news-monitor

name

news-monitor

description

Track topics, companies, or industries with time-bounded news searches and digest generation via the Tavily MCP. Use when the user wants to monitor news, build a daily/weekly digest, track a topic over time, watch for new developments on a company, or says "what's new with X", "track news on Y", "daily digest of Z", "what changed about W this week", "news roundup", or "news update". Pairs naturally with scheduled tasks for recurring digests. For one-off deep dives, use deep-research instead.

News Monitor

Build news digests on topics, companies, or industries using tavily_search with time-bounded queries. Best paired with Cowork scheduled tasks for recurring delivery.

When to use this skill

• "What's new with {company}?"
• "Daily digest of AI news / climate news / fintech news / …"
• "Track news on {topic} for me."
• "Has anything changed about {X} this week?"
• "Weekly news roundup on {category}."

For one-off deep dives, use deep-research. For tracking a competitor specifically (pricing, hiring, product launches), use competitive-intel.

Workflow

1. Define scope: get the topic(s), the time window (day, week, month), and the audience.
2. Query: call tavily_search with topic: "news", the appropriate time_range, and max_results: 10 for digest-style coverage.
3. De-noise: drop syndicated copies and press-release reposts. Cluster items by theme.
4. Optionally extract: for the 2–3 most important items, call tavily_extract to get the full body for richer summaries.
5. Format the digest (see template below).
6. Deliver: save to ~~document store, or post to ~~chat, or email via ~~email — match the user's request.

Search patterns

User intent	Query shape
Company watch	query: "{Company} news", time_range: "week", topic: "news"
Industry watch	query: "{Industry} latest developments", time_range: "week"
Topic watch	query: "{Topic}", time_range: "day", topic: "news"
Geographic angle	Add country: "{ISO}"
Trusted-source-only	Add include_domains: "reuters.com,bloomberg.com,..."

For multi-aspect monitoring (e.g. one company + their competitors), issue parallel searches and merge.

Digest template

# {Subject} — News Digest
*{Time window}, {YYYY-MM-DD}*

## Top story
**{Headline}** — {1–2 sentence summary}. [Source]({url})

## What else moved
- **{Headline}** ({source}, {date}) — {1-line summary}. [link]({url})
- …

## Themes this {window}
1. **{Theme}** — {synthesis across multiple stories with citations}
2. …

## Watch list (no movement, but worth knowing)
- {Topic / company} — last update was {when}, what to watch for next.

## Sources
[1] {url}
[2] {url}

Recurring digests

When the user wants this regularly, propose creating a scheduled task:

"I can run this digest every weekday at 8am and drop it in {chat / doc store / email}. Want me to set that up?"

Use the Cowork scheduled-tasks tool to create the schedule. Each run should re-execute this skill with the same scope and post the result.

Tone for digests

Default to factual and crisp — Reuters voice, not blog voice. Don't editorialize. Lead with what changed, not why it matters; the user can decide what matters. Keep each item to one or two lines unless the user specifically asks for analysis.

Failure modes

• Same stories every day (because nothing happened): say so. Don't recycle prior coverage as if it were new.
• All press releases / SEO spam: tighten include_domains to trusted outlets.
• Mixed signals on a company: flag explicitly — "Reuters reports X, but the company's own statement says Y."