← Back to Skills Marketplace
Bilibili Analytics
by
pcjinglang
· GitHub ↗
· v1.0.0
906
Downloads
0
Stars
7
Active Installs
1
Versions
Install in OpenClaw
/install bilibili-analytics
Description
Bilibili视频搜索与数据分析。抓取指定关键词的搜索结果,生成统计报告。支持多页面抓取、数据分析和可视化报告。适用于:(1) 搜索B站视频 (2) 分析视频数据 (3) 生成统计报告 (4) 追踪热门趋势
README (SKILL.md)
Bilibili Analytics
搜索和分析B站视频数据,生成统计报告。
快速开始
# 搜索关键词并抓取数据
agent-browser open "https://search.bilibili.com/all?keyword=你的关键词"
# 获取数据
agent-browser eval '抓取脚本'
# 生成报告
# 分析数据并输出统计
完整工作流程
1. 搜索视频
agent-browser open "https://search.bilibili.com/all?keyword={关键词}" --timeout 15000
2. 抓取数据
使用 scripts/scrape_videos.sh 或手动执行:
agent-browser eval '
const videos = [];
document.querySelectorAll(".bili-video-card").forEach((card) => {
const title = card.querySelector(".bili-video-card__info--tit")?.textContent.trim() || "";
const author = card.querySelector(".bili-video-card__info--author")?.textContent.trim() || "";
const date = card.querySelector(".bili-video-card__info--date")?.textContent.trim() || "";
const stats = card.querySelectorAll(".bili-video-card__stats--item");
const playCount = stats[0]?.textContent.trim() || "0";
const commentCount = stats[1]?.textContent.trim() || "0";
videos.push({title, author, date, playCount, commentCount});
});
JSON.stringify(videos, null, 2);
'
3. 多页面抓取
# 翻页抓取
for page in 1 2 3 4 5; do
agent-browser open "https://search.bilibili.com/all?keyword={关键词}&page=$page"
agent-browser eval '抓取脚本' >> data.json
done
4. 数据分析
使用 scripts/analyze_data.py 或手动分析:
- 时间分布统计
- 作者活跃度排名
- 评论数分布
- 播放量分布
- 关键发现和建议
脚本说明
scripts/scrape_videos.sh
一键抓取脚本,支持指定关键词和页数。
./scripts/scrape_videos.sh "关键词" 页数
scripts/analyze_data.py
数据分析脚本,生成统计报告。
python scripts/analyze_data.py data.json
输出格式
统计报告模板
## 📊 Bilibili "{关键词}" 搜索结果统计报告
### 📈 总体数据
- 数据范围: 前N页搜索结果
- 视频总数: X个
- 采集时间: YYYY-MM-DD HH:MM
### 🕐 发帖时间分布
| 时间段 | 数量 | 占比 |
|--------|------|------|
### 👥 活跃作者 TOP 10
| 排名 | 作者 | 视频数 |
|------|------|--------|
### 💬 评论数分布
| 评论数范围 | 视频数 | 占比 |
|------------|--------|------|
### 👁️ 播放量分布
| 播放量范围 | 视频数 | 占比 |
|------------|--------|------|
### 🎯 关键发现
1. ...
2. ...
### 📝 建议
- ...
注意事项
- 反爬虫: B站有反爬虫机制,建议控制抓取频率
- 数据准确性: 数据实时变化,报告仅代表抓取时刻状态
- 隐私合规: 仅抓取公开数据,不涉及用户隐私
错误处理
- 浏览器启动失败:检查 agent-browser 安装
- 数据抓取失败:检查页面是否加载完成
- 分析脚本错误:检查数据格式是否正确
Usage Guidance
This skill appears to be what it claims (scrape Bilibili search results and analyze them), but check these before installing or running:
- Missing declared dependencies: the scripts and SKILL.md require agent-browser (a third-party CLI) and the shell script prints usage of `jq`—these are not listed in the registry metadata. Install and vet agent-browser separately before use.
- Origin & provenance: the package homepage and source are unknown/placeholder; prefer skills from a known repository or inspect the full code locally before running.
- Rate limits and legality: scraping public pages is visible to the target site and may violate terms of service; use conservative scraping rates and confirm compliance.
- Requirements file mismatch: INSTALL.md references requirements.txt (pip) but the manifest doesn't include it—verify Python deps before running analyze_data.py.
- Run in a restricted environment: because the skill opens web pages and writes files, run initially in a sandbox or isolated account and inspect outputs (JSON/report files). If you allow the agent to invoke this autonomously, be aware it can run multi-page scrapes automatically—consider limiting invocation or requiring manual approval.
If you want to proceed, obtain and inspect the real upstream repository (or request the author/homepage), ensure agent-browser and jq are trusted and installed, and test on a small page set first.
Capability Analysis
Type: OpenClaw Skill
Name: bilibili-analytics
Version: 1.0.0
The skill bundle is designed for legitimate Bilibili data analysis. However, the `scripts/scrape_videos.sh` file contains a shell injection vulnerability. The `KEYWORD` argument is directly embedded into the `agent-browser open "$URL"` command without proper sanitization, allowing an attacker to inject arbitrary shell commands if they can control the `KEYWORD` input. This is a critical vulnerability, but there is no clear evidence of intentional malicious behavior by the skill developer, thus classifying it as suspicious rather than malicious.
Capability Assessment
Purpose & Capability
The name/description (Bilibili scraping + analysis) align with the included scripts and SKILL.md. However the package metadata lists no required binaries while both SKILL.md, INSTALL.md, and scripts clearly require the external tool `agent-browser` (and the scrape script also expects `jq` for the final count). This is an omission/incoherence in declared requirements.
Instruction Scope
Runtime instructions and scripts limit activity to opening Bilibili search pages, executing browser-evaluation JavaScript to extract public page data, saving JSON files, and running a local Python analysis. The instructions do not reference unrelated system files, environment secrets, or external endpoints beyond Bilibili and local filesystem. They do instruct multi-page scraping and use sleep to slow requests, but they give the agent discretion to perform repeated requests—so rate limits and anti-scraping behavior should be considered.
Install Mechanism
There is no formal install spec in the registry (instruction-only), which is lower-risk; INSTALL.md provides manual install steps including downloading/unzipping or git cloning. No external binary downloads from untrusted URLs are present in included files. Still, the INSTALL.md asks the user to run `agent-browser install` and to pip-install requirements (requirements.txt is referenced but not present in the file manifest), which are operational gaps to verify.
Credentials
The skill declares no environment variables or credentials and the scripts do not attempt to read secrets. The requests and filesystem writes (creating bilibili_data_*.json and report files) are proportional to the stated data-collection purpose.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and only writes output files in the working directory. It runs on-demand via agent-browser and local scripts; no elevated or persistent platform-level privileges are requested.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install bilibili-analytics - After installation, invoke the skill by name or use
/bilibili-analytics - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
首个版本:B站视频搜索与数据分析
Metadata
Frequently Asked Questions
What is Bilibili Analytics?
Bilibili视频搜索与数据分析。抓取指定关键词的搜索结果,生成统计报告。支持多页面抓取、数据分析和可视化报告。适用于:(1) 搜索B站视频 (2) 分析视频数据 (3) 生成统计报告 (4) 追踪热门趋势. It is an AI Agent Skill for Claude Code / OpenClaw, with 906 downloads so far.
How do I install Bilibili Analytics?
Run "/install bilibili-analytics" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Bilibili Analytics free?
Yes, Bilibili Analytics is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Bilibili Analytics support?
Bilibili Analytics is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Bilibili Analytics?
It is built and maintained by pcjinglang (@pcjinglang); the current version is v1.0.0.
More Skills