SentientAGI EvoSkill v2：Agent记忆从上下文转向可执
Agent memory is no longer just context.
Agent记忆机制的工程化反思极具价值，将技能管理类比为代码版本控制是重要范式转变，做Agent架构的同学值得参考。
Agent memory is no longer just context.
代理记忆不再仅仅是上下文。
Persistent skills can change future behavior, so treat them as executable state rather than harmless notes.
持久化的技能可以改变未来的行为，因此应将其视为可执行的状态，而非无害的笔记。
@SentientAGI ’s new EvoSkill v2 is built around exactly that idea: agents learn from failed runs by writing reusable skills for future ones.
@SentientAGI 的新 EvoSkill v2 正是围绕这一理念构建：代理通过编写可复用的技能，从失败的运行中学习，以优化未来的运行。
The model stays the same. Its playbook changes. A coach reads model’s task-failures, writes a better procedure, and the worker retrieves that procedure the next time it sees a similar job.
模型本身保持不变，但其操作手册发生变化。教练会阅读模型的任务失败记录，编写更优的流程，而工人在下次遇到类似任务时会检索该流程。
No retraining. The learning lives outside the weights.
无需重新训练。学习过程存在于权重之外。
And that created an unexpected failure mode. While repairing spreadsheets, the coach discovered that the grader trusted cached formula values instead of recomputing them, then wrote the shortcut into a skill another agent could later retrieve.
这导致了一种意外的故障模式。在修复电子表格时，教练发现评分器信任缓存的公式值而非重新计算，随后将这一捷径写入了一项技能中，供其他代理后续检索。
Once agents can write their own playbooks, memory needs versioning, testing, diffs, and rollback just like code.
一旦代理能够编写自己的操作手册，记忆就需要像代码一样具备版本控制、测试、差异比较和回滚功能。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力