agentleFS
grandameniumgrandamenium/cortextos/skills/business-news-monitor/SKILL.md

business-news-monitor

Skill106 starsChanged 7 months agoMITView on GitHub

Daily cron-driven scan of news/forums/social in a domain to surface market shifts, new competitors, regulatory changes, and net-new opportunities. Use when: running a scheduled daily/weekly news pass, or when an idea-grooming pass needs fresh signal. Domain config (sources, keywords, thresholds) lives in agent config.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 10 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    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
  1. Distinctive:

    Read your agent's config for the domain overlay:

    line 16
  2. Distinctive:

    If domain config is missing, surface the gap to the operator and don't silently produce a generic scan.

    line 25
  3. Very specific:

    Every run produces ONE markdown file at <output_path>/<YYYY-MM-DD>.md with this shape:

    line 31
  4. Distinctive:

    Does it affect a current pursue idea or active work?: If yes → material.

    line 75
  5. Very specific:

    Does it match a keywords_critical keyword?: If yes → material.

    line 76
  6. Distinctive:

    Does it represent a new competitor entering the space?: If yes → material.

    line 77
  7. Distinctive:

    Hard cap: 5 material items per day: If the day genuinely has more than 5 material shifts, that's worth a Telegram alert to the operator …

    line 83
  8. Distinctive:

    Don't surface yesterday's news.

    line 89
  9. Distinctive:

    Don't paraphrase headlines.

    line 90
  10. Distinctive:

    Don't make predictions. ("This will lead to X" is out of scope.) Stick to what changed and why we care.

    line 91
  11. Distinctive:

    Don't pad. If only 2 things are material, output 2 items.

    line 92
  12. Distinctive:

    Don't paste raw source content.

    line 93

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Business News Monitor

You scan the domain your agent operates in for material shifts that affect the agent's active work — competitor moves, regulatory changes, technology breakthroughs, new entrants, audience-behavior shifts. Output goes to a structured log that other skills (idea-grooming re-grill, planning convos) can read.

This is not a news aggregator. The output isn't "here are the day's headlines." It's "here are the 1-3 things in your domain that materially changed and why you care."


Inputs (from agent config)

Read your agent's config for the domain overlay:

  • domains: [<list>] — e.g. [financial-newsletters, retail-investor-tools] for fingers, [perfumery, scent-design] for Kevin's perfumer agent
  • news_sources: [<{name, url, kind}>] — RSS feeds, subreddit URLs, newsletter archives, GitHub trending paths. Each entry has a kind (rss / html / subreddit / hn / etc.) so you know how to parse.
  • competitors: [<list>] — known incumbents and direct comparables to track explicitly
  • keywords_critical: [<list>] — phrases that should trigger an alert if they appear (e.g. "Mailchimp acquisition" for fingers, "Firmenich layoffs" for perfumer)
  • keywords_relevant: [<list>] — phrases that should be surfaced if they appear (broader, lower bar)
  • output_path: <path> — where the daily log gets written (typically <vault>/Operating System/News Monitor/<YYYY-MM-DD>.md or similar)

If domain config is missing, surface the gap to the operator and don't silently produce a generic scan.


Output rubric

Every run produces ONE markdown file at <output_path>/<YYYY-MM-DD>.md with this shape:

---
date: YYYY-MM-DD
domains: [<from config>]
sources_scanned: <N>
items_seen: <N>
items_kept: <N>
materiality_threshold: medium
---

# Domain news monitor — <date>

## Material shifts (the only section the operator NEEDS to read)

For each material item (cap at 5 — discipline matters):

### <One-line headline of the shift>

- **What changed:** 1-2 sentences, plain language
- **Why we care:** how it affects active work / pursue ideas / current strategy
- **Source:** <link>, <publication>, <date>
- **Confidence:** high | medium | low (how solid the signal is)
- **Suggested action:** flag a re-grill on idea X / update positioning / no action / investigate further

## Surface (interesting, not material)

Bullet list. 3-10 items. Headlines + 1-line takeaway each. No expansion. These are FYI, not action triggers.

## Quiet today

If the scan found nothing material, say so explicitly. "Quiet today" is a valid output. Do NOT manufacture materiality to fill the section.

## Sources scanned

Inline list of every source visited + count of items pulled. Lets the operator audit coverage.

Materiality test

For every candidate item, ask:

  1. Does it affect a current pursue idea or active work? If yes → material.
  2. Does it match a keywords_critical keyword? If yes → material.
  3. Does it represent a new competitor entering the space? If yes → material.
  4. Does it represent a regulatory or platform change (Apple, Mailchimp, Anthropic, etc.) that touches our stack? If yes → material.
  5. Does it surface a net-new opportunity (new audience, new tech enabling X, new market emerging)? If yes → material.

Otherwise → surface (not material).

Hard cap: 5 material items per day. If the day genuinely has more than 5 material shifts, that's worth a Telegram alert to the operator separately — drop the rest into surface.


Anti-patterns (don't do)

  • Don't surface yesterday's news. Each item must be net-new since the last run.
  • Don't paraphrase headlines. Quote the source headline verbatim, then add the "what changed" interpretation.
  • Don't make predictions. ("This will lead to X" is out of scope.) Stick to what changed and why we care.
  • Don't pad. If only 2 things are material, output 2 items. Empty padding is noise.
  • Don't paste raw source content. Parse, summarize, link. The output should be skimmable in 60 seconds.

Cron pattern

Suggested cadence per domain:

  • Fast-moving domains (financial markets, AI, crypto): daily, weekday morning before the operator's first session
  • Slow-moving domains (perfumery, brand strategy, niche B2B): 2x/week

Cadence lives in the agent's cron config, not here. This skill just defines how the scan runs once invoked.


Cross-skill links

  • Material items with Suggested action: flag a re-grill on idea X should trigger idea-grooming to re-run on that idea
  • The output log is also read by multi-perspective-grilling when generating new perspectives on existing strategy

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.