agentleFS
Sign inStart free
tinchak0207tinchak0207/Hippocampus/SKILL.md

hippocampus

Skill2 starsChanged 26 days agoMITView on GitHub

让 agent 吸取教训：踩过的坑写进硬盘，下次不再踩。在开始一个非平凡任务前调用，用 `mem` 查这类任务以前的决策和坑；在犯错被纠正、查明一个问题的原因、或做出一个决定之后调用，用 `mem add` + `mem build` 把教训沉淀下来。Use before starting a task to look up past lessons, and after a mistake, correction, or decision to record the lesson.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. reescalder/claude-skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repositoryOne other file it gives its agents.

The file

Create a free account to mount
Files
SKILL.md

Hippocampus：让 agent 吸取教训

踩过的坑写进 notes/，mem build 编译进索引，下次同类任务开始前先查到上次怎么栽的。CLI 叫 mem。

记忆根目录 $MEM_HOME（默认 ~/.claude/memory）。全部是纯文本，mem build 从 notes/ 生成其余三层。

何时使用

  • 开始一个非平凡任务前：查这件事以前有没有做过、有没有决策或坑。
  • 用户提到「上次」「之前说过」「按老规矩」时。
  • 派 subagent 前：把相关记忆写进 spec。
  • 完成一个可复用的结果、解决一个反复出现的问题、被用户纠正一次之后：写一篇 note。

四层模型

层文件内容谁写
Tier-0MEMORY.md路由表：读写协定、pack 列表、最近 12 篇；有字节预算（默认 12000）mem build
Tier-1packs/<id>.md每个领域一页，每篇 note 一行 description + 路径mem build
Tier-2catalog.tsv每篇 note 一行：pack date type name title bytes gistmem build
Tier-3notes/*.md原文，带 YAML front-mattermem add 建文件，你填内容

分类规则在 $MEM_HOME/.memorypack/packs.json：front-matter 里 metadata.pack 写了合法 id 就用它，否则按关键词打分（strong +3，keywords +1），零分进 misc。

读路径（按顺序，够用就停）

  1. 读 MEMORY.md 全文（它很小，就是为了被整篇读）。
  2. 按「什么时候读」列找到领域 → mem pack <id> 扫读该领域。
  3. 需要定位具体条目 → mem find <关键字> [--pack <id>] [--limit N]，只返回 catalog 行。
  4. 只对命中的 1-3 篇 → mem show <name> 读原文。

记忆与现场证据（代码、日志、命令输出）冲突时，以现场为准，并告诉用户哪篇记忆过期了。

写路径（只有这一条）

mem add <slug> --type project|feedback|reference [--pack <id>] [--title "标题"]
# 打开打印出的路径，填 description 一行 + FACT / 决策 / 坑 / 相关
mem build
  • description 是唯一会出现在上三层的内容，写成「这是什么 + 什么时候要用」的一句话。
  • 改已有 note：直接编辑 notes/ 里那一篇，然后 mem build。
  • 写完可以跑 mem doctor 检查预算、断链、缺 description。

禁令

  • 不手改 MEMORY.md、packs/*、catalog.tsv：它们是生成物，下次 build 会被覆盖。
  • 不批量读 notes/（cat notes/*、ls 后逐篇读、对整个目录 grep 原文）：用 mem find。
  • 不删除 note，除非用户明确要求；过期就在原文标注并更新 description。
  • 不把 secret、token、密码、私钥、个人身份信息写进 note。
  • 不另开第二个记忆库；其他工具的旧路径用 symlink 指向 $MEM_HOME。

参考

  • 日常维护：prompts/maintain.md
  • 会话末复盘与主动建设：prompts/proactive-build.md
  • 可粘贴进 AGENTS.md / CLAUDE.md 的规则段：prompts/agents-snippet.md
  • hooks 示例：prompts/hooks/

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free