写作Writing
小团队的 agent 工作流需要清楚入口、可见检查点和严格的交接路径。A small-team agent workflow needs clear entry points, visible checks, and a strict handoff path.
小团队并不需要什么 agent 魔法。他们真正需要的，是可以重复、也不会让工作更难被信任的回路。
Agent 更像一个贡献者，而不是替代判断的东西。
如果一条工作流没法在一屏里讲清楚，那对小团队来说，它大概率已经太重了。
我通常就是从这里开始删东西的。
Small teams do not need agent magic. They need repeatable loops that do not make the work harder to trust.
The agent is a contributor, not a replacement for judgment.
If the workflow cannot be explained in one screen, it is probably too much for a small team.
That is usually where I start trimming.
和这篇笔记关联的项目与案例研究。Projects and case studies connected to this note.
项目Projects
把混乱的客户请求整理成可范围化、可交付的 intake 与分流工作流。A structured intake and triage workflow that turns messy client requests into scoped delivery plans.
一组本地 Agent 实验，用来验证研究、编码、评估与交付回路。A set of local agent experiments for research, coding, validation, and repeatable delivery loops.
案例研究Case studies
一个可重复的 intake 管道，减少上下文丢失，让范围定义更快。A repeatable intake pipeline that cuts context loss and makes proposals faster to scope.
一个在进入客户项目之前，用来测试 Agent 流程、trace 与评估回路的内部实验室。An internal lab for testing agent workflows, traces, and evaluation loops before client use.
和这个主题相邻的更多笔记。More notes close to this topic.
一个实用的 AI 工作环境：能起草、测试和发布，又不会把流程缠在一起。A practical setup for drafting, testing, and shipping AI work without letting the workflow get tangled.
怎样组织案例研究，才能证明判断力，而不只是展示结果。How to structure a case study so it proves judgment, not just output.
公共索引真正有用，往往不是因为页面多复杂，而是因为内容模型足够小、足够清楚、足够容易改。A public index gets more useful when the content model is small, explicit, and easy to revise.