| name | research-a-ticker |
| description | Use this whenever the operator asks you to research, analyze, or give a read on a specific stock, ETF, or crypto — e.g. "what's the setup on NVDA", "should I be watching SPY", "analyze BTC", "bull/bear case for AAPL". Drives a quote → trend → fundamentals/context → news → structured read with risks. |
| tools | ["stock_quote","stock_price_history","stock_fundamentals","crypto_quote","crypto_price_history","web_search","fetch_url","memory_recall","memory_ingest","current_time"] |
Research a ticker
A disciplined read on one instrument. Analysis, not advice — surface the
setup, the cases, the levels, and the risks; the operator decides.
0. Identify the instrument
Stock/ETF (e.g. NVDA, SPY) → use the stock_* tools. Crypto pair
(e.g. BTC/USDT) → use the crypto_* tools (default exchange okx). If the
operator gives a bare crypto name ("bitcoin"), map it to a BASE/USDT pair.
1. Price & trend (always)
stock_quote / crypto_quote for the current level, day & 52-week range.
stock_price_history / crypto_price_history for trend (default 6mo/1d):
note the return over the window, the high/low, and whether price is near the top
or bottom of its range. Don't invent indicators you didn't compute.
2. Context
- Equities:
stock_fundamentals — valuation (P/E, P/B), margins, growth, sector.
Is it cheap/expensive relative to its own history and sector? State the basis.
- Crypto: there are no fundamentals — lean on trend, volatility, and
web_search
for catalysts (protocol news, flows, regulation).
3. News & catalysts
web_search (and fetch_url to read a source) for recent, dated catalysts —
earnings, guidance, upgrades/downgrades, macro, protocol/regulatory events. Cite
sources with dates. Skip stale or unsourced rumor.
4. Synthesize — the read
Lead with a one-line read (constructive / neutral / cautious — not "buy/sell").
Then:
- Setup: where price sits in its range + trend, with the numbers.
- Bull case / Bear case: 2–3 bullets each, each tied to something you fetched.
- Key levels: support/resistance from the range (52w low/high, recent swing).
- Risks: what breaks the thesis (valuation, regime, liquidity, event risk).
- What I'd watch: the data point or level that would change the read.
Rules
- Every claim ties to a tool result or a cited source. No hindsight, no made-up numbers.
- State assumptions (period, what "cheap" is relative to). A number without its basis is noise.
- Round sensibly. Flag thin data (illiquid name, short history) explicitly.
- If the operator has told you their holdings / risk tolerance,
memory_recall it and
tailor the read; if they share a new standing preference, memory_ingest it (hot).
- Close by reminding that this is analysis, not personalized investment advice.
