IB 0DTE Credit Spreads

⚠️ **Trading warning / Avertissement trading** : informational only, not investment advice. This is the highest-risk skill in this catalogue batch: 0DTE options expire the same day, a credit spread can move from a small credit to its full maximum loss in minutes, and this skill can place real orders on a real Interactive Brokers account when its execute flag is passed. Stops reduce average losses but cannot guarantee an exit price in a fast move or a gap, so the budget-capped maximum loss is the real floor. Paper trading first is not a suggestion here, it is the only sane way to learn this tool. No output here is a promise of return. Curated by Skill Harbor: a finder and executor for zero-days-to-expiration credit spreads (bear call, bull put or iron condor) on cash-settled indices (SPX, NDX, RUT, VIX and others) and any optionable stock or ETF, with all data from IB. The default route is a regime strategy: it reads 30-minute IB bars, checks the volatility index (VXN for NDX and QQQ with a cutoff of 35, VIX otherwise with a cutoff of 20, and both the intraday reading and the prior close must pass), takes the direction from the last EMA 9/21 cross, and skips the trade entirely when the regime fails or the volatility reading is unavailable; two optional confirmation gates (a red-red bar check and a time gate anchored to the morning bars) are off by default. A manual route lets you name the spread type directly. Either way the finder ranks candidates by expected value with the probability of profit shown, sizes the position from a budget that defaults to half of the account's excess liquidity read live from IB, and, only when the execute flag is passed, places the spread as one native combo order with an automatically attached exit bracket: a profit target, a level-anchored stop, and a time exit late in the session, in one cancel-the-others group, with a repair command that checks open 0DTE spreads for missing stops. Every run, dry or executed, is saved as a timestamped JSON trade log. It runs as Python scripts (scripts/ema_vix_0dte.py and scripts/zero_dte.py, needs the ib-async library) from the staskh/trading_skills repo, against TWS or IB Gateway on your own machine. From the staskh/trading_skills repository (MIT). Honest caveats, several of them stated in the skill itself: the expected-value model rests on a realized-volatility ratio of 0.85 that the author explicitly calls an unvalidated prior, not a measured quantity, so at a ratio of 1.0 the modeled edge goes to roughly zero and only the ranking between candidates deserves trust; IB paper accounts often will not fill multi-leg 0DTE index combos at any marketable price, a simulator limit that can make paper results look worse, or simply emptier, than live ones; index options need their own market data entitlement or quotes will not populate; and a dry run is the default precisely because the execute path is one flag away, so read every proposal, its budget source and its maximum loss before ever adding that flag. Skill Harbor never reviews the code, review it yourself before use.