news-monitor
Skill10 starsChanged 2 months agoMITView on GitHub
Monitor recent developments on a topic and produce a chronological, source-read briefing. Use when recency or a stated time window is the core question.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- skills-agents-co/skills-and-agents-librarytheir own version, first seen Nov 11, 2026
- AgnesAI-Labs/agnet-skills-cntheir own version, first seen Nov 1, 2026
- claude-office-skills/skillstheir own version, first seen Jan 30, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- claude-office-skills/skills★ 433same name
- AgnesAI-Labs/agnet-skills-cn★ 1same name
- skills-agents-co/skills-and-agents-library★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- article-auditSkill
- competitive-intelSkill
- deep-researchSkill
- fact-checkSkill
- geopolitical-analystSkill
- tech-advisorSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 7 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
Before using tools, read source-access.md completely.
- Very specific: line 20
Use an available news-search tool, preferring a callable name ending in duckduckgo_news_search.
- Very specific: line 32
If uvx cannot run packaged news search, use ddgs without touching the current project environment:
- Distinctive: line 44
If packaged page fetch fails, use Jina Reader:
- Very specific: line 67
Start with 1–2 news searches with distinct angles and [current year] where useful.
- Distinctive: line 88
Complete when: every retained item represents a distinct in-window event with a development date, separate publication date, and candidate …
- Distinctive: line 100
Complete when: every published event has at least one fetched source supporting what happened, its development date, why it matters, and …
- Distinctive: line 117
Use an expanded briefing with 2–4 sentences per event only when the user asks for comprehensive coverage or an event needs that detail to …
- Distinctive: line 122
If successful discovery and filtering find no significant developments, say so and give the searched window.
- Distinctive: line 126
Complete when: all events are distinct and in-window, every event and explanatory claim cites a read source, dates are explicit, and …
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
News Monitor
Find distinct recent events, read a source for every published event, and deliver a chronological briefing.
Tools
Before using tools, read source-access.md completely.
Use an available news-search tool, preferring a callable name ending in
duckduckgo_news_search. Use regular search ending in duckduckgo_search only to fill
coverage gaps. Read pages with a fetch tool ending in web_fetch. Client-added prefixes
vary, so inspect the tools in the session.
Without session tools, run the packaged CLI in an isolated environment:
uvx --python '>=3.10,<3.14' web-forager news "your query" --max-results 10 --output-format json
uvx --python '>=3.10,<3.14' web-forager fetch "https://example.com" --format markdownIf uvx cannot run packaged news search, use ddgs without touching the current
project environment:
uv run --no-project --python '>=3.10,<3.14' --with 'ddgs>=9.5.2' python - <<'PY'
from ddgs import DDGS
results = DDGS().news(query="your query", max_results=10)
for r in results:
print(r["date"], r["title"], r["url"], r["source"])
PYIf packaged page fetch fails, use Jina Reader:
curl -s "https://r.jina.ai/https://example.com"For a YouTube caption failure, report the access gap without forwarding the video URL.
The workflow requires both search and fetch; state what is missing if either is unavailable.
Workflow
1. Set the window
Identify the topic, time frame, and angle. Use the user's window when supplied; otherwise cover the last 2–4 weeks. Ask only when an unresolved ambiguity would materially change the search.
Complete when: the inclusion window and topic boundaries are explicit.
2. Search for events
Start with 1–2 news searches with distinct angles and [current year] where useful.
Combine the results before selecting stories. Check their distribution across the
requested period, relevant subtopics, and independent outlets. If results cluster on
one date, outlet, or subtopic, expand with narrower subperiods, focused queries, and
primary announcements. Use regular search to recover coverage gaps, including a failed
news provider. Keep failed searches separate from successful searches with no matches.
Event counts are presentation targets, not evidence that discovery is complete.
Complete when: relevant subperiods and subtopics have been searched, additional queries mostly repeat known events, and any remaining coverage gaps or tool failures are explicit. One query suffices only when it demonstrably exhausts a narrow topic.
3. Deduplicate and filter
Group coverage of the same event and record both the event date and publication date. Filter by the date of the actual development, then select the best candidate source for each event. A recent republication or retrospective does not make an old event new. Include newly reported historical events only when the disclosure itself is a material in-window development; label it as a new disclosure and distinguish its date from the underlying event date. Aim for 3–7 distinct events.
Complete when: every retained item represents a distinct in-window event with a development date, separate publication date, and candidate source. Unknown dates are marked explicitly; items without a supported in-window development remain candidates until source reading resolves the date, or are omitted with a coverage gap.
4. Read every published event
Fetch at least one authoritative source for every event included in the briefing, including headlines and one-line additional developments. Prefer primary announcements and direct reporting. Search snippets may support headline discovery only; they are not evidence for explanatory claims. If an event's source cannot be read, replace it or omit that event and record the coverage gap.
Complete when: every published event has at least one fetched source supporting what happened, its development date, why it matters, and any stated next step. Recheck window eligibility against the read source before retaining the event.
5. Deliver the briefing
Present most recent first by development date, using the disclosure date for newly reported historical events and showing the underlying date or its uncertainty. The default briefing uses a single event list:
- period covered, current update date, and a one-sentence bottom line;
- 3–5 priority events, each with a headline and 1–2 sourced sentences covering what happened and why it matters;
- remaining significant events, when present, as cited one-line additional developments;
- a short “What to watch” section grounded in sourced upcoming events or unresolved threads.
Use an expanded briefing with 2–4 sentences per event only when the user asks for comprehensive coverage or an event needs that detail to avoid a misleading summary. Each event headline serves as both navigation and detail heading, so the event appears once.
If successful discovery and filtering find no significant developments, say so and give the searched window. If failures or unread sources prevent adequate coverage, report that coverage is incomplete rather than asserting a quiet period.
Complete when: all events are distinct and in-window, every event and explanatory claim cites a read source, dates are explicit, and inference is labeled.
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.

