bookmark

分类: 内容与多媒体 | 上传者: kengio | 下载: 0 | 版本: v1.0（最新）

快速 URL 书签捕获：粘贴链接，AI 生成名称和描述，建议类别，以 Awesome-list 格式保存到 Bookmarks.md。当用户想要保存 URL 供以后使用时调用 - 没有其他上下文的裸 URL 默认为此。请勿用于：立即深入处理或总结 URL 内容（使用汇总）、保存不是 URL 的注释（使用捕获）或从头开始研究主题（使用研究）。

更新日志: Source: GitHub https://github.com/kengio/onebrain

目录结构

当前层级: tree/main/.claude/plugins/onebrain/skills/bookmark/

📄 SKILL.md 3.5 KB

SKILL.md

---
name: bookmark
description: "Quick URL bookmark capture : paste a link, AI generates name and description, suggests category, saves to Bookmarks.md in awesome-list format. Invoke when user wants to save a URL for later — bare URL with no other context defaults to this. Do NOT use for: deeply processing or summarizing the URL content now (use summarize), saving a note that is not a URL (use capture), or researching a topic from scratch (use research)."
---

# Bookmark

Save a URL to your `Bookmarks.md` file in one step. Paste the link : the AI fills in the name and description, picks a category, and saves immediately.

Usage: `/bookmark [url]`

---

## Step 1: Get the URL

If a URL was provided after the command, use it directly.

If not, ask:
> What URL do you want to bookmark?

---

## Step 2: Fetch the Page

Fetch the page content. Extract:

- **Name**: page title or product name (prefer `<title>` or `<h1>`)
- **Description**: 1-line summary (~15 words) from meta description or page body

**If fetch fails**, ask:
> I couldn't fetch that URL. What should I call it, how would you describe it in one line, and what category fits best?

Use the user's answers and continue normally.

---

## Step 3: Pre-Save Checks

**Duplicate check:** grep `[resources_folder]/Bookmarks.md` for the URL (`[resources_folder]` from Step 2). If already present, tell the user and stop (unless they confirm to save again).

**Existing summary note:** grep `[resources_folder]/**/*.md` for `url: [URL]` in frontmatter. If found, record its title for use as a wikilink in the entry.

---

## Step 4: Pick Category

Infer the best category and optional subcategory from the content : do not ask. Use the structure:

- **Level 1 (`##`)**: broad domain : `AI Tools`, `Design`, `Dev Utilities`, `Productivity`, `Reading`, `Learning`, `Finance`, `Health`, `Reference`
- **Level 2 (`###`)**: optional refinement : only when it adds meaningful grouping

Create new categories freely when none fit.

---

## Step 5: Save to Bookmarks.md

Build the entry:

```markdown
- **[Name](URL)** : Description. → [[Summary Note Title]]
```

(Omit the wikilink if no summary note was found in Step 3.)

File path: `[resources_folder]/Bookmarks.md` (`[resources_folder]` from Step 2).

**If the file does not exist**, create it:

```markdown
---
tags: [bookmarks, resources]
created: YYYY-MM-DD
updated: YYYY-MM-DD
---

# Bookmarks
```

Append under the correct `##` / `###` section (alphabetical order). Create missing sections in alphabetical position. Refresh `updated` in frontmatter.

---

## Step 6: Confirm

Say in one line:
🔖 Saved to `Bookmarks.md` under `## {Category}`. [If subcategory: `/ ### {Subcategory}`]

---

## Recategorize

If the user asks to move or recategorize a bookmark:

1. Resolve file path using `[resources_folder]`
2. Find the entry by name, or use the last bullet for "last bookmark"
3. Remove from current section; append to target section (create `##` / `###` as needed)
4. Refresh `updated` in frontmatter
5. Confirm in one line:
   🔖 Moved {Name} from `## {Old}` → `## {New}`. [If subcategory: `/ ### {Sub}`]

---

## Known Gotchas

- **Fetch failure with a bare URL.** If the page cannot be fetched (timeout, 404, auth-gated), ask the user for name, description, and category rather than stopping. A bookmark with manual metadata is still useful.

- **Category auto-inference.** Infer category from page content, not just the domain. A GitHub repo for a productivity tool belongs under `Productivity` or `Dev Utilities`, not under a generic `GitHub` category.

登录后下载/点赞/收藏 ❤ 6 | ★ 0

bookmark

目录结构

SKILL.md

举报内容

提示