About morning-ai
MorningAI is a daily AI news tracking skill for people following model releases, product launches, funding, and benchmarks. It gathers items from six source types, scores and cross-checks them, then produces a Markdown report with links.
Key features
- 01Collects AI news from six source types
- 02Uses two-stage scoring and evaluation
- 03Deduplicates and cross-checks high-scoring items
- 04Creates Markdown reports with source links
- 05Supports language, depth, and exclusion options
Use cases
- 01Preparing a daily AI news briefing
- 02Tracking model releases and product launches
- 03Following funding or benchmark updates
- 04Creating shareable AI news reports
Install
Copy the command and run it in a terminal.
$
npx skills add https:/ / github.com/ davepoon/ buildwithclaude --skill morning-aiAfter installing, type/morning-aiin Claude Code to confirm it loaded.
Details
From the project READMEMorningAI
Daily AI news tracker that collects updates from 80+ entities across 6 sources, scores and deduplicates them, and generates a structured Markdown report.
When to Use This Skill
- When the user wants a daily AI news briefing
- When tracking AI model releases, product launches, funding rounds, or benchmark results
- When generating AI news reports or infographics for sharing
What This Skill Does
- Collects data from 6 sources: Reddit, Hacker News, GitHub, HuggingFace, arXiv, X/Twitter
- Scores items using two-stage scoring (automated metadata + qualitative evaluation)
- Deduplicates and cross-verifies high-scoring items
- Generates a Markdown report with scored entries and source links
Installation
MorningAI requires its full repository for data collection scripts:
# Install as a Claude Code plugin
git clone https://github.com/octo-patch/MorningAI.git ~/.claude/plugins/MorningAI
# Or install as a skill
git clone https://github.com/octo-patch/MorningAI.git
cd MorningAI
How to Use
Basic Usage
/morning-ai
With Options
/morning-ai --lang zh
/morning-ai --depth deep
/morning-ai --exclude Funding
Requirements
- Python 3.9+
- No API keys required (all 6 sources are free)
Links
- Repository: https://github.com/octo-patch/MorningAI
- License: MIT
Comments
Markdown supported