Tech Intel

Tech Intel — Daily Market & Tech Pulse

Plain-language guide for aggregating and scoring tech/market news from RSS feeds. Claude fetches articles from 12 sources, scores them by keyword relevance, and presents the top stories as a curated digest.

When to Activate

Activate when the user asks about:

• Tech news, tech intelligence, market intelligence
• Daily newsletter, tech newsletter, market newsletter
• Send tech digest, send market digest
• RSS feeds, news aggregation
• Tech pulse, market pulse

RSS Feed Sources

Fetch from all 12 sources concurrently:

Fetch a Single Feed

curl -s "https://techcrunch.com/feed/" | \
  python3 -c "
import sys, xml.etree.ElementTree as ET
root = ET.fromstring(sys.stdin.read())
for item in root.findall('.//item')[:5]:
    title = item.findtext('title', '').strip()
    link = item.findtext('link', '').strip()
    pub = item.findtext('pubDate', '').strip()
    print(f'{pub[:16]}  {title[:80]}')"

Processing Pipeline

Work through these steps in order:

Step 1 — Filter to Last 24 Hours

Only include articles published within the last 24 hours. Parse <pubDate> (RSS) or <updated> (Atom) fields.

Common date formats:

• Mon, 15 Jan 2024 10:23:01 +0000 (RFC 1123)
• 2024-01-15T10:23:01Z (RFC 3339)
• 2024-01-15T10:23:01-07:00

Step 2 — Deduplicate by URL

Remove articles with the same URL (normalized: strip query params and trailing slashes, lowercase).

Step 3 — Score Each Article

Calculate a relevance score using keyword matching in title + description:

Scoring rules:

• Add base score for each keyword found in title OR description
• Double the score if keyword is in the title (title matches get bonus weight)
• Add +3 recency bonus for articles published in the last 6 hours
• Add +2 category bonus for articles from Markets category sources

Step 4 — Select Top Stories

Sort by score descending, then by recency descending. Select the top 25 articles.

Output Format

Present as a numbered digest:

=== Daily Market & Tech Pulse ===
Date: Monday, 15 Jan 2024
Sources: 12 | Articles scanned: 342 | Top articles: 25

 1. [Score:23] NVIDIA Acquires AI Startup for $2.1B in Semiconductor Push
    TechCrunch | Markets | 10:45 SGT
    https://techcrunch.com/...
    Keywords: acquisition, semiconductor, ai

 2. [Score:18] Fed Chair Signals Rate Cuts Amid Inflation Data Surprise
    CNBC Top News | Markets | 09:12 SGT
    https://cnbc.com/...
    Keywords: regulation, earnings

 3. [Score:15] OpenAI Raises $6B in Latest Funding Round
    The Verge | Tech | 08:30 SGT
    https://theverge.com/...
    Keywords: ai, funding

...

Keyword Scoring Logic (Detailed)

# Pseudocode for scoring one article
def score_article(title, description):
    text = (title + " " + description).lower()
    title_lower = title.lower()
    score = 0
    matched = []

    for keyword, points in keyword_scores.items():
        if keyword in text:
            score += points           # base score
            matched.append(keyword)
            if keyword in title_lower:
                score += points       # title bonus (doubles the score)

    # Recency bonus
    if hours_since_published < 6:
        score += 3

    # Category bonus
    if source_category == "Markets":
        score += 2

    return score, matched

Interpreting Results

High-score articles (>15): Multiple high-value keywords or a title match on a major keyword — these are the day's most significant stories.

Medium-score articles (8–14): Relevant but less impactful. Worth a quick skim.

Low-score articles (<8): Background noise. Include only if filling out the top 25.

What to highlight in the summary:

• Any acquisition or IPO news (market-moving)
• AI/semiconductor developments (sector trends)
• Regulatory changes (policy impact)
• Data breaches or cybersecurity incidents (risk)
• Earnings surprises (market reaction)

References

• TechCrunch RSS
• Hacker News RSS (points>100)
• CNBC RSS Feeds
• Yahoo Finance RSS