AI Agent Hub
Back to plugins
🧰

rss-digest

Web Tools Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install JohnXu22786/rss-digest

Paste the following prompt into your AI chat to install this plugin:

Install the plugin in DeepSeek Harness by running dsh plugin install JohnXu22786/rss-digest; the full source code is available at https://github.com/JohnXu22786/rss-digest.

About this plugin

Keeping up with a dozen technical blogs, industry feeds, and podcast RSS streams by hand is fragmented and slow, and the same story often resurfaces across multiple sources with only slight phrasing differences. dsh-rss-digest folds that repetitive overhead into a single pipeline: unified subscription management, background scheduled polling, automatic deduplication, LLM-powered summarization, and a daily Markdown briefing delivered straight into your conversation or written to a file at a set hour.

Across the full information lifecycle, the plugin covers several critical stages. Subscriptions accept RSS 2.0, Atom, and RSS 1.0 (RDF), all persisted in a single versioned JSON document with atomic writes and corruption quarantine. The fetcher is configurable for poll interval, per-request timeout, retry count, and item retention cap. Deduplication pairs exact-hash matching with CJK-aware bigram Jaccard similarity at an adjustable threshold. Summaries default to one batch model call per digest (or per item in single mode), output in Chinese or English, and degrade deterministically to extractive summaries whenever the model is unreachable. Delivery writes Markdown files, pushes them into live Agent sessions, or does both, honoring IANA timezones and DST transitions correctly.

It suits developers and researchers working inside the dsh ecosystem who want feed intelligence woven naturally into their dialogue workflows, as well as anyone who prefers a fully local, zero-third-party-dependency RSS reader. The standalone CLI and the dsh plugin share the same core service, and that core carries no runtime dependencies at all, keeping the tool independent of any particular vendor API.

Use Cases

  • Aggregating multiple tech blogs and news feeds into a single daily Chinese Markdown briefing
  • Subscribing to new feeds on the fly via rss_add in a dsh conversation with instant summarization
  • Degrading gracefully to extractive summaries when no LLM endpoint is reachable

Best For

  • Developers and researchers working within the dsh ecosystem
  • AI-assisted professionals who want feed intelligence woven into their daily dialogue workflow
  • RSS power users who prefer fully local, zero-third-party-service tooling