← 返回 Skills 市场
jayshna

flomo-archive

作者 JayShna · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
115
总下载
1
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install flomo-archive
功能描述
获取 Flomo 指定月份的完整笔记并评估质量。当用户说"查看flomo XX年XX月的全部笔记"、"获取flomo某月笔记"、"拉取flomo历史记录"或"评估flomo笔记质量"时触发使用。支持智能三级降级策略(周→天)确保完整覆盖,内置质量评估规则自动标记低质量笔记。
使用说明 (SKILL.md)

Flomo 笔记归档工具

功能

  1. 完整获取 - 获取 Flomo 指定月份的所有笔记,处理 API 50条/次的限制
  2. 质量评估 - 自动评估笔记质量,标记低质量/碎片化记录

核心策略

Flomo 的 memo_search API 单次最多返回 50 条笔记。本工具采用智能分层策略:

  1. 按周获取 - 默认粒度,大部分情况足够
  2. 降级按天 - 当某周返回 50 条时,降级为按天获取该周
  3. ID 去重 - 处理周边界重叠,确保笔记不重复

使用方法

获取某月笔记(带质量评估)

python3 scripts/fetch_month.py \x3C年> \x3C月>

示例:

python3 scripts/fetch_month.py 2025 7

仅评估已有数据

python3 scripts/evaluate_quality.py /tmp/flomo_YYYY_MM.json

质量评估规则

规则 结果 说明
有标签 + 字数≥50 高质量 已分类且内容充实
含链接/图片 高质量 有多媒体内容
字数\x3C20 低质量 字数过少
无标签 + 字数\x3C100 低质量 建议人工审阅
疑似流水账关键词 低质量 如"起床了""晚安"等

输出格式

  • 默认:统计摘要 + 质量报告 + 保存到 /tmp/flomo_YYYY_MM.json
  • JSON 模式:--json 输出完整 JSON
  • 详细模式:-v 显示每日获取进度

技术细节

  • API 限制:单次最多 50 条
  • 调用频率:每次调用间隔 100ms
  • 数据已包含:content, tags, word_count, created_at 等完整字段
  • 无需额外 batch_get 调用

注意事项

  • 需要 mcporter 已配置 flomo MCP 服务器
  • 结果文件保存在 /tmp/ 目录,需及时转存
安全使用建议
This skill largely does what it says, but be aware it will run the 'mcporter' CLI using a config file at ~/.openclaw/workspace/config/mcporter.json (the code assumes that path). The registry metadata did not declare that config path or any credentials, so installing or running the skill will cause the agent to use whatever MCP credentials are stored there. Before installing or allowing autonomous runs: (1) verify you trust the skill source; (2) inspect the included scripts yourself (they are small); (3) ensure mcporter is configured with an account you are comfortable the skill using (consider a limited/test account); (4) avoid running it with elevated or multi-service credentials stored in mcporter; (5) consider running the scripts manually rather than granting the agent autonomous invocation if you are unsure. If you need full assurance, ask the author to declare required config paths/credentials in the manifest and to avoid hard-coded config locations or use a configurable path/explicit credential environment variables.
功能分析
Type: OpenClaw Skill Name: flomo-archive Version: 1.0.0 The skill contains a shell injection vulnerability in `scripts/fetch_month.py` where `subprocess.run(shell=True)` is used with string concatenation to execute the `mcporter` CLI tool. While the script currently constrains inputs (year and month) to integers, the pattern of building shell commands from variable arguments is high-risk. Additionally, the script explicitly accesses the OpenClaw workspace configuration file at `~/.openclaw/workspace/config/mcporter.json` to facilitate API calls.
能力评估
Purpose & Capability
Name/description match the implementation: scripts fetch notes (via mcporter) and evaluate quality. Using an MCP/mcporter client is reasonable for contacting Flomo MCP APIs.
Instruction Scope
SKILL.md and scripts instruct the agent to call the 'mcporter' CLI with a hard-coded config path in the user's home (~/.openclaw/workspace/config/mcporter.json). The manifest declared no required config paths or credentials, but the runtime relies on that config and on mcporter being present. That mismatch is scope-creep: the tool will indirectly access the user's stored MCP credentials/config when mcporter runs.
Install Mechanism
No install spec; code files are included and run directly. There are no network downloads or extract steps in the install. This is lower-risk than fetching remote binaries.
Credentials
The skill declares no required env vars or config paths, yet the code expects a specific mcporter config file in the user's home. Running mcporter will use whatever credentials/config are present there (sensitive). The script also passes the full environment to subprocess.run and builds a shell command with shell=True, increasing the chance of unintended environment or shell-injection influence if values were different. These accesses are not declared in the registry metadata.
Persistence & Privilege
always is false, the skill does not request permanent platform-wide presence, and it only writes output files to /tmp. It does not modify other skills or system-wide settings.
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-archive
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /flomo-archive 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
- Initial release of flomo-archive. - Fetches all Flomo notes from a specified month, bypassing API 50-notes-per-request limit via week→day smart fallback. - Automatic note quality evaluation included, based on content length, tags, media, and keywords. - Outputs summary statistics, detailed quality report, and saves results as JSON to /tmp. - Console options for quality-only evaluation and JSON/verbose output. - Requires mcporter with flomo MCP server. - Handles duplicate entries and API rate limiting automatically.
元数据
Slug flomo-archive
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-archive 是什么?

获取 Flomo 指定月份的完整笔记并评估质量。当用户说"查看flomo XX年XX月的全部笔记"、"获取flomo某月笔记"、"拉取flomo历史记录"或"评估flomo笔记质量"时触发使用。支持智能三级降级策略(周→天)确保完整覆盖,内置质量评估规则自动标记低质量笔记。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 115 次。

如何安装 flomo-archive?

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

flomo-archive 是免费的吗?

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

flomo-archive 支持哪些平台?

flomo-archive 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-archive?

由 JayShna(@jayshna)开发并维护,当前版本 v1.0.0。

💬 留言讨论