| name | strategy-selector |
| description | Recommend an options strategy given the user's directional view, volatility view, risk tolerance, and capital. Use this skill when the user describes a market thesis but does NOT yet know which options structure to use. Triggers: "what strategy should I use", "how should I play this", "I think X will rally / sell off / stay flat", "I want to bet on a move but cap loss", "I want to collect premium", "what's the best way to express this view", or any question where the user gives a thesis and asks for the corresponding multi-leg structure. Do not activate if the user already specifies the strategy — use `options-payoff` instead.
|
Options Strategy Selector
Given a user's market view, return a ranked list of candidate strategies with rationale, then hand off to options-payoff to visualize.
Step 1: Extract User View
Probe for these dimensions (ask follow-ups for any that are missing):
| Dimension | Values |
|---|
| Direction | strongly bullish / mildly bullish / neutral / mildly bearish / strongly bearish |
| Volatility view | IV will rise / IV will fall / IV stays flat |
| Time horizon | days / weeks / months |
| Capital | rough $ size of position |
| Max loss tolerance | $ cap, or "uncapped OK" |
| Skill/experience | beginner / intermediate / advanced |
Refuse to recommend if the user can't articulate at least direction and time horizon — that's not "missing input", it's "no view".
Step 2: Map View to Strategy Grid
| Direction × Vol View | IV will rise | IV will fall | IV flat |
|---|
| Strongly bullish | Long call | Call debit spread, Short put | Call debit spread |
| Mildly bullish | Call debit spread | Bull put credit spread, Covered call | Bull put credit spread |
| Neutral | Long straddle / strangle | Iron condor, Short straddle (advanced) | Iron butterfly, Calendar |
| Mildly bearish | Put debit spread | Bear call credit spread | Bear call credit spread |
| Strongly bearish | Long put | Put debit spread, Short call | Put debit spread |
Capped-loss preference → always pick the debit-spread / credit-spread variant, not the naked leg.
Uncapped-loss OK + small account → reject. Position sizing fails on a single bad day.
Step 3: Apply Filters
Cross out candidates that violate the user's constraints:
- Strict $ cap on loss → eliminate naked shorts (naked put, naked call, short straddle/strangle, ratio spreads).
- Beginner: stick to debit spreads, covered calls, cash-secured puts. No calendars, no ratios, no naked.
- 0DTE preference: only suggest spreads with daily expiries available on the underlying. SPX/SPY have 0DTE; most names don't.
- Earnings catalyst within horizon: prefer long-vol structures (straddle/strangle) since IV crush after earnings will hurt short-vol ones unless you specifically want to SELL the earnings vol pump.
Step 4: Rank Top 3
For each candidate, surface:
- Why this matches the view (1 sentence)
- Approx max profit / max loss at user's capital
- Greeks profile (long/short on each of Δ, Γ, ν, Θ)
- Best- and worst-case scenarios
Step 5: Hand Off to options-payoff
Pick the top-1 candidate, propose strikes anchored to current spot, then call options-payoff to visualize the payoff curve. If the user wants to see alternatives, render one chart per finalist.
Step 6: Respond to User
Present as a comparison table, then ONE strong recommendation with the rationale and a note on the failure mode (what kills this trade). End with a question: "Want me to show the payoff for [top pick] with strikes set at [proposed K1, K2]?"
What This Skill Won't Do
- Predict direction. The user has to bring the thesis.
- Predict vol. Pair with
vol-term-structure / iv-surface if the user wants a vol-regime read.
- Tell the user whether to take the trade. Final go/no-go is theirs.
