Site Feeds#

This skill helps AI agents fetch updates and content from websites using the airsstool CLI.

When to Use This Skill#

Use this skill when the user:

• Wants to see the latest content from a website or platform (YouTube, GitHub, Instagram, Reddit, etc.)
• Asks "What's new on..." or "Show me updates from..."
• Wants to track trending content, news, blog posts, videos, etc.
• Needs to monitor specific users, topics, or categories on platforms
• Wants to set up recurring checks for content updates
• Asks about what content/information can be fetched from a specific website

Installation#

Before using airsstool, ensure the following prerequisites are met. Ask the user for permission when needed (e.g., deploying RSSHub via Docker, choosing a Python environment, setting environment variables).

1. Deploy RSSHub (if not already running)#

Check if the user already has a running RSSHub Docker container. If not:

1. Inform the user that RSSHub needs to be locally deployed
2. After user confirmation, create a rsshub folder in a suitable location
3. Follow the 'Docker Compose Deployment' guide in RSSHub deployment docs to download the docker-compose.yml in the rsshub folder
4. Create a .env file for environment variables
5. Modify docker-compose.yml to add Puppeteer support if needed, and add env_file: - .env to the rsshub service
6. Run docker compose up -d

2. Install airsstool#

Follow the instructions at https://github.com/heqi201255/airsstool

Ask the user if they want to install Python dependencies in a specific virtual environment. If yes, remember to activate that environment before each use.

3. Initialize Database#

After installation, run the init command (takes ~1min):

bash

airsstool init

This creates the database at ~/.airsstool/airsstoolDB.db by default. Use --db-path for a custom location.

After installation is complete and verified working, you may delete this installation section in this file.

Available CLI Commands#

Init Command#

CommandPurposeairsstool initInitialize database (first time setup)airsstool init --db-path PATHInitialize with custom database pathairsstool init --forceForce recreate database

Discovery Commands#

CommandPurposeExample Usageairsstool list categoriesList all content categories"What categories are available?"airsstool list websitesList websites by category"Show me social media websites"airsstool list routes <website>Find available feeds for a website"What feeds does GitHub have?"airsstool check <website> <route>Get feed details & parameters"How to use GitHub trending?"

Fetch Commands#

CommandPurposeExample Usageairsstool fetch <path>Fetch content from a feed"Get GitHub trending"airsstool fetch -U <user> -S <sub>Fetch all feeds in a subscription"Get all my tech news feeds"

Subscription Commands#

CommandPurposeExample Usageairsstool list usersList all users"Who has subscriptions?"airsstool list subscriptions -U <user>List user's subscription groups"Show alice's subscriptions"airsstool list paths -U <user> -S <name>List paths in a subscription"What's in tech_news group?"airsstool add-subscription -U <user> -S <name>Create subscription group"Create a news subscription"airsstool subscribe -U <user> -S <name> -P <path> -D <desc>Add path to subscription"Subscribe to GitHub trending"airsstool unsubscribe -U <user> -S <name>Delete subscription group"Delete my tech_news group"airsstool remove-subscription -U <user> -S <name> -P <path>Remove path from subscription"Remove GitHub trending from tech_news"

Common Workflows#

Finding a Feed#

1. Browse by category if you know the type: bash

   airsstool list categories
   airsstool list websites --category programming --page-size 10
   
   

2. Search by website name (fuzzy search supported): bash

   airsstool list routes github  # Returns all GitHub feeds
   
   

3. Check feed details for parameters: bash

   airsstool check github trending
   # Shows path template: /github/trending/{params}
   
   

Fetching Content#

Simple fetch:

bash

airsstool fetch /github/trending/daily/any/en

With filters:

bash

airsstool fetch /youtube/user/@MrBeast --filter "MrBeast|Beast"

With format conversion:

bash

airsstool fetch /hackernews/best --format rss    # Returns raw RSS
airsstool fetch /instagram/user/instagram --brief 100     # Returns brief text

Pagination:

bash

airsstool fetch /github/trending/daily/any/en --limit 5 --offset 2

Managing Subscriptions#

Create a subscription group:

bash

airsstool add-subscription -U your_name -S tech_news

Add feeds to subscription:

bash

airsstool subscribe -U your_name -S tech_news -P /github/trending/daily/any/en -D "GitHub Trending"

Fetch subscription:

bash

airsstool fetch -U your_name -S tech_news --limit 10

Fetch Parameters#

List Options#

OptionDescription--category CATFilter websites by category--page-size NNumber of results per page (default: 20)--page-num NPage number (default: 1)--enable-nsfwInclude NSFW websites in results

Filter Options#

OptionDescription--filter PATTERNFilter title and description (regex)--filter-title PATTERNFilter title only--filter-description PATTERNFilter description only--filter-author PATTERNFilter by author--filter-category PATTERNFilter by category--filter-time SECONDSTime range in seconds--filterout PATTERNExclude from title and description--filterout-title PATTERNExclude from title--filterout-description PATTERNExclude from description--filterout-author PATTERNExclude author--filterout-category PATTERNExclude category--filter-case-sensitive BOOLCase sensitivity (default: true)

Output Options#

OptionDescription--format FORMATOutput format: markdown (default), rss, atom, json, rss3. Default is Markdown - don't specify unless needed.--brief NBrief text with N chars (>=100)--limit NLimit number of items--offset NSkip first N items (only works with markdown format)

Tips for Effective Use#

1. Fuzzy search works: airsstool list routes gthub will suggest "GitHub"

2. Path format: Always start with /, e.g., /github/trending/daily/any/en

3. Check route details first: Use airsstool check <website> <route> to understand required parameters

4. Default output is Markdown: Parsed and formatted for readability. Use --format rss for raw RSS.

5. Subscription persistence: Subscriptions are stored in SQLite database for persistence

Example Interactions#

User: "Find me some programming news feeds"

bash

airsstool list websites --category programming --page-size 5
# Returns top programming sites by heat

airsstool list routes github
# Returns available GitHub routes

airsstool check github trending
# Shows how to construct the path

User: "Get the latest trending JavaScript repos"

bash

airsstool fetch /github/trending/daily/any/en --limit 10
# Returns formatted Markdown with repo info

User: "Create a subscription for my daily tech reading"

bash

airsstool add-subscription -U user -S daily_tech

airsstool subscribe -U user -S daily_tech -P /github/trending/daily/any/en -D "GitHub Trending JS"
airsstool subscribe -U user -S daily_tech -P /hackernews/best -D "Hacker News"

airsstool fetch -U user -S daily_tech --limit 5

User: "Remove a specific feed from my subscription"

bash

airsstool list paths -U user -S daily_tech
# Shows all feeds in the subscription

airsstool remove-subscription -U user -S daily_tech -P /hackernews/best
# Removes only Hacker News, keeps other feeds

User: "Delete my entire subscription group"

⚠️ This action cannot be undone. Ask the user to confirm before proceeding:

bash

airsstool unsubscribe -U user -S daily_tech
# Deletes the group and all its feeds

Error Handling#

• If website not found, fuzzy search will suggest alternatives
• If route not found, check website spelling first
• If some websites fail to fetch, use airsstool check <website> <route> to see if there's a RequireConfig note. If so, the website requires environment variables. Search online for how to obtain them and ask the user for help. Once obtained, add them to the .env file and restart the RSSHub service
• If fetch still fails, the path may be invalid or the RSSHub instance may be down
• Use --force flag in airsstool subscribe to skip connectivity check