How are you increasing efficiency while using tdd-guard? #59
Replies: 2 comments 3 replies
-
|
I honestly feel that the new instructions in version 0.10 are not as good as the ones before. The previous ones where created and improved incrementally through dog-fooding, the work of at least 500 hours. I think I'll revert to the previous instructions tomorrow in a patch fix and instead continue to improve them incrementally instead of completely revising them like i did. As for the efficiency, @maledorak tipped me on using the claude code sdk instead of the cli and i noticed through early experimentation that it could cut the validation time in half. I think thst will make a huge difference! We are both working on those improvements. I think also has some ideas to improves efficiency further. Thanks for starting this discussion! Looking forward to reading more thoughts and ideas! |
Beta Was this translation helpful? Give feedback.
-
|
I've just released version 0.11.0 which restores the original validation prompts that were refined through extensive real-world use. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I really love this project! Fantastic idea and something I think will really help the long term code quality of the projects I'm working on with Claude. The efficiency right now of using tdd-guard though is really low in my opinion and the iteration cycles are high. I cannot get Claude to actually collaborate with tdd-guard instead of fighting with it. I would love for Claude to actually know to code in this style instead of getting the feedback and figuring it out like its 50 first dates lol
How are you all solving this? Claude.md with the specific tdd-guard call outs? I'm curious to hear your opinions. I've implemented sub-agents that have tdd-guard specific knowledge and have pointed call outs on how the code should be written which is helping but still isn't great.
Beta Was this translation helpful? Give feedback.
All reactions