Date: 2026-11-03. State: accepted by the later instruction to implement the roadmap without interruptions. The values are engineering defaults for simulation, not an invented answer to Q7–Q9 nor an investment recommendation.
Decision#
Keep Brazil/US and exclusively fictitious capital. Adopt the watchlist PETR4, VALE3, ITUB4, AAPL, NVDA, TSLA; long-only and no leverage; closed five-minute candles; US source IEX and B3 CSV import. Three strategy families, immutable configs and disabled by default. Adopt a modular architecture with web and worker, PostgreSQL without mandatory extensions and shadcn/ui Radix.
Risk defaults: 50 bps per trade; daily loss 100 bps; exposure 5000 bps; three positions; simulated cost 5 bps per side and slippage 5 bps. The policy is validated and editable with the account blocked; strategies/accounts start without executing.
Consequences#
Q7–Q9 no longer block implementation; their history remains as not individually answered. Do not reduce Brazil to the US. B3 has a calendar revised for 2026, rejecting years not yet supported; expand with verified sources. No real-time B3 schedule or external B3 broker. Do not connect real money.
Google recovery is an audited CLI and revokes sessions; there is no public bypass. Deploy is authorized as a goal, but requires the owner's credentials. The P6 quantitative observation remains pending real data/period.
Review#
Change defaults upon proven need, with tests and documentation. A change of market, real capital or offer to third parties requires a new scope.