brooksy4503/rss-agent-viewer
Overview
This skill is a CLI RSS/Atom feed viewer designed for AI agents to discover, subscribe, read, and search feeds from the terminal. It includes feed discovery, caching, full-text search, and OPML import/export so agents can manage subscriptions and fetch articles programmatically. The tool is implemented in TypeScript and optimized for automation and scripted workflows.
How this skill works
The skill scans websites for RSS/Atom links, subscribes to feeds, and stores articles in a local database with caching for fast reads. It provides commands to discover feeds via web search, import/export OPML, run full-text searches across all subscriptions, and clean up broken or duplicate feeds. Search can run locally or use an external provider (Exa) when an API key is configured.
When to use it
- When an agent needs to discover RSS/Atom feeds for a website
- To subscribe and read articles from the terminal or automated pipeline
- To run full-text searches across multiple feeds and surface relevant posts
- When importing or exporting subscriptions via OPML for migration or backups
- To automate feed discovery, subscription, and reading in scripts or CI tasks
Best practices
- Initialize the config and database before bulk discovery to avoid duplicate entries
- Use cached reads for fast, offline access and fresh reads when new content is required
- Limit parallel fetches and set sensible --limit values when reading many feeds
- Use discover-search with --auto-add sparingly; verify feeds before automatic subscription
- Protect external API keys (EXA_API_KEY) and prefer local search when privacy is a concern
Example use cases
- Agent discovers feeds for a target domain, adds the best candidates, and reads the latest articles
- Run a scheduled job to fetch and cache all subscribed feeds, then run full-text search to generate summaries
- Import an OPML file from another reader, deduplicate and clean feeds, then export a curated OPML
- Use discover-search with an external provider to find topical feeds, auto-add them, and immediately fetch top results
- Clean up broken or duplicate feeds from the database before syncing subscriptions to another system
FAQ
Yes. Use the search command to run local full-text queries, or discover-search to perform web-backed discovery plus search.
How do I avoid repeated network requests when reading feeds?
Use the --cached flag to read only the database cache, or rely on the built-in cache and run fresh reads only when necessary.
Can I automate discovery and subscription?
Yes. discover-search supports flags like --auto-add and --read for end-to-end automation; combine with environment variables for provider settings.
