News analysis

News Analysis

Purpose

Source fresh news for the active ticker, persist it to the workspace news panel, then deliver a structured analysis: what happened, why it matters, and what it signals for the investment thesis.

⛔ NEVER do these

What the agent might do	What to do instead
Ask "Should I fetch fresh news or use the cache?"	Always call fetch_latest_news first — it refreshes the panel
Summarize only headlines without any analysis	Add a Signals section with investment-relevant take-aways
Ask which ticker to use	It is always in the session context — never ask
Add a disclaimer about news being incomplete	If data is sparse, note it in chat only

---

Steps

Step 1 — Fetch fresh news

fetch_latest_news(ticker=TICKER, limit=15)

• This refreshes the News panel in the UI (same as clicking "Reload now?").
• Use the ticker from the session context.
• Default limit is 15. Use 20+ if the user wants a deep news scan.

Step 2 — Scan for notable headlines

From the returned list, identify:

• Catalyst events — earnings, guidance, M&A, product launches, regulatory decisions.
• Management / insider signals — executive comments, Form 4 activity, buybacks.
• Macro / sector cross-winds — broader trends affecting the thesis.

If a headline looks highly relevant, fetch the full article body:

fetch_url(url=ARTICLE_URL, max_chars=4000)

Only do this for 1–2 most relevant articles — not for every headline.

Step 3 — Classify sentiment

For each notable headline classify as: Positive / Negative / Neutral for the stock. Do not hedge classifications. Be decisive.

Step 4 — Write the analysis

Output a concise structured analysis:

## News Analysis — {TICKER} ({DATE})

### Key Headlines
- {DATE} | {SOURCE} | {HEADLINE} — **{POSITIVE/NEGATIVE/NEUTRAL}**
  *Why it matters: one sentence.*
- ...

### Signals
- **Catalyst present?** YES / NO — brief explanation.
- **Sentiment trend:** Broadly positive / Mixed / Broadly negative.
- **Thesis impact:** One sentence on what this means for the current thesis.

### Raw headlines (for reference)
{paste full list from fetch_latest_news output}

---

Output rules

• Do NOT include full article bodies in the Thesis — only the analysis.
• Keep "Key Headlines" to the 3–5 most investment-relevant items.
• "Signals" must always be populated, even if news is thin.
• If no news is found: state it clearly in chat → do not create an empty section in the Thesis.

---

When to escalate to deeper research

After the news analysis, suggest follow-up actions if you find:

• A missed earnings catalyst → suggest fetch_latest_news + financials review.
• Insider buying/selling mentioned → suggest fetch_insider_transactions.
• Analyst upgrade/downgrade → suggest fetch_technical_insights.
• M&A rumour → suggest running the activist-feasibility-analysis skill.