2.6k Downloads
Overview
Prioritize, read, and summarize high-signal posts from a user’s RSS feeds using the `feed` CLI, producing a focused reading digest.
Key Advantages
1.Agentic end-to-end workflow: fetches feeds, triages entries, reads content, and synthesizes into a digest without manual command orchestration.
2.Built-in bootstrap: if the user has no feeds configured, it can import a curated OPML (e.g., popular HN-adjacent blogs) and retry automatically.
3.High-signal filtering: explicitly designed to pick a small set of unusually insightful or relevant posts rather than dumping all unread items.
4.Token-efficient design: prefers table outputs and fetching full articles via URL instead of loading entire posts into the LLM context.
5.Clear operational constraints: avoids marking entries as read, leaving state changes to the user and reducing unintended side effects.
Use Cases
- Catching up on a large backlog of RSS items by surfacing only the most worthwhile posts.
- Creating a daily or weekly RSS digest grouped by themes for personal reading.
- Exploring a new feed set via the starter OPML when a user has no existing RSS subscriptions configured in `feed`.
- Researching a topic across multiple blogs by combining `feed search` with triage and summarization.
- Producing brief executive-style summaries of long-form blog posts pulled from RSS feeds.
Evaluation Scores
7.9
/ 10
Reliability
7.2
Functionality
8.0
Usability
8.3
Safety
8.7
Performance
7.8
Compatibility
6.5
Based on 1 evaluation · Latest: 3/19/2026
Download Trend
Loading...
Evaluation History (1)
7.9/103/19/2026▼
OS: linux-x64LLM: google/gemini-2.5-flash-lite
**Quick judgment:** Strong, focused skill for turning RSS queues into concise reading digests, especially for users already using or willing to install the `feed` CLI. It’s well-scoped, opinionated, and designed to surface only the most valuable posts rather than everything new.
**Strengths:**
- Automates a full workflow: fetch → triage → read → summarize → present.
- Sensible defaults (unread limit, starter OPML import, table outputs) and clear guidance not to mark items as read.
- Token-conscious design by preferring URL fetching and tabular outputs, which should help performance and cost.
**Risks / Limitations:**
- Hard dependency on the external `feed` CLI (`brew install odysseus0/tap/feed`); environments without Homebrew or that CLI will break the workflow.
- Reliability is sensitive to external web fetching tools and RSS source stability; content extraction quality may vary.
- No explicit content filtering: contrarian or surprising articles might include controversial or sensitive material, relying on the underlying model’s general safety behavior.
**Recommended scenarios:**
- You have an existing RSS setup (or are willing to use the provided OPML starter set) and want a compact, high-signal reading list instead of a raw feed dump.
- You regularly fall behind on blogs or tech news and want an automated agent to scan and summarize the most interesting items.
- You’re building a personal “reading brief” workflow where an agent prepares a theme-grouped summary of notable posts from your subscriptions.
Comments (0)
No comments yet. Be the first!