fetch-news
Skill99 starsChanged 6 months agoApache-2.0View on GitHub
Pulls the latest Google News and Hacker News items for every topic and genre in the reader's interests, deduped against every item shown in previous runs.
In this repository27 other files it gives its agents.
- ai-radio/AGENTS.mdAGENTS.md
- customer-support/AGENTS.mdAGENTS.md
- data-analyst/AGENTS.mdAGENTS.md
- document-processor/AGENTS.mdAGENTS.md
- news-curator/AGENTS.mdAGENTS.md
- repo-maintainer/AGENTS.mdAGENTS.md
- seo-auditor/AGENTS.mdAGENTS.md
- audio-mixingSkill
- cover-image-generationSkill
- metadata-generationSkill
- music-generationSkill
- researchSkill
- script-writingSkill
- tts-generationSkill
- memorySkill
- scannerSkill
- data-explorerSkill
- python-dataSkill
- pdf-parsingSkill
- reconciliationSkill
- slide-creatorSkill
- vendor-verificationSkill
- briefingSkill
- curationSkill
- deliverSkill
- github-researcherSkill
- git-workflowSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Use this skill at the start of every sweep.
- Distinctive: line 82
Run this exactly once per sweep.
- Very specific: line 84
Treat items as the complete candidate pool for the sweep.
- Very specific: line 86
If source_errors is non-empty, name the degraded source in the briefing — the counts only cover the sources that responded.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Fetch News Skill
Use this skill at the start of every sweep. It queries two public sources — no account, credentials, or API key required for either — and owns the cross-run dedupe state:
- Google News RSS (
news.google.com) — a search feed per topic plus a headline feed per genre (WORLD,NATION,BUSINESS,TECHNOLOGY,ENTERTAINMENT,SCIENCE,SPORTS,HEALTH) - Hacker News via the Algolia search API (
hn.algolia.com) — stories and comments per topic, for the community-buzz angle - Medium RSS (
medium.com) — long-form posts per tag frominterests.json'smedium_tags, for the practitioner-writing angle
Usage
python3 skills/fetch-news/scripts/fetch_news.pyTopics, genres, and the per-topic cap all come from
.agents/workspace/interests.json — change that file, not the script.
The script:
- Reads the interests profile.
- Fetches every topic (both sources) and every genre (Google News). If a
source fails for one topic, the sweep continues and records the error in
source_errors. - Drops every item whose id is already in
.agents/workspace/seen_items.json, then adds the new ids to that file (bootstrapping it on the first run). - Writes the fresh items to
.agents/workspace/new_items.jsonand prints them to stdout, with a one-line summary on stderr.
Output format (new_items.json)
{
"topics": ["Gemini API"],
"genres": ["TECHNOLOGY"],
"fetched_at": "2026-07-15T09:00:03Z",
"source_errors": [],
"new_count": 42,
"items": [
{
"id": "gn:CBMiXyz…",
"source": "news",
"outlet": "The Verge",
"title": "…headline…",
"context": null,
"url": "https://news.google.com/rss/articles/…",
"published": "2026-07-15T08:12:44Z",
"engagement": null,
"topic": "Gemini API"
},
{
"id": "hn:44581234",
"source": "community",
"outlet": "Hacker News (comment by example_dev)",
"title": "…comment text…",
"context": "Title of the story the comment is on",
"url": "https://news.ycombinator.com/item?id=44581234",
"published": "2026-07-15T07:03:10Z",
"engagement": 12,
"topic": "Gemini API"
}
]
}source is news (Google News), community (Hacker News), or longform
(Medium, with ids prefixed md: and topic: "medium:<tag>"). engagement
is only present for community items (points or comment count). Genre items
carry topic: "genre:<name>".
Instructions for the Agent
- Run this exactly once per sweep. Running it twice in one sweep marks items as seen without you having curated them.
- Treat
itemsas the complete candidate pool for the sweep.new_count: 0means a quiet interval — still write the briefing. - If
source_errorsis non-empty, name the degraded source in the briefing — the counts only cover the sources that responded.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

