daily-news-report
Scrapes content based on a preset URL list, filters high-quality technical information, and generates daily Markdown reports.
Author
Category
Development ToolsInstall
Download and extract to your skills directory
Copy command and send to OpenClaw for auto-install:
Daily News Report - Intelligent Daily News Aggregation and Generation Tool
Skill Overview
Daily News Report is an intelligent news aggregation skill based on a primary-agent orchestration architecture. It can concurrently fetch content from multiple preset technical sources, use AI quality scoring to filter high-value information, and automatically generate a structured Markdown daily report.
Use Cases
Automatically fetch frontier tech news daily from high-quality sources like Hacker News, HuggingFace Papers, and OneUsefulThing, saving time on manual browsing and filtering and enabling quick grasp of industry dynamics and competitive intelligence.
Supports team-shared unified news source configuration and automatically archives filtered technical content into a knowledge base, helping team members stay technically aware and discover directions worth deeper research.
Targets deep content sources such as academic papers and technical blogs, intelligently filters out low-quality marketing content, and prioritizes research results with deep analysis and practical value.
Core Features
Uses a three-layer architecture of primary agent + sub-agents and supports Wave batched parallel execution. The first wave simultaneously fetches core sources like HackerNews and HuggingFace Papers; if high-quality content is insufficient, second and third waves automatically kick in to supplement crawling. A browser rendering engine is used to handle JS-intensive sites like ProductHunt.
Each fetched item is evaluated with multi-dimensional quality scores (1–5) and automatically ranked by depth, usefulness, and timeliness. Built-in filtering rules exclude low-value content such as marketing soft articles, pure academic repackaging, and job postings, ensuring the report focuses on "cutting-edge tech, deep tech, productivity improvements, and actionable information."
Cross-source deduplication is achieved through exact URL matching and title similarity comparison, combined with a cache.json historical fingerprint to avoid duplicate pushes. Supports append mode so multiple runs in a day intelligently merge newly fetched content into the existing report.
Frequently Asked Questions
What news sources does Daily News Report support?
The skill ships with multiple tiered sources by default. Tier 1 core sources include Hacker News (Top 10), HuggingFace Papers (high-vote papers), OneUsefulThing, and Paul Graham's Essays. Tier 2 supplementary sources cover James Clear, FS Blog, HackerNoon, Scott Young, and others. For sites requiring JS rendering (such as ProductHunt, Latent Space), the skill automatically invokes a browser rendering engine for crawling. Users can customize sources by editing sources.json to add or remove news sources.
How does it avoid fetching duplicate or low-quality content?
The skill uses a three-layer deduplication mechanism: exact URL matching, title similarity threshold (80%), and cache.json historical fingerprint comparison. On quality, every item is AI-scored (1–5) and built-in filters exclude types like general science news, marketing soft articles, overly academic rewrites, and pure job postings. The skill follows a "quality over quantity" principle and will rather have fewer report items than include low-quality content.
How long does it take to generate a report?
Execution time depends on source responsiveness and whether browser rendering is required. In the best case (Tier 1 sources sufficient and no browser rendering), about 2 minutes; under normal conditions (Tier 2 supplementation needed), about 3–4 minutes; if JS rendering sites are involved, about 5–6 minutes. The skill includes an "early stop" mechanism: once 20 high-quality items are collected, crawling terminates to avoid unnecessary waiting.