bilibili-search

Real-time Bilibili video search that retrieves and extracts top trending video titles as structured JSON via a local API.

Security Scan

The skill is internally consistent: it exposes a local FastAPI/Playwright-based search API for Bilibili and asks the agent to call that local endpoint; it does not request unrelated credentials or install actions.

Purpose & Capability

Name and description match the implementation: SKILL.md describes a local FastAPI service that uses Playwright to scrape Bilibili and return structured JSON. Requiring a local service and browser engine is proportionate to 'real-time search' functionality.

Instruction Scope

Runtime instructions are limited to calling a local API (http://127.0.0.1:8000) and running a local Playwright-based service. The SKILL.md does not instruct reading unrelated files or environment variables. Caution: allowing an agent to call localhost endpoints can expose other local services if the local API is misconfigured or malicious, so the local server's code should be reviewed before running.

Install Mechanism

Instruction-only skill with no install spec or packaged downloads reduces supply-chain risk. The user must manually run their own FastAPI/Playwright service—nothing in the skill automatically installs external code.

Credentials

No environment variables, credentials, or config paths are requested. This is proportionate for a local scraping/search helper that relies on a user-run local service.

Persistence & Privilege

always is false and the skill does not request persistent/system-wide privileges. The skill does require a user-run local service but does not attempt to modify other skills or agent configuration.

Assessment

This skill delegates all web scraping to a local Python FastAPI service using Playwright. Before using: (1) verify and review the skill_api.py source you will run — the agent will call http://127.0.0.1:8000 so malicious or poorly written service code could access local resources; (2) run the service in an isolated environment (container or VM) if you are unsure about its origin; (3) ensure Playwright and browser downloads come from official sources; (4) confirm the OpenAPI endpoint (/api/search_bilibili) is the only route exposed and that the service is bound to localhost so it cannot be accessed remotely without your intent.

Bilibili 热门视频实时搜索与分析技能 (Bilibili Search Skill)

🌟 技能简介 (Introduction)

这是一个专为 AI 代理 (Agent) 打造的 Bilibili (B站) 数据获取与分析插件。 大模型通常受限于训练数据的知识截止日期，无法获取当天的互联网热点。本技能通过调用本地部署的 Python 自动化接口 (基于 Playwright 和 FastAPI)，让大模型能够实时检索 B 站的最新视频和全站热门内容，从而填补 AI 在时效性上的短板。

🎯 核心功能 (Key Features)

• 实时检索：根据大模型传入的关键词（如“具身智能”、“人形机器人”），实时去 B 站进行搜索。
• 精准去重与提取：跳过广告和无关信息，精准提取搜索结果中排名前列的视频标题。
• JSON 结构化输出：将爬取到的非结构化网页数据转化为标准 JSON 格式，方便大模型进行二次语义分析和内容提炼。
• 静默运行：底层采用 Headless 无头浏览器模式，完全在后台执行，不会干扰用户的桌面操作。

💡 典型使用场景与提示词 (Use Cases & Example Prompts)

当您在聊天框中与 AI 代理沟通时，可以使用以下提示词直接触发本技能：

1. 行业热点追踪：

   "请帮我调用 Bilibili 搜索技能，查一下今天关于‘具身智能’最新的 5 个爆款视频，并总结一下大家都在关注什么技术点？"

2. 特定 UP 主动态：

   "用 B站搜索查一下‘波士顿动力’最近发布的视频标题。"

3. 竞品分析：

   "搜索一下‘宇树科技’，看看他们最新的机器狗测试视频有哪些？"

🛠️ 环境依赖与运行说明 (Prerequisites)

注意：本技能非纯云端应用，它依赖于您本地电脑或服务器上运行的专属 API 服务。 在使用本技能前，请确保：

1. 本地的 FastAPI 服务 (skill_api.py) 已经成功启动。
2. 终端显示 Uvicorn running on http://127.0.0.1:8000。
3. 本地环境已正确安装 playwright 并下载了对应的 Chromium 浏览器内核。

📝 更新日志 (Changelog)

• v1.0.0: 初始版本发布，支持基础关键词搜索与标题提取。