compare_tickers

Side-by-side comparison of two or more tickers — financial metrics, recent events, and a qualitative read.

When to use

Trigger this skill when the user wants to compare multiple tickers:

• "compare AAPL and MSFT"
• "compare NVDA vs AMD"
• "AAPL GOOGL MSFT — which has the most stable fundamentals?"
• "AAPL MSFT NVDA which is the best buy"

Note: the router LLM handles non-English inputs naturally; equivalent phrasings in other languages route here too.

Do not use when:

• The user only mentions one ticker → use generate_report.
• The user wants raw numbers, no analysis → answer from RAG / direct query.
• The user mentions multiple tickers in passing but is asking about

something else (e.g. "did AAPL or MSFT report yesterday?") → RAG.

Inputs

• tickers (array of strings, required, min 2): ticker symbols, e.g.

["AAPL", "MSFT"]. Case-insensitive; will be uppercased.

• focus (string, optional): comparison angle, e.g. "valuation",

"earnings growth", "recent catalysts". Free-form — passed into the prompt as a hint. Defaults to "overall fundamentals".

Output

• content: a Markdown report with a metrics table + per-ticker

observations + a final ranking/recommendation paragraph.

• artifacts: none (comparisons are ad-hoc; we don't persist them).
• meta: {tickers, focus, ticker_chars: {ticker: int, ...}}.

Notes for routing

When extracting tickers from a free-text message, prefer all-caps sequences of 1–5 letters that look like symbols. If the user only mentions one ticker, do not pick this skill — pick generate_report instead.