[jakozaur]

See x thread for rationale: https://x.com/mitchellh/status/2014433315261124760?s=46&t=FU...

“ Ultimately, I want to see full session transcripts, but we don't have enough tool support for that broadly.”

I have a side project, git-prompt-story to attach Claude Vode session in GitHub git notes. Though it is not that simple to do automatic (e.g. i need to redact credentials).

[ollien]

Not sure how I feel about transcripts. Ultimately I do my best to make any contributions I make high quality, and that means taking time to polish things. Exposing the tangled mess of my thought process leading up to that either means I have to "polish" that too (whatever that ends up looking like), or put myself in a vulnerable position of showing my tangled process to get to the end result.

[radarsat1]

I've thought about saving my prompts along with project development and even done it by hand a few times, but eventually I realized I don't really get much value from doing so. Are there good reasons to do it?

[simonw]

For me it's increasingly the work. I spend more time in Claude Code going back and forth with the agent than I do in my text editor hacking on the code by hand. Those transcripts ARE the work I've been doing. I want to save them in the same way that I archive my notes and issues and other ephemera around my projects.

My latest attempt at this is https://github.com/simonw/claude-code-transcripts which produces output like the is: https://gisthost.github.io/?c75bf4d827ea4ee3c325625d24c6cd86...

[radarsat1]

Right, I get that writing prompts is "the work", but if you run them again you don't get the same code. So what's the point of keeping them? They are not 'source code' in the same sense as a programming language.

[simonw]

That's why I want the transcript that shows the prompts AND the responses. The prompts alone have little value. The overall conversation shows me exactly what I did, what the agent did and the end result.

[radarsat1]

> shows me exactly what I did

I get that, but I guess what I'm asking is, why does it matter what you did?

The result is working, documented source code, which seems to me to be the important part. What value does keeping the prompt have?

I'm not trying to needle, I just don't see it.

[simonw]

It's like issues in that it helps me record why I had the agent solve problems in a particular way.

It's also great for improving my prompting skills over time - I can go back and see what worked.

[fragmede]

It's not for you. It's so others can see how you arrived to the code that was generated. They can learn better prompting for themselves from it, and also how you think. They can see which cases got considered, or not. All sorts of good stuff that would be helpful for reviewing giant PRs.

[Ronsenshi]

Sounds depressing. First you deal with massive PRs and now also these agent prompts. Soon enough there won't be any coding at all, it seems. Just doomscrolling through massive prompt files and diffs in hopes of understanding what is going on.

[verdverm]

I suspect this future will not play out. Mitchell is definitely leaning to one side on this debate.

To me, quality code is quality code no matter how it was arrived at. That should be the end of it

[verdverm]

Using them for evals at a future date.

I save all of mine, including their environment, and plan to use them for iterating on my various system prompts and tool instructions.

[awesan]

If the AI generated most of the code based on these prompts, it's definitely valuable to review the prompts before even looking at the code. Especially in the case where contributions come from a wide range of devs at different experience levels.

At a minimum it will help you to be skeptical at specific parts of the diff so you can look at those more closely in your review. But it can inform test scenarios etc.

[optimalsolver]

>I want to see full session transcripts, but we don't have enough tool support for that broadly

I think AI could help with that.

[stevenhuang]

simow wrote a tool that does this for Claude code

https://simonw.substack.com/p/a-new-way-to-extract-detailed-...

[couchdb_ouchdb]

You should be able to attach the plan file that you and Claude develop in Plan mode before even starting to code. This should be the source of truth.

[empath75]

On our team, we have discussed attaching claude transcripts to jira tickets, not github PRs (though the PRs are attached to tickets)