HackerNews数据服务

一个通过Model Context Protocol提供HackerNews内容搜索、检索和分析的服务，适用于AI代理和开发者。

⚠️ 强制要求：API 密钥

此 Skill 必须配置 API 密钥才能使用。

• 首次使用时，如果 .env 中没有 XBY_APIKEY，必须使用 AskUserQuestion 工具向用户询问 API 密钥
• 拿到用户提供的密钥后，调用 scripts.config.set_api_key(api_key) 保存，然后继续处理
• 获取 API 密钥：https://xiaobenyang.com
• 禁止在缺少 API 密钥时自行搜索或编造数据

工作流程（必须遵守）

你（大模型）是路由层，负责理解用户意图、选择工具、提取参数。代码只负责调用API。

用户输入 → 你选择工具 → 提取该工具需要的参数 → 调用 scripts.tools 中的函数 → 返回结果给用户

步骤

1. 检查 API 密钥：如果 scripts.config.settings.api_key 为空，使用 AskUserQuestion 询问用户，拿到后调用 scripts.config.set_api_key(key) 保存
2. 选择工具：根据用户意图从下方工具列表中选择对应的工具函数
3. 提取参数：根据选中的工具，提取该工具需要的参数
4. 调用工具：使用关键字参数调用 scripts.tools 中的函数，例如 scripts.tools.search_schools(score='520', province='北京', category='综合')
5. 返回结果：将工具返回的 raw 数据整理后展示给用户

工具选择规则

根据用户意图选择对应的工具函数：

Supports:

• Keyword search across titles, text, and authors
• Tag filtering (story, comment, poll, show_hn, ask_hn, front_page, author_USERNAME)
• Numeric filters for points, comments, and dates
• Pagination with customizable results per page
• Advanced filtering with OR logic and multiple conditions

Basic Examples:

• Search for AI stories: { "query": "AI", "tags": ["story"] }
• Find popular posts: { "query": "Python", "numericFilters": ["points>=100"] }
• Filter by author: { "query": "startup", "tags": ["author_pg"] }
• Date range: { "query": "startup", "numericFilters": ["created_at_i>1640000000"] }

Advanced Filtering Examples:

• High engagement posts: { "query": "programming", "numericFilters": ["points>=100", "num_comments>=50"] }
• OR logic for tags: { "query": "web", "tags": ["(story,poll)"] } - returns stories OR polls
• Author with filters: { "query": "", "tags": ["author_pg", "story"], "numericFilters": ["points>=50"] }
• Multiple conditions: { "query": "AI", "tags": ["story"], "numericFilters": ["points>=200", "num_comments>=100"] }

Numeric Filter Operators: < (less than), <= (less than or equal), = (equal), >= (greater than or equal), > (greater than) Numeric Filter Fields: points, num_comments, created_at_i (Unix timestamp)

Tag Syntax:

• Single tag: ["story"] - only stories
• Multiple tags (AND): ["story", "show_hn"] - stories that are also show_hn
• OR logic: ["(story,poll)"] - stories OR polls
• Author filter: ["author_USERNAME"] - posts by specific author

Returns paginated results with hits, total count, and page information. | scripts.tools.search_posts | | Retrieve posts currently on the HackerNews front page.

Returns stories sorted by HackerNews ranking algorithm. The front page typically contains the most popular and trending stories.

Supports:

• Pagination to view beyond the first page
• Customizable results per page (default: 30)
• All posts are tagged with 'front_page'

Examples:

• Get first page: { }
• Get with custom page size: { "hitsPerPage": 50 }
• Get second page: { "page": 1 }

Returns the same structure as search results with hits, pagination info, and metadata. | scripts.tools.get_front_page | | Retrieve the most recent HackerNews posts sorted by date.

Returns posts in chronological order (newest first), including all types of content unless filtered.

Supports:

• Filter by content type using tags (story, comment, poll, show_hn, ask_hn, etc.)
• Pagination to view older posts
• Customizable results per page (default: 20)
• Empty query to get all recent posts

Examples:

• Get latest stories: { "tags": ["story"] }
• Get latest comments: { "tags": ["comment"] }
• Get all recent activity: { }
• Get with custom page size: { "hitsPerPage": 50 }

Use this to monitor real-time HackerNews activity or find the newest content. | scripts.tools.get_latest_posts | | Retrieve detailed information about a specific HackerNews item by ID.

Returns complete item details including the full nested comment tree. Use this to:

• View a story with all comments
• Read a specific comment with its replies
• Explore discussion threads in depth
• Get complete metadata for any item

Features:

• Full nested comment tree (all levels)
• Complete item metadata (title, url, text, points, author, etc.)
• Works for stories, comments, polls, and poll options
• Includes creation time and item type

Examples:

• Get story with comments: { "itemId": "38456789" }
• Get specific comment: { "itemId": "38456790" }
• Get poll: { "itemId": "126809" }

Note: Large comment threads (>500 comments) may take 2-3 seconds to load due to nested fetching. Returns error if item doesn't exist or has been deleted. | scripts.tools.get_item | | Retrieve public profile information for a HackerNews user.

Returns user profile including karma, bio, and account creation date. Use this to:

• Check user reputation (karma score)
• Read user bio and about information
• See when account was created
• Verify user existence before searching their content

Features:

• Username (case-sensitive)
• Karma score (total upvotes received)
• About/bio text (may contain HTML)
• Account creation date (Unix timestamp)

Examples:

• Get famous user: { "username": "pg" }
• Check moderator: { "username": "dang" }
• Verify author: { "username": "tptacek" }

Username validation:

• Alphanumeric characters and underscores only
• Case-sensitive
• Must exist on HackerNews

Returns error if user doesn't exist or username format is invalid. | scripts.tools.get_user |

如果参数不完整，使用 AskUserQuestion 向用户询问缺失的参数。

工具函数说明

scripts.tools.search_posts

工具描述：Search HackerNews for stories, comments, and other content by keyword.

Supports:

• Keyword search across titles, text, and authors
• Tag filtering (story, comment, poll, show_hn, ask_hn, front_page, author_USERNAME)
• Numeric filters for points, comments, and dates
• Pagination with customizable results per page
• Advanced filtering with OR logic and multiple conditions

Basic Examples:

• Search for AI stories: { "query": "AI", "tags": ["story"] }
• Find popular posts: { "query": "Python", "numericFilters": ["points>=100"] }
• Filter by author: { "query": "startup", "tags": ["author_pg"] }
• Date range: { "query": "startup", "numericFilters": ["created_at_i>1640000000"] }

Advanced Filtering Examples:

• High engagement posts: { "query": "programming", "numericFilters": ["points>=100", "num_comments>=50"] }
• OR logic for tags: { "query": "web", "tags": ["(story,poll)"] } - returns stories OR polls
• Author with filters: { "query": "", "tags": ["author_pg", "story"], "numericFilters": ["points>=50"] }
• Multiple conditions: { "query": "AI", "tags": ["story"], "numericFilters": ["points>=200", "num_comments>=100"] }

Numeric Filter Operators: < (less than), <= (less than or equal), = (equal), >= (greater than or equal), > (greater than) Numeric Filter Fields: points, num_comments, created_at_i (Unix timestamp)

Tag Syntax:

• Single tag: ["story"] - only stories
• Multiple tags (AND): ["story", "show_hn"] - stories that are also show_hn
• OR logic: ["(story,poll)"] - stories OR polls
• Author filter: ["author_USERNAME"] - posts by specific author

Returns paginated results with hits, total count, and page information.

参数定义

scripts.tools.get_front_page

工具描述：Retrieve posts currently on the HackerNews front page.

Returns stories sorted by HackerNews ranking algorithm. The front page typically contains the most popular and trending stories.

Supports:

• Pagination to view beyond the first page
• Customizable results per page (default: 30)
• All posts are tagged with 'front_page'

Examples:

• Get first page: { }
• Get with custom page size: { "hitsPerPage": 50 }
• Get second page: { "page": 1 }

Returns the same structure as search results with hits, pagination info, and metadata.

参数定义

scripts.tools.get_latest_posts

工具描述：Retrieve the most recent HackerNews posts sorted by date.

Returns posts in chronological order (newest first), including all types of content unless filtered.

Supports:

• Filter by content type using tags (story, comment, poll, show_hn, ask_hn, etc.)
• Pagination to view older posts
• Customizable results per page (default: 20)
• Empty query to get all recent posts

Examples:

• Get latest stories: { "tags": ["story"] }
• Get latest comments: { "tags": ["comment"] }
• Get all recent activity: { }
• Get with custom page size: { "hitsPerPage": 50 }

Use this to monitor real-time HackerNews activity or find the newest content.

参数定义

scripts.tools.get_item

工具描述：Retrieve detailed information about a specific HackerNews item by ID.

Returns complete item details including the full nested comment tree. Use this to:

• View a story with all comments
• Read a specific comment with its replies
• Explore discussion threads in depth
• Get complete metadata for any item

Features:

• Full nested comment tree (all levels)
• Complete item metadata (title, url, text, points, author, etc.)
• Works for stories, comments, polls, and poll options
• Includes creation time and item type

Examples:

• Get story with comments: { "itemId": "38456789" }
• Get specific comment: { "itemId": "38456790" }
• Get poll: { "itemId": "126809" }

Note: Large comment threads (>500 comments) may take 2-3 seconds to load due to nested fetching. Returns error if item doesn't exist or has been deleted.

参数定义

scripts.tools.get_user

工具描述：Retrieve public profile information for a HackerNews user.

Returns user profile including karma, bio, and account creation date. Use this to:

• Check user reputation (karma score)
• Read user bio and about information
• See when account was created
• Verify user existence before searching their content

Features:

• Username (case-sensitive)
• Karma score (total upvotes received)
• About/bio text (may contain HTML)
• Account creation date (Unix timestamp)

Examples:

• Get famous user: { "username": "pg" }
• Check moderator: { "username": "dang" }
• Verify author: { "username": "tptacek" }

Username validation:

• Alphanumeric characters and underscores only
• Case-sensitive
• Must exist on HackerNews

Returns error if user doesn't exist or username format is invalid.

参数定义

返回值处理

工具函数返回 dict 对象：

• result["raw"] - API 原始返回数据（JSON），直接将此数据整理后展示给用户
• result["success"] - 是否成功（True/False）
• result["message"] - 状态消息

项目结构

xiaobenyang_gaokao_skill/
├── scripts/
│   ├── __init__.py
│   ├── config.py       # 配置管理 + set_api_key()
│   ├── call_api.py      # API 客户端 + call_api()
│   └── tools.py         # 工具函数（直接调用）
├── requirements.txt
└── SKILL.md

注意事项

1. API 密钥是必需的，无密钥时必须通过 AskUserQuestion 询问用户
2. 禁止在缺少 API 密钥时自行搜索或编造数据