my-skill

• Multi-language Support (ZH/EN/JA): All public skills support Chinese, English, and Japanese output
• Zero-dependency Runtime: Core functionality works with Python 3.10+ only, no external packages required
• Graceful Degradation: Works without LLM/network; optional AI-powered translation
• Production-ready Templates: Battle-tested skill templates with i18n support built-in

One-line Install (Recommended)

Install skills directly using [skills.sh](https://skills.sh):

```bash

# Install all skills from this repository

npx skills add deletexiumu/AgentSkills-Hub

# Or install a specific skill

npx skills add deletexiumu/AgentSkills-Hub/ai-news-digest

npx skills add deletexiumu/AgentSkills-Hub/smart-data-query

npx skills add deletexiumu/AgentSkills-Hub/x-ai-digest

```

Manual Install

```bash

# For Claude Code

mkdir -p ~/.claude/skills

git clone https://github.com/deletexiumu/AgentSkills-Hub.git

cp -R AgentSkills-Hub/skills/public/* ~/.claude/skills/

# For Codex

mkdir -p ~/.codex/skills

cp -R AgentSkills-Hub/skills/public/* ~/.codex/skills/

```

For more details, see: docs/skill-installation.md.

---

Create a new skill (example: place it in skills/public):

```bash

python3 scripts/init_skill.py my-skill --path skills/public --resources scripts,references,assets

```

Validate and package:

```bash

python3 scripts/validate_skill.py skills/public/my-skill

python3 scripts/validate_i18n.py skills/public/my-skill # i18n check

python3 scripts/package_skill.py skills/public/my-skill dist

```

For more workflows and specifications, see: docs/skill-workflow.md.

---

| Skill | Description | Highlights |

|-------|-------------|------------|

| ai-news-digest | Multi-source AI news aggregation and digest generation | 20+ sources, auto-dedup, 5 categories, image export |

| smart-data-query | Intelligent data warehouse Q&A with executable SQL output | Auto-iteration, business questionnaire, Q&A logging |

| x-ai-digest | X platform AI news scraper with reply suggestions | Browser automation, share card generation |

---

> Multi-source AI news aggregation and digest generation with deduplication, classification, and source tracing.

Automatically fetches news from 20+ AI sources (OpenAI, Anthropic, DeepMind, Google AI, TechCrunch, Hugging Face, etc.), intelligently deduplicates and categorizes, and generates daily briefings in Markdown/JSON/Image format.

Why ai-news-digest?

| Advantage | Description |

|-----------|-------------|

| 20+ Curated Sources | OpenAI, Anthropic, DeepMind, Google AI, Meta AI, TechCrunch, Hugging Face, arXiv, and more |

| Smart Deduplication | Same news from multiple sources automatically merged with cross-references |

| 5 Theme Categories | Research / Products / Open Source / Funding / Policy - auto-classified |

| Multi-language Output | Chinese, English, Japanese - use --lang zh/en/ja |

| Natural Language Time | "today", "yesterday", "2026-01-20" - all supported |

| Zero Dependencies | Only Python 3.10+ required for core functionality |

| Multiple Output Formats | Markdown, JSON, shareable PNG images |

| Graceful Degradation | Works without LLM; optional AI translation (Anthropic/OpenAI) |

Usage Examples

In Claude Code / Codex conversation:

```

# Chinese

"Use ai-news-digest to generate today's AI news digest"

# English

"Use ai-news-digest to generate today's AI news in English"

# Japanese

"ai-news-digest で今日のAIニュース要約を日本語で作成して"

```

CLI:

```bash

cd skills/public/ai-news-digest/scripts

# Default (Chinese)

python run.py --day today

# English output

python run.py