← 返回 Skills 市场
theopenbase

Obsidian Read

作者 theopenbase · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
175
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install obsidian-read
功能描述
读取 Obsidian Publish 发布的文档链接。当用户发送 publish.obsidian.md 链接,或需要读取 Obsidian 笔记文章时使用。 适用于从 Obsidian Publish 平台获取完整文章内容、笔记、Essay 等场景。 优先使用 web_fetch,失败时使用 browser...
使用说明 (SKILL.md)

Obsidian Read

读取 Obsidian Publish 发布的文档内容。

工作流

步骤 1:优先尝试 web_fetch

web_fetch(url="链接", maxChars=15000)

如果成功返回完整内容,直接使用。

步骤 2:使用 browser 方案(当 web_fetch 失败时)

当 web_fetch 返回内容不完整、404 或解析错误时,使用以下流程:

2.1 打开链接获取 targetId

browser(action="open", url="https://publish.obsidian.md/a/essay/xxx")

返回结果包含 targetId,例如:"579ECBB9218EE0FA843D5CACD2BD99F0"

2.2 读取页面内容

browser(action="snapshot", targetId="上一步获取的targetId")

2.3 提取文本

从 snapshot 结果中提取 paragraphheading 等文本元素,返回给用户。

注意事项

  • 必须用 open 获取 targetId:直接使用 snapshot 而不传 targetId 会导致失败
  • 静态内容优先:Obsidian Publish 是静态站点,snapshot 通常能完整抓取
  • 不需要截图:snapshot 是获取可访问性树(文本内容),不是截图

示例

用户发送:https://publish.obsidian.md/a/essay/202603181132

  1. 先尝试 web_fetch → 失败/不完整
  2. browser.open → 获取 targetId
  3. browser.snapshot → 读取完整文章
  4. 总结文章内容返回给用户
安全使用建议
This skill appears coherent and minimal: it only fetches public Obsidian Publish pages using web_fetch and a browser snapshot fallback. Before installing, consider: (1) any link you or other users supply will be fetched (so avoid pasting private links you do not want retrieved); (2) the browser tool will load the provided URL and extract text — ensure users understand that content fetched from arbitrary URLs may be presented to the agent. No credentials or installs are required, so there are no hidden requests for secrets.
功能分析
Type: OpenClaw Skill Name: obsidian-read Version: 1.0.0 The obsidian-read skill is designed to extract content from Obsidian Publish links using standard web_fetch and browser tools. The instructions in SKILL.md provide a logical workflow for retrieving document text and do not contain any evidence of malicious intent, data exfiltration, or unauthorized execution.
能力评估
Purpose & Capability
The name/description (read Obsidian Publish links) matches the runtime instructions which only call web_fetch and a browser tool to open and snapshot publish.obsidian.md pages. No unrelated binaries, env vars, or config paths are requested.
Instruction Scope
SKILL.md is focused and prescriptive: try web_fetch first, then use browser.open to obtain a targetId and browser.snapshot to extract paragraph/heading text. It does not instruct reading local files, other env vars, or exfiltrating data to third parties outside the target site.
Install Mechanism
No install spec or code files are present (instruction-only), so nothing will be written to disk or downloaded. This is the lowest-risk install profile and is appropriate for the described capability.
Credentials
The skill declares no environment variables, secrets, or config paths. That matches its need to fetch public Obsidian Publish pages; no extra credentials are requested.
Persistence & Privilege
always is false and the skill does not request to modify agent config or persist credentials. Autonomous model invocation is allowed (platform default) but that is appropriate for a web-fetching skill.
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 obsidian-read
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /obsidian-read 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 - 读取 Obsidian Publish 文档
元数据
Slug obsidian-read
版本 1.0.0
许可证 MIT-0
累计安装 2
当前安装数 1
历史版本数 1
常见问题

Obsidian Read 是什么?

读取 Obsidian Publish 发布的文档链接。当用户发送 publish.obsidian.md 链接,或需要读取 Obsidian 笔记文章时使用。 适用于从 Obsidian Publish 平台获取完整文章内容、笔记、Essay 等场景。 优先使用 web_fetch,失败时使用 browser... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 175 次。

如何安装 Obsidian Read?

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

Obsidian Read 是免费的吗?

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

Obsidian Read 支持哪些平台?

Obsidian Read 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 Obsidian Read?

由 theopenbase(@theopenbase)开发并维护,当前版本 v1.0.0。

💬 留言讨论