WHAT NEW IN X11.01 SUN

X 上的 AI 合集档案 · 22 个合集 · 63,170 条

jevVIDEO

让 Jev 决定 coding agent 该跑哪些测试

演示 agent 按改动文件判断测试范围:只改文档就跳过、动到支付就跑全量,跳过时置信度低于 0.8 就照跑不误。

CyrilXBT
@cyrilXBT19 天前收录原推发布于 2026.10.11
CyrilXBT @cyrilXBT原推发布于 2026.10.11
在 X 上打开原推

POV: your coding agent decides which tests to run "Changed http://auth.py. Running the auth tests, not all 900." "Touched payments. Full suite. No shortcuts." "Docs only. Skipping tests." "Confidence 0.71 on skipping. Below 0.8. Running them anyway." Fast when it's safe. Paranoid when it isn't. That's Jev making the call. How to build it 👇

这条是引用着另一条写的:x.com/1373665408193036293/status/210869352490011…

x.com/cyrilxbt/status/2109211212735230102采集于 2026.10.12 10:05