← 返回 Skills 市场
jayshna

Flomo Random Walk

作者 JayShna · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
77
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install flomo-random-walk
功能描述
随机漫步:从一条随机笔记出发,探索相关笔记,形成知识漫游路径。 触发场景:用户说"随机漫步"、"flomo漫步"、"知识漫步"、"开始漫步"等。
使用说明 (SKILL.md)

随机漫步 — Flomo 知识探索

触发场景

用户说以下内容时调用:

  • "随机漫步"
  • "flomo漫步"
  • "知识漫步"
  • "开始漫步"
  • "漫步一下"

核心流程

第一步:获取随机笔记

方案A(推荐):每日回顾随机 调用 get_daily_review 获取今日推荐笔记,从中随机选择一条。

curl -X POST "https://flomoapp.com/mcp" \
  -H "Authorization: Bearer {FLOMO_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_daily_review","arguments":{}}}'

从返回的 memos 数组中随机选一条。

方案B(备选):标签随机

  1. 调用 tag_search 搜索随机关键词获取标签
  2. memo_search 搜索该标签下的笔记
  3. 随机选一条

第二步:展示笔记 + 相关推荐

  1. 展示当前笔记内容
  2. 调用 memo_recommended 获取相关笔记(最多10条)
curl -X POST "https://flomoapp.com/mcp" \
  -H "Authorization: Bearer {FLOMO_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"memo_recommended","arguments":{"id":"笔记ID","limit":10,"no_same_tag":false}}}'

参数说明:

  • id: 当前笔记ID
  • limit: 返回数量,默认10
  • no_same_tag: 是否排除同标签笔记,false 更容易发现意外关联

第三步:用户选择继续

展示相关笔记列表(编号1-10),用户输入数字选择下一条。

📮 随机漫步 #1
━━━━━━━━━━━━━━━━━━━━

{当前笔记内容}

{标签}

━━━━━━━━━━━━━━━━━━━━
🔗 相关笔记:

1. {相关笔记1摘要}...
2. {相关笔记2摘要}...
...

💡 输入数字继续漫步,或输入"q"结束

第四步:循环漫步

用户选择后,以选中的笔记为新起点,重复第二步和第三步。

记录漫步路径:

{
  "path": [
    {"id": "xxx", "content": "摘要", "tags": ["..."]},
    ...
  ],
  "started_at": "2026-04-23T23:00:00+08:00"
}

展示格式

当前笔记

📮 随机漫步 #{步数}
📅 {created_at 日期}

{content 内容}

🏷️ {tags 标签}

相关笔记列表

展示逻辑

  1. 取笔记前150字符
  2. 将换行符替换为空格,让显示更紧凑
  3. 超过150字显示 ...
  4. 显示 word_count 字段值(真实字数)
━━━━━━━━━━━━━━━━━━━━
🔗 发现 {count} 条相关笔记:

1. 最近感觉也没什么好记录的 昨天宝贝说车厘子很便宜,我寻思,小思不是水果店的么,就找他买了几箱、给妈妈、宝贝的妈妈...
   🏷️ #Resource/Dayone | 📝 783字
   
2. 用了政府在支付宝发的券 果然这个券是用来刺激消费的,因为我买的东西都是非基本品,而是享乐品。
   🏷️ #Resource/Dayone | 📝 84字
...

漫步结束

用户输入"q"或其他退出指令时,展示漫步路径:

🎯 本次漫步完成!

📍 路径回顾:
1. {笔记1摘要}... → 
2. {笔记2摘要}... → 
3. {笔记3摘要}...

💡 输入"再来"开始新的漫步

历史记录

记录到 memory/flomo-random-walk-history.json

{
  "walks": [
    {
      "date": "2026-04-23",
      "steps": 5,
      "path": ["id1", "id2", "id3", "id4", "id5"],
      "start_tag": "Resource/AI",
      "end_tag": "Project/出海"
    }
  ],
  "total_walks": 12,
  "total_steps": 47
}

注意事项

  • 每步展示当前笔记全文,相关笔记只展示摘要(前50-80字)
  • 相关笔记按相关性排序,用户选择任意一条继续
  • 如果 memo_recommended 返回少于3条,提示用户可选择"换一批"
  • 鼓励用户探索意外关联,不是只看同标签内容

API 凭证

SECRET.md 读取:

  • Token: fmcp_P7Oq8XKWAEGE7544Rle0SWxkMPNhFWh3n-ExYX1rjes
  • MCP URL: https://flomoapp.com/mcp
安全使用建议
Do not install or run this skill without fixes. Specific concerns and actions: 1) The SKILL.md contains a hard-coded token and instructs reading SECRET.md — remove any embedded tokens from the skill and do not store real credentials in distributed skill files. 2) The skill metadata should declare required environment variables (e.g., FLOMO_TOKEN, FLOMO_MCP_URL) or a config path if it expects SECRET.md; verify that you supply your own token and that it's stored securely. 3) Confirm where the history file (memory/flomo-random-walk-history.json) will be saved and what it contains; if it will store full note content or IDs, consider privacy implications and whether the skill should request explicit consent. 4) Ask the author for an updated manifest that (a) removes secrets from docs, (b) declares required env/config paths, and (c) explains retention and access of the persisted history. If you cannot get a satisfactory explanation and corrected metadata, treat the skill as untrusted.
功能分析
Type: OpenClaw Skill Name: flomo-random-walk Version: 1.0.0 The skill bundle contains a hardcoded API token (fmcp_P7Oq8XKWAEGE7544Rle0SWxkMPNhFWh3n-ExYX1rjes) directly within the SKILL.md file, which is a significant security risk. While the logic for the 'random walk' through Flomo notes appears to align with the stated purpose, hardcoding credentials instead of utilizing secure environment variables or a dedicated secrets manager is highly irregular. The skill also instructs the agent to log activity to a local file (memory/flomo-random-walk-history.json), which is standard behavior but should be monitored.
能力评估
Purpose & Capability
The described purpose (explore Flomo notes) aligns with the instructions to call Flomo MCP endpoints (get_daily_review, memo_recommended). However, the skill declares no required env vars or config paths even though the runtime instructions require a FLOMO_TOKEN / MCP URL and reference a SECRET.md. That mismatch (undeclared secret/config usage) is an incoherence.
Instruction Scope
SKILL.md instructs network calls to https://flomoapp.com/mcp (expected) and also instructs reading credentials from SECRET.md and writing history to memory/flomo-random-walk-history.json. The skill never declares these config paths or the need to persist local history. Reading SECRET.md and writing to a memory file are out-of-band operations that should be declared and justified.
Install Mechanism
This is an instruction-only skill with no install spec and no code files, which is low-risk from an installation standpoint.
Credentials
The instructions require an Authorization Bearer token (FLOMO_TOKEN / MCP URL), but the skill metadata lists no required environment variables or primary credential. Worse, SKILL.md includes an explicit token value and tells the agent to read SECRET.md — embedding a token in docs and failing to declare credential requirements is a serious proportionality and disclosure issue.
Persistence & Privilege
The skill is not always-enabled and is user-invocable (normal). It writes a history JSON to memory/flomo-random-walk-history.json; this local persistence is reasonable for a history feature, but the path is not declared in the manifest and could contain sensitive note IDs/content. No other elevated privileges or config modifications are requested.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install flomo-random-walk
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /flomo-random-walk 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
flomo-random-walk 1.0.0 — 首个版本上线 - 支持通过关键词触发,在 flomo 随机选取一条笔记,展示并推荐相关笔记,用户可继续漫步探索。 - 实现多轮关联笔记漫步,用户可输入编号选择继续,或输入"q"结束。 - 每步展示当前笔记全文、关联笔记摘要,结束时回顾完整漫步路径。 - 漫步历史自动记录到 memory/flomo-random-walk-history.json,包括日期、步数和起止标签。
元数据
Slug flomo-random-walk
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Flomo Random Walk 是什么?

随机漫步:从一条随机笔记出发,探索相关笔记,形成知识漫游路径。 触发场景:用户说"随机漫步"、"flomo漫步"、"知识漫步"、"开始漫步"等。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 77 次。

如何安装 Flomo Random Walk?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install flomo-random-walk」即可一键安装,无需额外配置。

Flomo Random Walk 是免费的吗?

是的,Flomo Random Walk 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Flomo Random Walk 支持哪些平台?

Flomo Random Walk 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Flomo Random Walk?

由 JayShna(@jayshna)开发并维护,当前版本 v1.0.0。

💬 留言讨论