x-collect-skill 1.0.0 → 1.0.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +22 -2
- package/README.zh-CN.md +148 -0
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -1,7 +1,22 @@
|
|
|
1
|
+
<div align="center">
|
|
2
|
+
<a href="https://github.com/mindfold-ai/Trellis">
|
|
3
|
+
<img src=".github/trellis-banner.png" width="600" alt="Trellis">
|
|
4
|
+
</a>
|
|
5
|
+
<p>
|
|
6
|
+
This skill is built with <a href="https://github.com/mindfold-ai/Trellis"><b>Trellis</b></a> — an all-in-one AI framework & toolkit for Claude Code, Cursor & iFlow.<br>
|
|
7
|
+
<a href="https://github.com/mindfold-ai/Trellis">Try it out here</a>
|
|
8
|
+
</p>
|
|
9
|
+
</div>
|
|
10
|
+
|
|
11
|
+
---
|
|
12
|
+
|
|
1
13
|
# x-collect
|
|
2
14
|
|
|
3
15
|
[](https://www.npmjs.com/package/x-collect-skill)
|
|
4
16
|
[](./LICENSE)
|
|
17
|
+
[](https://discord.gg/tjCB6tVPAm)
|
|
18
|
+
|
|
19
|
+
English | [中文](./README.zh-CN.md)
|
|
5
20
|
|
|
6
21
|
X/Twitter topic intelligence skill for Claude Code. Uses Playwright browser automation to search x.com directly, scraping real tweets with engagement metrics.
|
|
7
22
|
|
|
@@ -36,10 +51,11 @@ claude mcp add --scope user playwright -- npx @playwright/mcp@latest
|
|
|
36
51
|
|
|
37
52
|
## Install
|
|
38
53
|
|
|
39
|
-
### Option A:
|
|
54
|
+
### Option A: Claude Plugin Marketplace (Recommended)
|
|
40
55
|
|
|
41
56
|
```bash
|
|
42
|
-
|
|
57
|
+
claude plugin marketplace add SamCuipogobongo/x-collect
|
|
58
|
+
claude plugin install x-collect
|
|
43
59
|
```
|
|
44
60
|
|
|
45
61
|
### Option B: npm global install
|
|
@@ -123,6 +139,10 @@ The skill uses Playwright MCP to:
|
|
|
123
139
|
|
|
124
140
|
No third-party API needed — reads directly from x.com via Playwright browser automation.
|
|
125
141
|
|
|
142
|
+
## Community
|
|
143
|
+
|
|
144
|
+
Questions, feedback, or feature requests? Join our [Discord](https://discord.gg/tjCB6tVPAm).
|
|
145
|
+
|
|
126
146
|
## License
|
|
127
147
|
|
|
128
148
|
MIT
|
package/README.zh-CN.md
ADDED
|
@@ -0,0 +1,148 @@
|
|
|
1
|
+
<div align="center">
|
|
2
|
+
<a href="https://github.com/mindfold-ai/Trellis">
|
|
3
|
+
<img src=".github/trellis-banner.png" width="600" alt="Trellis">
|
|
4
|
+
</a>
|
|
5
|
+
<p>
|
|
6
|
+
本 Skill 由 <a href="https://github.com/mindfold-ai/Trellis"><b>Trellis</b></a> 构建 — Claude Code、Cursor 与 iFlow 的一站式 AI 框架与工具集。<br>
|
|
7
|
+
<a href="https://github.com/mindfold-ai/Trellis">立即体验</a>
|
|
8
|
+
</p>
|
|
9
|
+
</div>
|
|
10
|
+
|
|
11
|
+
---
|
|
12
|
+
|
|
13
|
+
# x-collect
|
|
14
|
+
|
|
15
|
+
[](https://www.npmjs.com/package/x-collect-skill)
|
|
16
|
+
[](./LICENSE)
|
|
17
|
+
[](https://discord.gg/tjCB6tVPAm)
|
|
18
|
+
|
|
19
|
+
[English](./README.md) | 中文
|
|
20
|
+
|
|
21
|
+
X/Twitter 话题情报采集工具，Claude Code 专用 Skill。通过 Playwright 浏览器自动化直接搜索 x.com，抓取真实推文和互动数据。
|
|
22
|
+
|
|
23
|
+
## 功能
|
|
24
|
+
|
|
25
|
+
`/x-collect [话题]` 打开 x.com，执行 3 轮搜索，提取真实推文数据：
|
|
26
|
+
|
|
27
|
+
1. **热门推文** — x.com 搜索 "Top" 标签页，算法排序的高热推文
|
|
28
|
+
2. **24h 趋势推文** — Top 标签页 + `min_faves:50 since:昨天`，过去 24 小时最热内容
|
|
29
|
+
3. **KOL 推文** — `min_faves:100` 过滤，只看高互动账号
|
|
30
|
+
|
|
31
|
+
可选附加轮次：
|
|
32
|
+
|
|
33
|
+
4. **Hook 研究** — 纯文字推文（`-filter:media -filter:links`），500+ 赞，用于研究文案模式
|
|
34
|
+
5. **对话引爆点** — 高回复推文（`min_replies:30`），找到引发讨论的话题
|
|
35
|
+
|
|
36
|
+
输出：JSONL + Markdown 保存在 `./x-collect-data/`，包含账号、推文正文、点赞、转发、回复、浏览量，以及内容机会分析。
|
|
37
|
+
|
|
38
|
+
## 前置条件
|
|
39
|
+
|
|
40
|
+
| 依赖 | 用途 | 必需 |
|
|
41
|
+
|------|------|------|
|
|
42
|
+
| [Claude Code](https://claude.ai/claude-code) | 运行环境 | 是 |
|
|
43
|
+
| [Playwright MCP](https://github.com/microsoft/playwright-mcp) | 浏览器自动化 | 是 |
|
|
44
|
+
| 已登录的 X/Twitter 会话 | 浏览器需已登录 x.com | 是 |
|
|
45
|
+
|
|
46
|
+
### 安装 Playwright MCP
|
|
47
|
+
|
|
48
|
+
```bash
|
|
49
|
+
claude mcp add --scope user playwright -- npx @playwright/mcp@latest
|
|
50
|
+
```
|
|
51
|
+
|
|
52
|
+
## 安装
|
|
53
|
+
|
|
54
|
+
### 方式 A：Claude Plugin Marketplace（推荐）
|
|
55
|
+
|
|
56
|
+
```bash
|
|
57
|
+
claude plugin marketplace add SamCuipogobongo/x-collect
|
|
58
|
+
claude plugin install x-collect
|
|
59
|
+
```
|
|
60
|
+
|
|
61
|
+
### 方式 B：npm 全局安装
|
|
62
|
+
|
|
63
|
+
```bash
|
|
64
|
+
npm install -g x-collect-skill
|
|
65
|
+
```
|
|
66
|
+
|
|
67
|
+
### 方式 C：一行脚本
|
|
68
|
+
|
|
69
|
+
```bash
|
|
70
|
+
curl -fsSL https://raw.githubusercontent.com/SamCuipogobongo/x-collect/main/install.sh | bash
|
|
71
|
+
```
|
|
72
|
+
|
|
73
|
+
### 方式 D：手动安装
|
|
74
|
+
|
|
75
|
+
```bash
|
|
76
|
+
git clone https://github.com/SamCuipogobongo/x-collect.git
|
|
77
|
+
mkdir -p ~/.claude/skills/x-collect
|
|
78
|
+
cp x-collect/SKILL.md ~/.claude/skills/x-collect/
|
|
79
|
+
```
|
|
80
|
+
|
|
81
|
+
## 使用
|
|
82
|
+
|
|
83
|
+
```bash
|
|
84
|
+
# 在 Claude Code 中（确保 Chrome 已打开并登录 x.com）：
|
|
85
|
+
/x-collect Claude Code # 采集 "Claude Code" 话题
|
|
86
|
+
/x-collect vibe coding # 采集 "vibe coding" 话题
|
|
87
|
+
/x-collect # 交互模式（会询问话题）
|
|
88
|
+
```
|
|
89
|
+
|
|
90
|
+
## 输出
|
|
91
|
+
|
|
92
|
+
数据保存在 `./x-collect-data/`：
|
|
93
|
+
|
|
94
|
+
| 文件 | 说明 |
|
|
95
|
+
|------|------|
|
|
96
|
+
| `intel.jsonl` | 每行一条推文 JSON，按 URL 去重 |
|
|
97
|
+
| `intel.md` | 格式化报告，含互动数据 + 内容机会分析 |
|
|
98
|
+
|
|
99
|
+
### JSONL 字段
|
|
100
|
+
|
|
101
|
+
```json
|
|
102
|
+
{
|
|
103
|
+
"url": "https://x.com/bcherny/status/123",
|
|
104
|
+
"text": "I'm Boris and I created Claude Code...",
|
|
105
|
+
"intel_type": "viral_post",
|
|
106
|
+
"topic": "Claude Code",
|
|
107
|
+
"account": "@bcherny",
|
|
108
|
+
"display_name": "Boris Cherny",
|
|
109
|
+
"angle": "创始人分享 15+ 并行会话的原始配置",
|
|
110
|
+
"format": "thread",
|
|
111
|
+
"likes": 5800,
|
|
112
|
+
"retweets": 1200,
|
|
113
|
+
"replies": 340,
|
|
114
|
+
"views": 6500000,
|
|
115
|
+
"posted_at": "2026-01-15T10:30:00Z",
|
|
116
|
+
"key_takeaway": "创始人幕后故事 + 实用技巧 = 超高互动",
|
|
117
|
+
"collected": "2026-02-08"
|
|
118
|
+
}
|
|
119
|
+
```
|
|
120
|
+
|
|
121
|
+
### 情报类型
|
|
122
|
+
|
|
123
|
+
| 类型 | 来源 |
|
|
124
|
+
|------|------|
|
|
125
|
+
| `viral_post` | x.com 搜索 "Top" 标签页 |
|
|
126
|
+
| `trending_post` | Top 标签页 + `min_faves:50 since:昨天` |
|
|
127
|
+
| `kol_post` | `min_faves:100` 过滤 |
|
|
128
|
+
| `hook_study` | 纯文字推文（`-filter:media -filter:links`），500+ 赞 |
|
|
129
|
+
| `conversation_starter` | 高回复推文（`min_replies:30`） |
|
|
130
|
+
|
|
131
|
+
## 工作原理
|
|
132
|
+
|
|
133
|
+
通过 Playwright MCP 浏览器自动化：
|
|
134
|
+
1. 导航到 x.com 搜索页（Top 标签页 + 各种过滤条件）
|
|
135
|
+
2. 等待推文加载
|
|
136
|
+
3. 通过 DOM 查询或无障碍树快照提取推文数据
|
|
137
|
+
4. 按推文 URL 去重
|
|
138
|
+
5. 输出 JSONL + Markdown，附带内容机会分析
|
|
139
|
+
|
|
140
|
+
无需第三方 API，直接通过 Playwright 浏览器自动化读取 x.com。
|
|
141
|
+
|
|
142
|
+
## 社区
|
|
143
|
+
|
|
144
|
+
问题、反馈或功能建议？加入我们的 [Discord](https://discord.gg/tjCB6tVPAm)。
|
|
145
|
+
|
|
146
|
+
## License
|
|
147
|
+
|
|
148
|
+
MIT
|