원클릭으로
kyrgyz-news
Search, browse, and explore Kyrgyzstan news from multiple agencies via AI-generated summaries
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Search, browse, and explore Kyrgyzstan news from multiple agencies via AI-generated summaries
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | kyrgyz-news |
| description | Search, browse, and explore Kyrgyzstan news from multiple agencies via AI-generated summaries |
| homepage | https://fuddy-duddy.org |
| mcp_server | https://mcp.fuddy-duddy.org/ |
Access Kyrgyzstan news from multiple news agencies through AI-generated summaries, semantic search, and story clustering.
MCP Server: https://mcp.fuddy-duddy.org/
Semantic search across news summaries using natural language queries.
language: "RU" for Kyrgyz-language results (default), language: "EN" for English translationslimit (1-20, default 10)fromDate and toDate in ISO 8601 formatBrowse recent news summaries with pagination.
page (0-based) and pageSize (1-20) for paginationlanguage: "RU" or "EN"Get the full text of a specific news summary by its ID.
search_news or get_latest_news resultsGet trending news clusters — groups of similar articles from multiple sources covering the same story.
language: "RU" or "EN"period: today, yesterday, this_week (default), last_week, this_month, last_monthget_news_articleFind news on a topic:
search_news with a natural language queryget_news_article on interesting results to read the full summary and find related storiesBrowse latest news:
get_latest_news to see recent summariesget_news_article for full details on any itemSee what's trending:
get_top_clusters to find the most-covered storiesget_news_article on the lead article ID for the full story and related coveragelanguage: "RU" for the broadest coveragelanguage: "EN") are available for most summariesAll tools return structured JSON responses with declared outputSchema values (MCP spec v1.0). Each response is a typed object — no markdown parsing required. Empty results preserve query metadata (e.g. query, fromDate, toDate, datesExtracted) so the client always knows what was requested. Not-found cases return { notFound: true, id, similarities: [] } instead of error text.