context. Great framing here.
I think skills may be a bit overkill to be honest - if you copy/paste some of these prompts it should be sufficient. If you want, you can turn them into a skill (Claude Code and Codex have good skill creator UX)
Thanks for having me on to share! I hope these tips help ICs and leaders alike get the most out of AI :)
Appreciate you for sharing your insights Jeff! This will definitely help many to get started faster/easier.
Love this! I'm literally taking on a new project on Monday and will need to get up to speed fast, I'll definitely steal some of the prompts.
One of the things I did before is that I told Claude code to build me an interactive tutorial through the codebase that teaches me both about the repo and the language used (since I have never used it before). It created a RAG application that walks me through the key parts but also lets me ask questions as I go. Highly recommend that approach as well!
Great to hear that! That's an interesting approach, definitely something to try, thanks for sharing Kacper.
This is super cool! I hadn't thought of asking for an interactive tutorial
Great prompts, thanks for sharing. One adaption I like is to ask it to write to HTML and use SVG to create diagrams. That allows a more visual and interactive output, at the cost of some tokens.
The onboarding use case is underrated as an entry point for AI adoption in engineering teams. From an IT budget perspective, reducing time-to-productivity for new engineers has a clear, measurable ROI — which makes it easier to justify the AI tool spend internally. The pattern we see is that companies who start with clear, bounded use cases like this build the internal confidence to expand AI tooling more broadly. The cognitive load problem you're describing (turning tangled complexity into structured understanding) is exactly what AI excels at when given the right context. Great framing here.
Is there a skill on git you have that you can share?
I think skills may be a bit overkill to be honest - if you copy/paste some of these prompts it should be sufficient. If you want, you can turn them into a skill (Claude Code and Codex have good skill creator UX)
Thanks
Great stepsc and prompts. Thanks for sharing!
Any pointers on how to handle large codebases? Where there are hundreds of libraries all separate projects.
I think these same principles apply, with some extra emphasis on your harness. I do like Augment Code (especially intent) for big codebases