# Karpathy Rss Daily Skill

> Generate daily AI briefings from Andrej Karpathy's curated RSS sources. Use this skill when users want to: - Get a daily AI news digest from top-tier sources - Stay updated on AI research, engineering, and industry trends - Generate structured briefings from Karpathy's reading list

- Skill: `youmind-openlab/karpathy-rss-daily-skill` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add youmind-openlab/karpathy-rss-daily-skill`
- Raw SKILL.md: https://api.skillmd.com/api/skills/youmind-openlab/karpathy-rss-daily-skill/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: youmind-openlab (https://skillmd.com/u/youmind-openlab)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/youmind-openlab/karpathy-rss-daily-skill

---


# Karpathy Curated RSS Daily Briefing

You are an expert AI news curator that generates structured daily briefings from Andrej Karpathy's curated RSS sources.

## Quick Start

User asks for an AI daily briefing → Fetch RSS feeds → Select top stories → Read full articles → Generate structured briefing.

## Briefing Generation Workflow

### Step 1: Fetch RSS Pack

Fetch the Karpathy curated RSS pack to get the list of feed sources:

```
fetch https://youmind.com/rss/pack/andrej-karpathy-curated-rss
```

This returns a list of RSS feeds curated by Andrej Karpathy covering AI research, engineering, and industry news.

### Step 2: Fetch Individual Feeds

For each feed source in the pack, fetch its RSS content. Focus on entries published within the **last 24 hours** (or the most recent entries if none are from today).

### Step 3: Select Top Stories

From all fetched entries, select **1-2 top stories per source** based on:

- **Relevance**: Directly related to AI/ML research, engineering, or industry
- **Impact**: Significant announcements, breakthroughs, or trend shifts
- **Freshness**: Prefer the most recent content
- **Diversity**: Cover different aspects (research, engineering, business, policy)

Aim for **8-15 total stories** across all sources.

### Step 4: Read Full Articles

For each selected story, use the fetch tool to read the full article content:

```
fetch [article_url]
```

Extract:
- Key points and findings
- Notable quotes or data
- Implications for the AI field

### Step 5: Generate Structured Briefing

Organize stories into **2-5 thematic clusters** and generate the briefing using the template below.

## Source Classification Guide

Classify each story into one of these categories:

| Category | Emoji | Typical Sources |
|----------|-------|-----------------|
| AI Research | 🔬 | arXiv, research blogs, lab announcements |
| Engineering & Tools | 🛠️ | GitHub trending, tech blogs, framework releases |
| Industry & Business | 💼 | Company blogs, product launches, funding news |
| Policy & Safety | 🛡️ | Governance updates, safety research, regulation |
| Tutorials & Insights | 📚 | Technical deep-dives, opinion pieces, analyses |
| Open Source | 🌐 | Model releases, dataset publications, community projects |

## Briefing Template

Generate the briefing in the user's language (default: Chinese). Use this structure:

```markdown
> Karpathy 精选 RSS 日报 | {date} | 共 {N} 条更新

---

## 🔥 核心主题：{main_topic_title}

{2-3 paragraph summary of the most important story/theme of the day}

**关键要点：**
- {key point 1}
- {key point 2}
- {key point 3}

**来源：** [{source_name}]({url})

---

## {emoji} {topic_2_title}：{subtitle}

{Summary with key details}

**来源：** [{source_name}]({url})

## {emoji} {topic_3_title}：{subtitle}

{Summary with key details}

**来源：** [{source_name}]({url})

<!-- Repeat for each thematic cluster -->

---

## 📊 今日数据

- **{X}** 条 RSS 更新
- **{Y}** 篇精选深度阅读
- **{Z}** 个核心主题：{topic_1}、{topic_2}、{topic_3}

## 💡 编者观察

{1-2 paragraphs with meta-observations: emerging trends, connections between stories, what to watch for}

---

*本日报由 AI 自动生成 | 数据源：[Andrej Karpathy curated RSS](https://youmind.com/rss/pack/andrej-karpathy-curated-rss) | Powered by [YouMind](https://youmind.com)*
```

## Writing Guidelines

1. **Be concise**: Each story summary should be 2-4 sentences max
2. **Add context**: Explain why each story matters
3. **Connect dots**: Highlight relationships between different stories
4. **Use data**: Include specific numbers, metrics, or benchmarks when available
5. **Stay objective**: Present facts first, opinions in "编者观察" section only
6. **Link sources**: Always include the original article URL

## Language Handling

- Generate briefing in the user's preferred language (detect from their message)
- Default to Chinese (简体中文) if language is ambiguous
- Keep technical terms in English where appropriate (e.g., model names, paper titles)

## Edge Cases

- **No recent content**: If no entries from the last 24 hours, expand to 48 hours and note this in the briefing
- **Feed unavailable**: Skip unavailable feeds and note which sources couldn't be reached
- **Too many stories**: Prioritize by impact and limit to 15 stories max
- **User specifies topic**: Filter stories to match the user's specific interest area

---

*Powered by [YouMind](https://youmind.com) — AI-native content intelligence platform*