@akshay_pachaar: Hermes /learn explained. agents usually learn the hard way. they struggle through a task live, fail a few times, find t…

X AI KOLs Timeline Tools

Summary

Hermes Agent by Nous Research introduces /learn, a command that lets the agent deliberately create skills from documentation, code, or instructions without needing to first fail at the task, turning any source into a reusable skill.

Hermes /learn explained. agents usually learn the hard way. they struggle through a task live, fail a few times, find the path that works, and only then write down what they figured out. the lesson costs you a painful session before it becomes reusable. Hermes Agent by Nous Research just shipped /learn, and it removes that struggle from the front. point it at a source and it builds a skill before it has ever run the task. a local SDK directory, a docs URL, a workflow you just walked it through, or a procedure you paste in as plain notes. it reads the material, writes a 𝗦𝗞𝗜𝗟𝗟.𝗺𝗱, tests the skill live, and saves it. so /learn an internal REST client, tell it to focus on auth and pagination, and you get back a skill that covers exactly that, ready to invoke as a slash command. here is the part worth understanding. Hermes already created skills on its own. after a complex task or a recovery from a dead end, a background pass would quietly capture what it learned. that loop only fires on completed work, so it learns from trajectories the agent has already finished. /learn is the deliberate version. you invoke it, and it learns from material the agent has never touched. docs it has not read, a repo it has not run, someone else's runbook. the mechanism is the cleanest part. there is no separate ingestion engine. /learn builds a standards-guided prompt and hands it to the agent as a normal turn. the agent gathers the material with tools it already has, then authors the skill and saves it through the same skill tooling. because it is just a turn, it works everywhere the agent does. the CLI, the messaging gateways, the dashboard, with nothing new to deploy. the advantages stack up from there. → onboarding collapses to one command. a private API that meant re-reading the docs every session becomes a skill your whole team invokes. → a repo turns into a playbook. point /learn at a codebase and it captures the patterns and workflows, so the next session starts from how to work with it. → a one-time walkthrough becomes repeatable. deploy the staging server once, /learn it, and the procedure outlives the session. → the output is verifiable. every skill ships with a verification section and gets tested live before saving, so you get a draft that already ran. memory remembers facts. skills remember how to do the work. /learn is the front door to the second one, and it lets you fill that store on purpose instead of waiting for the agent to earn each entry the hard way. the best builders stopped re-teaching the agent the same procedure every session. they hand it the source once and keep the skill. That said, if you’re looking to set up Hermes, I wrote a full deep dive covering the Hermes agent’s architecture, memory system, self-evolving skills, GEPA optimization, and how to set up multiple specialized agents. The article is quoted below. You can also watch my YouTube crash course on the Hermes agent: https://youtube.com/watch?v=bNp6YcKBLgY…
Original Article
View Cached Full Text

Cached at: 06/25/26, 03:23 PM

Hermes /learn explained.

agents usually learn the hard way.

they struggle through a task live, fail a few times, find the path that works, and only then write down what they figured out. the lesson costs you a painful session before it becomes reusable.

Hermes Agent by Nous Research just shipped /learn, and it removes that struggle from the front.

point it at a source and it builds a skill before it has ever run the task. a local SDK directory, a docs URL, a workflow you just walked it through, or a procedure you paste in as plain notes.

it reads the material, writes a 𝗦𝗞𝗜𝗟𝗟.𝗺𝗱, tests the skill live, and saves it.

so /learn an internal REST client, tell it to focus on auth and pagination, and you get back a skill that covers exactly that, ready to invoke as a slash command.

here is the part worth understanding. Hermes already created skills on its own.

after a complex task or a recovery from a dead end, a background pass would quietly capture what it learned. that loop only fires on completed work, so it learns from trajectories the agent has already finished.

/learn is the deliberate version. you invoke it, and it learns from material the agent has never touched. docs it has not read, a repo it has not run, someone else’s runbook.

the mechanism is the cleanest part. there is no separate ingestion engine.

/learn builds a standards-guided prompt and hands it to the agent as a normal turn. the agent gathers the material with tools it already has, then authors the skill and saves it through the same skill tooling.

because it is just a turn, it works everywhere the agent does. the CLI, the messaging gateways, the dashboard, with nothing new to deploy.

the advantages stack up from there.

→ onboarding collapses to one command. a private API that meant re-reading the docs every session becomes a skill your whole team invokes.

→ a repo turns into a playbook. point /learn at a codebase and it captures the patterns and workflows, so the next session starts from how to work with it.

→ a one-time walkthrough becomes repeatable. deploy the staging server once, /learn it, and the procedure outlives the session.

→ the output is verifiable. every skill ships with a verification section and gets tested live before saving, so you get a draft that already ran.

memory remembers facts. skills remember how to do the work.

/learn is the front door to the second one, and it lets you fill that store on purpose instead of waiting for the agent to earn each entry the hard way.

the best builders stopped re-teaching the agent the same procedure every session. they hand it the source once and keep the skill.

That said, if you’re looking to set up Hermes, I wrote a full deep dive covering the Hermes agent’s architecture, memory system, self-evolving skills, GEPA optimization, and how to set up multiple specialized agents.

The article is quoted below.

You can also watch my YouTube crash course on the Hermes agent: https://youtube.com/watch?v=bNp6YcKBLgY…


TL;DR

Hermes agent 是一个拥有三层记忆系统、自我进化技能(GAPA)和身份层(soul.md)的开源 AI 代理框架,你使用越多它越强大,通过记忆、技能和身份的无缝协作实现持久化、个性化与自动进化。


Hermes Agent 概述

Hermes agent 上线仅两个月便收获 16 万颗星标,成为增长最快的开源项目之一。它区别于普通 Agent 的关键在于:会话关闭后,Hermes 不会忘记一切;相反,它会记住你与它做过的所有事情,使用越多就越强大。本大师课将涵盖三层记忆系统、自我进化技能(GAPA 技术,已被 ICLR 2026 接收),以及三个可 24/7 为你工作的 Agent 实例:程序员 Neo、设计师 Pixel 和深度研究员。


架构基础

单一 AI Agent 类与多入口

所有功能都流经一个核心的 run_agent.py 中的 AI Agent 类。你可以通过 CLI、Telegram、批处理运行器或 IDE 等入口点访问它。这种设计实现了平台无关性——同一 Agent 可在不同界面下工作。

模型无关的翻译层

Hermes 几乎可以与任何大语言模型配合使用。翻译层通过三种 API 格式路由所有提供者,因此你可以在 cloud code、GPT、Gemini 之间轻松切换,甚至可以使用 Olama 连接本地模型,而不会破坏任何功能。

90 轮硬上限安全措施

AI Agent 类周围有虚线框代表每个任务 90 轮的硬上限。这是防止 Agent 陷入循环、重试失败 API 或重新读取同一文件时耗尽 API 积分的关键保护。子 Agent 也共享同一预算。此上限可以按需调整,但团队默认提供了这一重要护栏。


身份层 (soul.md)

记忆告诉 Agent 它知道什么,技能告诉 Agent 它做什么,但两者都没有告诉 Agent 它是谁。没有身份层,每个 Agent 都像戴着不同帽子的同一个 Agent。

soul.md 存储在根 Hermes 文件夹中,占据系统提示中的第一个位置(位置一),之后依次是记忆(位置二)、技能文件(位置三),最后是对话历史(位置四)。 soul.md 是你写一次、随时间调整,并在每个项目、每个会话中保持一致的固定框架。记忆和技能是可移动的部分,而知识整合是通过 soul.md 的视角发生的,正如反馈箭头所表示的那样。


三层记忆系统

Hermes 没有单一記憶層,而是三层,各自有着不同的持久化策略。

第一层:即时记忆(始终在上下文)

由两个微小的 Markdown 文件组成:

  • memory.md:保持约 2200 字符,记录 Agent 关于环境、项目约定、工具以及通过艰难方式学到的笔记。
  • user.md:保持约 1375 字符,存储你的个人资料,如名字、沟通偏好、技能水平、你想避免的事情。

这两个文件在会话开始时作为冻结快照摄入系统提示,因此每轮对话都在上下文中,无需额外代码。

第二层:全文搜索 SQLite

你与 Hermes 的每一次对话(CLI、Telegram 等)都存储到启用了全文搜索的 SQLite 数据库中。即使搜索几周前的聊天记录也是完全可搜索的,按需进行。权衡是:第一层始终在上下文中但很小,第二层无限但需要搜索 + LLM 总结。

第三层:外部记忆提供者

Hermes 与八个即插即用的外部记忆提供者合作。你可以连接你最喜欢的记忆存储,如基于知识图谱的记忆、时态知识图谱等。集成很容易,文档链接可在描述中找到。

底线:关键事实存在于记忆中(第一层),其他一切可搜索(第二层),当你需要更深入的持久化时连接外部提供者(第三层)。


技能与自我进化

技能定义与渐进式揭示机制

技能是一个 Markdown 文件,以 YAML 前置元数据开头,包含技能名称和描述。你可能有数百个技能,但不会全部加载到上下文中以避免浪费 token。在级别零,Agent 加载所有技能的前置元数据(一小部分),然后确定哪些技能与当前任务相关。在级别一,它逐步揭示技能的主体——过程、陷阱和验证。大多数任务在级别一完成。如果需要,技能还可以引用其他内容。

自我进化技能(GAPA)

Hermes 具有自我进化技能的机制(这在其他开源 Agent 中通常不存在)。GAPA(一种在不改变模型权重的情况下改进提示并发挥最佳性能的技术)论文已被 ICLR 2026 接收。通过 GAPA,Agent 可以自行改进技能文件,根据失败和反馈优化技能定义,从而不断提升性能。


三个实战 Agent

Neo - 程序员

Neo 是一个内部使用 cloud code 的程序员。它可以在你分配任务时首先生成计划,问几个后续问题,确定计划并记录规格后才开始构建。例如,它曾对我进行深度研究(我做什么、管理哪些平台、公司、 newsletter 等),然后建立了一个个人着陆页(index.html),包含我发布 AI 和机器学习内容的平台、影响力等,并包含了写作、课程、GitHub 等交互标签。你可以通过 Telegram 随时随地委托它。

Pixel - 设计师

Pixel 了解我的品牌,知道如何创建设计。我写了一篇文章,其中所有设计都是由 Pixel 自己创建的。例如,我让它创建一张非常漂亮的手绘图表说明 LLM 推理中的预填充和解码阶段。它选对了我预先创建的自定义技能(包含我的品牌价值、设计风格、示例),然后生成了一个干净、符合品牌规范的插图。如果你有自己的一系列设计,也可以作为示例让 Pixel 理解并创建一致的技能。

深度研究员

深度研究员扫描所有最新的 GitHub 仓库、论文、AI 和机器学习领域的热门新闻,让我了解世界正在发生的事情。它可以自定义焦点领域,24/7 工作。


如何开始

你可以直接跳到“启动和运行”部分,命令可以独立运行。但理解技能进化、记忆构成和 GAPA 的价值是区分将 Hermes 用作带有节点的聊天工具还是用作随时间累积的系统的关键。你使用得越多,Agent 越能与你配合。


Source: YouTube 视频

Similar Articles