I then added a few more personal preferences and suggested tools from my previous failures working with agents in Python: use uv and .venv instead of the base Python installation, use polars instead of pandas for data manipulation, only store secrets/API keys/passwords in .env while ensuring .env is in .gitignore, etc. Most of these constraints don’t tell the agent what to do, but how to do it. In general, adding a rule to my AGENTS.md whenever I encounter a fundamental behavior I don’t like has been very effective. For example, agents love using unnecessary emoji which I hate, so I added a rule:
从“十五五”规划建议提出“持续巩固拓展脱贫攻坚成果”，到2026年中央一号文件明确提出“实施常态化精准帮扶”，着眼的正是确保长久守住不发生规模性返贫致贫底线。
。关于这个话题，谷歌浏览器【最新下载地址】提供了深入分析
Crucially, this distribution of border points is agnostic of routing speed profiles. It’s based only on whether a road is passable or not. This means the same set of clusters and border points can be used for all car routing profiles (default, shortest, fuel-efficient) and all bicycle profiles (default, prefer flat terrain, etc.). Only the travel time/cost values of the shortcuts between these points change based on the profile. This is a massive factor in keeping storage down – map data only increased by about 0.5% per profile to store this HH-Routing structure!
The Ultrahuman Ring Pro comes with a snazzy Pro Charging Case for up to 45 days of additional battery life. | Image: Ultrahuman