news-fetch
Searches for and archives significant global news, focusing on AI, trade, and conflicts.
Install
mkdir -p .claude/skills/news-fetch && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/12534" && unzip -o skill.zip -d .claude/skills/news-fetch && rm skill.zipInstalls to .claude/skills/news-fetch
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Search and archive significant global events (7-day window). Topics: international conflicts, trade/China policy, and AI/Agent updates. Include entity tracking, semantic dedupe against news.jsonl, and atomic archival via recorder.Key capabilities
- →Dynamically anchor dates for a 7-day window
- →Confirm search scope and entities
- →Perform multi-path parallel retrieval by topic
- →Curate candidate lists into a refined set
- →Perform semantic deduplication against historical records
- →Atomically write individual news records
How it works
The skill anchors the search to a 7-day window, confirms the scope and entities, and performs parallel retrieval by topic using `web.run`. It curates results into a set, semantically deduplicates against historical records, and atomically writes each news item.
Inputs & outputs
When to use news-fetch
- →Tracking AI industry updates
- →Archiving trade policy news
- →Researching global conflict reports
About this skill
News Fetch
用于一次性抓取“最近7天”的重大国际事件/政策/AI Agent 动态,并落库到 news 模块。
工作流(保持简洁)
-
动态日期锚定
- 以系统当前日期为 T,窗口为 T-7 至 T(含首尾)
-
确认范围
- 默认范围:国际冲突、贸易政策、中国政策、AI/Agent(含公司/实验室/论文)
- 默认关注实体:OpenAI、Google/DeepMind、阿里、字节、腾讯、智谱、清华相关实验室、幻方量化
- 若用户指定主题(例如“Agent 论文”),以用户主题为主,并加入
scope.topics - 任务结束后将新增主题写回本技能的默认 topics 清单(保证下次可复用)
-
多路并行检索(按 Topic 拆分 Query)
- 必须使用
web.run,时间窗限定最近 7 天 - 每个 topic 单独构建查询,避免热点霸屏
- AI 领域优先:公司官方博客/研究发布 + arXiv
- 政策领域优先:政府官网/央行/监管机构/权威媒体
- 必须使用
-
候选列表 → 精选集合(List + Set)
- 先将检索结果放入“待看 list”(仅收集标题/日期/来源链接)
- 逐条阅读并判断是否纳入“提炼 set”
- set 采用语义去重(AI 判定是否同一事件)
-
语义去重(Smart Dedupe)
- 读取
workspace/records/news/news.jsonl历史记录 - 由 AI 判断是否为同一事件(允许标题/日期轻微差异)
- 来源权威度更高者优先保留
- 写入
dedupe字段(策略+剔除数量)
- 读取
-
原子化写入(Atomic Storage)
- 每条新闻单独写一条
news记录(便于检索/RAG) - 不写
news_digest,仅保留原子化条目 - 使用
recorder脚本:record_jsonl.py --record-type news - 推荐通过
--extra写入结构化内容
- 每条新闻单独写一条
-
News Fetch 自用 TODO
- 每次调用 News Fetch,都先基于模板更新
workspace/records/news/news-fetch_todo_list.md - 清单至少包含:分 Topic 检索、去重、写入 news、补充 coverage_gap、产出摘要
- 完成后将 TODO 归档到
workspace/records/news/news_archive/YYYY/MM/DD/news-fetch_todo.md
- 每次调用 News Fetch,都先基于模板更新
-
List / Set 文件位置
- 待看 list:
workspace/records/news/staging_candidates.jsonl - 提炼 set:
workspace/records/news/curated_set.jsonl
- 待看 list:
-
归档 List / Set(完成任务后)
- 将
staging_candidates.jsonl与curated_set.jsonl归档到:workspace/records/news/news_archive/YYYY/MM/DD/staging_candidates.jsonlworkspace/records/news/news_archive/YYYY/MM/DD/curated_set.jsonl
- 归档完成后再重置两个文件为空白(保留文件)。
- 将
最小可执行步骤(PowerShell)
# 1) 初始化/清空 list 与 set
"" | Set-Content -Path "workspace/records/news/staging_candidates.jsonl" -Encoding utf8
"" | Set-Content -Path "workspace/records/news/curated_set.jsonl" -Encoding utf8
# 2) 初始化 TODO
@'
# News Fetch TODO
- 分 Topic 检索
- 语义去重
- 写入 news(原子化)
- 补充 coverage_gap
- 产出摘要(可选)
'@ | Set-Content -Path "workspace/records/news/news-fetch_todo_list.md" -Encoding utf8
推荐记录结构(news)
{
"date": "YYYY-MM-DD",
"category": "AI 自动判定(可随主题变化)",
"summary": "...",
"tags": ["news", "ai", "agent"],
"sources": [{"name":"...","url":"..."}],
"entities": ["..."],
"source_rank": "official|media|preprint"
}
记录示例(PowerShell)
$extra = @'
{"time_window":{"start":"2026-01-13","end":"2026-01-19","timezone":"local"},"scope":{"topics":["国际冲突","贸易政策","中国政策","AI与Agent"],"entities_watch":["OpenAI","Google/DeepMind","阿里","字节","腾讯","智谱","清华相关实验室","幻方量化"]},"dedupe":{"strategy":"title+source+date","dropped":0}}
'@
python .\\.codex\\skills\\recorder\\scripts\\record_jsonl.py --record-type news --title "单条新闻标题" --summary "单条新闻摘要" --tags "news,ai,agent" --module "news" --source "web" --extra $extra
When not to use it
- →When searching for news outside the 7-day window
- →When topics are not international conflicts, trade/China policy, or AI/Agent updates
- →When not archiving significant global events
Limitations
- →Limited to a 7-day search window
- →Default scope includes specific topics and entities
- →Requires `web.run` for retrieval
How it compares
This workflow includes semantic deduplication and atomic archival, ensuring that news records are unique and stored individually, unlike a simple web search that might return duplicates or unstructured data.
Compared to similar skills
news-fetch side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| news-fetch (this skill) | 0 | 6mo | No flags | Intermediate |
| status | 0 | 6mo | No flags | Beginner |
| literature-review | 559 | 2mo | Review | Advanced |
| openalex-database | 48 | 7mo | Review | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
status
Defiect
Check the progress of the current or most recent deep research run. Shows stage, source counts, claim counts, and any issues.
literature-review
K-Dense-AI
Conduct comprehensive, systematic literature reviews using multiple academic databases (PubMed, arXiv, bioRxiv, Semantic Scholar, etc.). This skill should be used when conducting systematic literature reviews, meta-analyses, research synthesis, or comprehensive literature searches across biomedical, scientific, and technical domains. Creates professionally formatted markdown documents and PDFs with verified citations in multiple citation styles (APA, Nature, Vancouver, etc.).
openalex-database
davila7
Query and analyze scholarly literature using the OpenAlex database. This skill should be used when searching for academic papers, analyzing research trends, finding works by authors or institutions, tracking citations, discovering open access publications, or conducting bibliometric analysis across 240M+ scholarly works. Use for literature searches, research output analysis, citation analysis, and academic database queries.
market-research-reports
davila7
Generate comprehensive market research reports (50+ pages) in the style of top consulting firms (McKinsey, BCG, Gartner). Features professional LaTeX formatting, extensive visual generation with scientific-schematics and generate-image, deep integration with research-lookup for data gathering, and multi-framework strategic analysis including Porter's Five Forces, PESTLE, SWOT, TAM/SAM/SOM, and BCG Matrix.
scientific-brainstorming
davila7
Research ideation partner. Generate hypotheses, explore interdisciplinary connections, challenge assumptions, develop methodologies, identify research gaps, for creative scientific problem-solving.
exa-search
benjaminjackson
Search the web for content matching a query with AI-powered semantic search. Use for finding relevant web pages, research papers, news articles, code repositories, or any web content by meaning rather than just keywords.